Compare commits
127
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
51993c0a2c | ||
|
|
b043f6f1a7 | ||
|
|
29c986cfdc | ||
|
|
171ad422b9 | ||
|
|
f4061f4255 | ||
|
|
8d139cdfc5 | ||
|
|
fae68ecb77 | ||
|
|
661c33a926 | ||
|
|
9b27bc9c4c | ||
|
|
de530e2eac | ||
|
|
7d81a43133 | ||
|
|
0067575b0c | ||
|
|
8cc90794d8 | ||
|
|
d82ed510b9 | ||
|
|
3f06362f9c | ||
|
|
d37e0cd026 | ||
|
|
38bdddded8 | ||
|
|
f0f6ca9f0c | ||
|
|
d47a6743e3 | ||
|
|
ab8389d598 | ||
|
|
5b10b3a23f | ||
|
|
f2d966afd1 | ||
|
|
182abca56f | ||
|
|
c1c9831378 | ||
|
|
6418940761 | ||
|
|
72af858975 | ||
|
|
80e5221403 | ||
|
|
f658ac58b0 | ||
|
|
adb1eaacc4 | ||
|
|
b0f8b5ec55 | ||
|
|
a3984bf656 | ||
|
|
232e1dcdb7 | ||
|
|
ab67a1f77f | ||
|
|
2af3fd9c83 | ||
|
|
9986ded109 | ||
|
|
197ac04a8c | ||
|
|
be2a2a1444 | ||
|
|
76eb4fc7e7 | ||
|
|
057f2bd05a | ||
|
|
172252ab02 | ||
|
|
ebc81ea57e | ||
|
|
bc235e96e5 | ||
|
|
6572b49386 | ||
|
|
d593e01fc9 | ||
|
|
009aa95d98 | ||
|
|
b61b3efc94 | ||
|
|
6cfed25a8e | ||
|
|
27e175aa08 | ||
|
|
6224bae13c | ||
|
|
bd49afae19 | ||
|
|
13e4d73dbd | ||
|
|
37584519e8 | ||
|
|
1ed921b6fb | ||
|
|
15e5803a06 | ||
|
|
a0327f6ddd | ||
|
|
e65612bb28 | ||
|
|
11d79b252d | ||
|
|
7e5cadaa06 | ||
|
|
94d302e112 | ||
|
|
fb26f3ae06 | ||
|
|
11638b0557 | ||
|
|
ad81fbcca1 | ||
|
|
9cc91ffba8 | ||
|
|
0450466422 | ||
|
|
d5442cd659 | ||
|
|
9a1096149a | ||
|
|
0401d9dd34 | ||
|
|
8961827079 | ||
|
|
dea2eac8cd | ||
|
|
3c1eeaa486 | ||
|
|
cf0eec8d31 | ||
|
|
8618d8ca56 | ||
|
|
46dde36185 | ||
|
|
a29ca76fe8 | ||
|
|
2908eac20c | ||
|
|
39a7599fd9 | ||
|
|
c4e519565b | ||
|
|
f18b8f6ca3 | ||
|
|
64ab3e9761 | ||
|
|
065f2cb14c | ||
|
|
3d99ddd35a | ||
|
|
224b4af1da | ||
|
|
0a7177b151 | ||
|
|
b31561d1a0 | ||
|
|
ba4ab486c3 | ||
|
|
4603256ebf | ||
|
|
52b23691dc | ||
|
|
df773658da | ||
|
|
d4636bb5e3 | ||
|
|
83f59cac93 | ||
|
|
431941fc9d | ||
|
|
f5eab116c0 | ||
|
|
6008d31653 | ||
|
|
434817bc76 | ||
|
|
cdd98b13a3 | ||
|
|
f9683347be | ||
|
|
a2853a0ae1 | ||
|
|
b2eb11996b | ||
|
|
36fcf9db4f | ||
|
|
cb21a14424 | ||
|
|
f9ca96f648 | ||
|
|
2d7dd74fe2 | ||
|
|
8a379940bd | ||
|
|
a87ffb84b7 | ||
|
|
89bda6ba11 | ||
|
|
e191f0fc40 | ||
|
|
090dcdad62 | ||
|
|
47d1c4f739 | ||
|
|
a92017cff4 | ||
|
|
cfbc42799a | ||
|
|
799d731ab5 | ||
|
|
5fc8cffc9f | ||
|
|
4dffaf3991 | ||
|
|
bc474f1532 | ||
|
|
c206d9e9da | ||
|
|
b4657e2b58 | ||
|
|
eba880cdee | ||
|
|
b6a67a4c34 | ||
|
|
27525ffdd8 | ||
|
|
eb668c448c | ||
|
|
9b3cd7c769 | ||
|
|
13c303178b | ||
|
|
f945feabb6 | ||
|
|
24ce4ffb06 | ||
|
|
980ed5db8b | ||
|
|
bf59887643 | ||
|
|
60be0ee865 |
@@ -0,0 +1,6 @@
|
||||
{
|
||||
"attribution": {
|
||||
"commit": "",
|
||||
"pr": ""
|
||||
}
|
||||
}
|
||||
@@ -24,6 +24,10 @@ SECURITY.md
|
||||
.claude/
|
||||
/docs/
|
||||
|
||||
# the skill an AI agent installs from the repository (npx skills add);
|
||||
# nothing in the pack loads it
|
||||
/skills/
|
||||
|
||||
# README media: GitHub shows it on the repository page, nothing in the
|
||||
# pack loads it
|
||||
/assets/
|
||||
|
||||
+367
-5
@@ -2,12 +2,374 @@
|
||||
|
||||
All notable changes to ComfyUI-AusBoss are documented here.
|
||||
|
||||
## Unreleased
|
||||
## 2.26.3 - 2026-11-17
|
||||
|
||||
- **ComfyUI starts when the pack is installed twice.** Manager installs it in
|
||||
`custom_nodes/ausboss-nodes` and a git clone in `custom_nodes/ComfyUI-AusBoss`.
|
||||
With both there, ComfyUI used to stop at startup. Now the second copy loads
|
||||
nothing, and the console names both folders so you can delete one.
|
||||
- **Tone match now works on inpaints.** On a Crop For Inpaint 🆎 crop,
|
||||
Stitch Inpaint 🆎 read the colour shift in the soft edge around the
|
||||
painted area. Workflows such as Krea 2 Inpaint Masked let the model paint
|
||||
a little wider than that edge, so the edge came back close to your picture
|
||||
and the slider hardly did anything: in a test on skin it took off a
|
||||
third of the shift at 1. It is now read just inside the painted area and
|
||||
takes off more than half, on a picture of any size. A workflow that already had Tone match on for
|
||||
an inpaint gives a slightly different picture, closer to the original's
|
||||
colour. When the painted area holds something new, it is matched to your
|
||||
picture right beside it where the two look alike, such as new skin where
|
||||
a shirt was. Something different from what is beside it, such as a
|
||||
swimsuit over skin, is left as painted; before, Tone match could tint the
|
||||
whole patch with the opposite colour. Outpaints, blend in and Tone match 0
|
||||
are unchanged.
|
||||
|
||||
- **Copy a crop node's picture without its handles.** Right-click the picture
|
||||
on Image Crop + Rotate + Pad and choose Copy image. By default it copies
|
||||
the original picture. To include crop, rotation, padding and resize, turn
|
||||
on Copy with crop, rotation and padding in the gear menu. Editing controls
|
||||
and guides are never copied. Open image provides a clean
|
||||
picture to copy with the browser on plain HTTP addresses too.
|
||||
|
||||
- **H3 Prompt Writer: a swap reads the new person's looks however you type
|
||||
them.** `@1 looks like ...` on the same line as the swap sentence, and
|
||||
clothes on a line of their own (`@1 is wearing a black suit.`), were
|
||||
left out of the prompt without a word, and the clip often came back
|
||||
unchanged. Both are now used.
|
||||
|
||||
- **H3 Prompt Writer no longer drops lines.** When one line held two
|
||||
people talking, like `The phone says: "Get a job." @1 looks up and says:
|
||||
"What?"`, only the first words reached the prompt. Each quote is its own
|
||||
moment now. A swap's `@1 looks like` line also keeps only the looks: a
|
||||
sentence after it that names an `@` again, like `The trees behind @1 stay
|
||||
the same.`, is read as its own line instead of going into the description
|
||||
with a raw `@1`. `"Hi," she says` is spoken as "Hi." instead of "Hi,.",
|
||||
and a Chinese or Japanese line keeps its own full stop: "。", not "。.".
|
||||
|
||||
- **H3 Prompt Writer can carry a line across a cut.** End the words with
|
||||
`...` and the person keeps talking into the next shot, without saying the
|
||||
words twice. On the last line, `...` lets the clip end mid-sentence. A line
|
||||
too long for its shot carries on the same way instead of a warning.
|
||||
|
||||
- **H3 Prompt Writer makes 4-second clips.** The shortest length went from
|
||||
5 to 4 seconds.
|
||||
|
||||
- **H3 Prompt Writer can put the person from your video somewhere new.** With
|
||||
Copy its moves and no person picture, the person in the video keeps their
|
||||
own look and moves. Give a picture the job place and type `He dances in @1.`
|
||||
Before, the node asked for a person picture and stopped. Copy its moves
|
||||
also takes its length from the video now, like Swap, Change something and
|
||||
New voice.
|
||||
|
||||
- **H3 Prompt Writer can restyle a video from a picture.** With Change
|
||||
something and a style or place picture, the look comes from the picture
|
||||
instead of the video. Type `It looks like @1.` A clip with nobody in it,
|
||||
like an empty hallway, no longer gets a made-up person in its prompt.
|
||||
|
||||
- **H3 Prompt Writer can make a thing do the moves.** With Copy its moves or
|
||||
Swap the person and no person picture, a thing picture your lines make do
|
||||
the action (a cartoon hand, a toy) takes the moves in its own look. A thing
|
||||
that is held or worn stays a prop.
|
||||
|
||||
- **H3 Prompt Writer's job lists take a click again.** Picking a job from a
|
||||
chip's list closed the list without changing the job.
|
||||
|
||||
- **H3 Prompt Writer's job lists match the node.** The pick list on each
|
||||
picture and on the video opens under its chip in the node's own colours
|
||||
and grows with the zoom. Before, it was the browser's small plain list.
|
||||
|
||||
- **An AI agent can now learn the video workflows from the repository.**
|
||||
`npx skills add ausboss/ComfyUI-AusBoss` gives your agent a skill for the
|
||||
MiniMax H3 Pika Person Swap and Pika Remix workflows: how to write the
|
||||
words, which model files to get, and how to run a workflow in your
|
||||
ComfyUI. The same page is at https://ausboss-examples.vercel.app/skill.md.
|
||||
|
||||
- **The "AusBoss was updated" message no longer asks for a reload by
|
||||
mistake.** After the Manager switched the pack to another version, a
|
||||
page reload before ComfyUI restarted brought up the message, telling you
|
||||
the page was older than the pack when it was newer. A restart was what
|
||||
was needed. Now the message only comes up when the pack really is newer
|
||||
than the page, and versions such as 2.25.10 and 2.25.9 are compared as
|
||||
numbers.
|
||||
|
||||
- **LoRA routes only open files on ComfyUI's LoRA list.** The LoRA
|
||||
Loader's info, hash and preview routes used to hand the name they were
|
||||
given to ComfyUI's file lookup and check the answer afterwards. Now the
|
||||
name is matched against the list first, and a name that is not on it is
|
||||
refused before anything on disk is looked up.
|
||||
- **Image Folder 🆎 checks a folder or picture name as text before the
|
||||
disk.** A drive letter in the middle of a name, or a control character,
|
||||
is refused like `..`, and the joined path is checked before it is
|
||||
resolved as well as after. The list no longer sends each picture's size
|
||||
and time; the thumbnail's cache tag is a short hash instead.
|
||||
- **Error messages no longer pass on server paths.** When a model,
|
||||
decoder or encoder fails, its message can name a file on the server's
|
||||
disk. SAM 3 masks, LoRA loading and video saving now keep the kind of
|
||||
error and drop its words when they name a path, since node errors reach
|
||||
the browser and stay in the history.
|
||||
|
||||
## 2.26.2 - 2026-11-14
|
||||
|
||||
- **Loaders never look up a file name they refused.** Before each run,
|
||||
Load Image 🆎, Load Image + Pad 🆎, Image Crop + Rotate + Pad 🆎 and both
|
||||
Video Crop + Rotate + Pad nodes check whether their file changed. When the
|
||||
name was one they do not accept, such as a whole path or a network share,
|
||||
that check still looked it up on disk. On Windows, looking up a share
|
||||
contacts its server and hands it the computer's login. Now a refused name
|
||||
is only compared as text.
|
||||
- **The pack's scripts are found without looking up the address.** The pack
|
||||
tags its scripts with a version as the browser asks for them. It found
|
||||
each one by joining the address the browser sent to its own folder and
|
||||
looking the result up on disk, and only then checked that the file was
|
||||
its own. Now it looks the address up in a list of its scripts made when
|
||||
ComfyUI starts, so no address reaches the disk.
|
||||
- **Save Image's Save button copies only pictures it holds.** With Auto
|
||||
save off, the list of a run's held pictures was a file in ComfyUI's temp
|
||||
folder. That folder takes uploads, so a planted list could copy another
|
||||
file on the server into the output folder. The list now stays in
|
||||
ComfyUI's memory, and a held picture is found only by its run, its number
|
||||
and one of Save Image's formats. The Save button works for the last 256
|
||||
runs, up to 10,000 pictures each; held pictures still go when ComfyUI
|
||||
restarts, as before.
|
||||
|
||||
## 2.26.1 - 2026-11-14
|
||||
|
||||
- **A path that starts with `~` is no longer expanded.** Image Folder used
|
||||
to turn a pasted `~/...` path into a full one before checking that it sits
|
||||
inside ComfyUI's folders; now only a path written out in full is checked.
|
||||
LaMa Inpaint only opens a model from ComfyUI's `models/lama` list. Nothing
|
||||
outside those folders could be read before either; this removes the code
|
||||
that looked as if it could.
|
||||
|
||||
## 2.26.0 - 2026-11-14
|
||||
|
||||
- **Missing LoRAs show in ComfyUI's missing-models list.** When a workflow
|
||||
needs a LoRA you don't have, ComfyUI's Issues panel now lists it with the
|
||||
other missing models, so **Download all** fetches it too. Before, the
|
||||
list skipped every LoRA in a LoRA Loader 🆎, and Run then stopped on
|
||||
"missing LoRA file". A LoRA that moved to another folder, or a row that is
|
||||
switched off, is not listed.
|
||||
|
||||
- **Link tiles and title tiles snap by the edge you see.** A tile draws a
|
||||
card inside a wider see-through frame, and Snap into line used that frame,
|
||||
so tiles lined up with a gap around them, and a tile you dragged never
|
||||
snapped at all. Now a dragged tile lines up, and other nodes line up with
|
||||
it, by the card's own edges.
|
||||
|
||||
- **Cards work in App mode.** App mode is ComfyUI's simple form view of a
|
||||
workflow: a few labelled fields and a Run button. A field from a node
|
||||
with a card (a picture or video loader, H3 Prompt Writer) used to show
|
||||
as an empty row there. Now it shows a real control: a file list with an
|
||||
upload button, a video list, a text box. Back in the graph, the node
|
||||
looks the same as before.
|
||||
- **The video crop nodes take uploads in App mode.** Video Crop + Rotate +
|
||||
Pad and its Clip version now tell ComfyUI that their video list takes
|
||||
uploads, as the picture nodes do. In App mode the clip field gets the same
|
||||
picker as a picture field: your videos as thumbnails, an Upload button and
|
||||
a folder button. On the node itself nothing changes.
|
||||
|
||||
- **Models you already have are picked for you.** When a workflow asks for
|
||||
a model file you keep in a subfolder (`minimax-h3/model.safetensors`),
|
||||
ComfyUI used to mark its loader red until you clicked **Use it** on the
|
||||
workflow's Guide. Now the Guide switches the loader to your copy as the
|
||||
workflow opens and says so in a short message. That matters most in App
|
||||
mode, which shows no Guide to click. With two copies in different
|
||||
subfolders it still asks. Turn it off in Settings, 🆎 AusBoss, Models.
|
||||
|
||||
- **The video clip's timing is three boxes: Start, Duration, Ends at.** On
|
||||
Video Crop + Rotate + Pad (Clip), the Length on/off switch is gone. Drag
|
||||
the start to move the clip, drag the end or type a Duration to change how
|
||||
long it is, and switch between seconds and frames. Details sit in hover
|
||||
text; a short note shows only when something needs fixing, with a button
|
||||
to fix it. Saved workflows open with the same values and run the same.
|
||||
- **A linked frame rate no longer stops the clip's start too early.** With
|
||||
a rate wired in, a typed start of frame 60 used to settle on frame 7; it
|
||||
now lands where you put it, and the start handle stops only where the
|
||||
Duration no longer fits. A rate set inside a subgraph now counts as
|
||||
known, so the clip shows its length before you run.
|
||||
- **Linked clip lengths say where they come from.** A box set by another
|
||||
node shows a **linked** button that jumps to that node. A cap that the
|
||||
video cannot fill says how many frames you get, and a count only known
|
||||
when the graph runs shows the room left after the start.
|
||||
- **H3 Prompt Writer says how much of the clip H3 uses.** When the video
|
||||
comes from the clip node, it shows how many of the clip's frames H3
|
||||
makes and where in the video that stops, and which Durations use all of
|
||||
it.
|
||||
- **Crop + Rotate + Pad panels fill the node's width.** The rows under the
|
||||
picture used to stop at 460 px and sit in the middle of a wide node.
|
||||
- **Load Video uses the same timing boxes as the clip node.** Start,
|
||||
Duration and Ends at replace the IN and OUT fields and the separate
|
||||
Limit row, with the same seconds | frames switch. Saved workflows open
|
||||
with the same values and run the same.
|
||||
- **Load Video counts the frames the file really has.** The face used to
|
||||
work the count out from the length and could say one frame too many (248
|
||||
for a 247-frame clip); it now reads it from the file.
|
||||
- **Load Audio has Start, Duration and Ends at boxes.** They replace the IN
|
||||
and OUT times and the separate Max length row, so the red "cut" stripe is
|
||||
gone: the Duration is the length you get.
|
||||
- **A number wired through a Reroute or a subgraph now shows on the face.**
|
||||
The clip node, Load Video, Load Audio and Layout Boxes read a linked
|
||||
number the same way, so it no longer drops out behind a Reroute, a
|
||||
subgraph or an older Primitive node.
|
||||
|
||||
## 2.25.3 - 2026-11-13
|
||||
|
||||
- **The video clip's frame readout explains itself.** On Video Crop + Rotate
|
||||
+ Pad (Clip), the small `fr 52 · 0:02.17` beside Play now says in plain
|
||||
words, when you point at it, which frame the picture shows and how many
|
||||
seconds into the video it is. The help page explains it too.
|
||||
|
||||
## 2.25.2 - 2026-11-13
|
||||
|
||||
- **The video clip's IN handle always drags.** On Video Crop + Rotate + Pad
|
||||
(Clip), a Length as long as the clip or longer used to lock IN, and the
|
||||
only way to trim the start was to turn Length off. Now IN drags anyway:
|
||||
the Length then only caps the clip, and the clip's own end is OUT. A
|
||||
Length shorter than the clip works as before.
|
||||
|
||||
## 2.25.1 - 2026-11-13
|
||||
|
||||
- **Edit Lock Restore 🆎 takes out the pink tint.** MiniMax H3 Image gives
|
||||
every edit a faint pink, slightly lighter tint over the whole picture.
|
||||
Edit Lock already removed it wherever it locked, but when it stepped
|
||||
aside (a new hair color, a new background, rain) the tint stayed. A new
|
||||
**Remove tint** switch, on by default, reads the tint from the part of
|
||||
the picture the edit left alone and takes it out of the whole picture.
|
||||
Edits meant to recolor everything, such as golden hour or red light, are
|
||||
left as they are. Turn it off to get the old behavior.
|
||||
|
||||
## 2.25.0 - 2026-11-12
|
||||
|
||||
- **New node: Load Image.** Loads a picture and sets its size in the same
|
||||
node, so a workflow needs no resize node after its loader. Size can be a
|
||||
megapixel count, a longest side, a scale, a box the picture fits inside, or
|
||||
a box it fills by cropping, rounded to a multiple you choose. It can also
|
||||
cut a piece or a shape such as 16:9 out of the picture without resizing,
|
||||
and you pick which part is kept: the center, a side or a corner. Arrows
|
||||
step through the pictures in the input folder. Two small pictures on the
|
||||
node show what goes in and what comes out, with both sizes and what is
|
||||
cut off, before you run. It also gives the file's name and its original
|
||||
size. With Size off the picture comes out exactly as the file has it.
|
||||
|
||||
- **Clear VRAM from the right-click menu.** Right-click the empty canvas
|
||||
and pick **Clear VRAM 🆎** to unload every model and free memory on the
|
||||
ComfyUI server, the same as ComfyUI's own "Unload Models and Execution
|
||||
Cache". A short message says when it is done. Turn the entry off in
|
||||
Settings, 🆎 AusBoss, Memory.
|
||||
|
||||
- **Callout 🆎: Info card.** Right-click a Callout and pick **Info card**
|
||||
for a small card with a title, a list of labels and values, and small
|
||||
print, such as a table of run times. Write a row as `label | value`; a
|
||||
row in **bold** is drawn in teal so the one to pick stands out. The first
|
||||
line is the title, and a ⏱️ ℹ️ ⚠️ 💡 or ✅ at its start becomes a small
|
||||
icon. An older pack shows the card as a plain note with its rows written
|
||||
out.
|
||||
|
||||
- **Save Image shows big pictures at once.** The node now shows a small
|
||||
copy of each picture, a few hundred kilobytes, where it used to load the
|
||||
whole file. A 30 MB sheet on a ComfyUI opened from another computer took
|
||||
half a minute to show, or never did; it now shows as the run ends. The
|
||||
saved file is the same full-size file as before.
|
||||
|
||||
- **Save Image: the Save button is hard to miss.** With **Auto save** off,
|
||||
a row over the picture now says **Not saved yet** and holds a solid
|
||||
**Save picture** button. Before, it was a small SAVE label in the corner.
|
||||
With several pictures the button reads **Save all 4**, or **Save this
|
||||
one** when the arrows are on one picture. After a save the row says the
|
||||
name the picture was saved under.
|
||||
|
||||
- **A picture that loads slowly no longer leaves the old one up.** When a
|
||||
result took more than about 16 seconds to reach the browser, the node
|
||||
kept the picture from the run before. Now the new picture always shows,
|
||||
and the old one dims while the new one loads. This fixes Select Frame,
|
||||
Mask Refine, Mask by Name and LaMa Inpaint as well as Save Image.
|
||||
|
||||
## 2.24.0 - 2026-11-12
|
||||
|
||||
- **New node: Edit Lock.** Keeps every part of your picture that an edit
|
||||
did not change exactly where it was, with the colors it had. Edit models
|
||||
redraw the whole picture, so untouched parts often come back a little
|
||||
zoomed, shifted or tinted. Put Edit Lock between the model and the
|
||||
sampler and give it the picture you are editing. It needs no mask. It
|
||||
steps aside when the edit changes the whole picture, and lets rain,
|
||||
snow and haze through while holding the picture in place. Tested on
|
||||
MiniMax-H3-Image.
|
||||
|
||||
- **New node: Edit Lock Restore.** Puts your picture's own pixels back in
|
||||
every part the edit did not change. Add it after the VAE Decode with
|
||||
Edit Lock's `edit_lock` output. It also gives a mask of what the edit
|
||||
changed and a one-line report.
|
||||
|
||||
- **New node: Text Lines.** It runs what comes after once for each line of
|
||||
your text. Type five prompts, one on each line, and one run makes five
|
||||
pictures. It also takes a list a text model wrote: empty lines are
|
||||
skipped, a leading number or bullet comes off, and **Limit** stops a list
|
||||
that ran long.
|
||||
|
||||
- **Save Image shows every picture of a run at once.** When a run saves
|
||||
several pictures, they now show side by side in the node, the way the
|
||||
core Save Image does. The arrows still go to each picture by itself, and
|
||||
round again to all of them. Select Frame, Mask Refine, Mask by Name and
|
||||
LaMa Inpaint show a batch the same way.
|
||||
|
||||
- **Save Image can wait for a click before it saves.** Switch **Auto save**
|
||||
off and a run only shows its pictures on the node. **SAVE ALL** over them
|
||||
keeps every one; step to one picture with the arrows and **SAVE** keeps
|
||||
just that one. Good for making a batch and keeping the best. Auto save is
|
||||
on unless you switch it off, so nothing changes for workflows you have.
|
||||
|
||||
- **New node: Prompt Variations.** A prompt box that picks one choice from
|
||||
each set in curly brackets, such as `a {red|blue|green} {sweater|jacket}`,
|
||||
on every run. The seed decides the picks, so the same seed gives the same
|
||||
prompt again. Queue 50 runs with a random seed and get 50 different
|
||||
prompts, with no wildcard pack.
|
||||
|
||||
- **Mask by Touch: type a name to take it out of the mask.** The name field
|
||||
under the picture now has a sign in front. Click it to switch between **+**
|
||||
(add by name) and **−** (take away by name), then type "face" or "hair" and
|
||||
press Enter. Typing "-face" takes it away whatever the sign says. Before,
|
||||
a typed name was only taken away while Erase was chosen, which nothing on
|
||||
the node said.
|
||||
|
||||
- **Mask by Touch: Ctrl-click only takes away what you point at.** With a
|
||||
sweater picked, a Ctrl-click on the face offered the whole person, and the
|
||||
sweater went with it. Now a spot where nothing is picked says so and
|
||||
changes nothing.
|
||||
|
||||
- **Mask by Touch works on big pictures.** A look at a 3840 pixel picture took
|
||||
twice as long as at 2048, and at 8192 pixels it ran a 32 GB graphics card
|
||||
out of memory. The node now asks SAM 3 about a copy of at most 2048 pixels
|
||||
and brings the answer back to full size. Pictures up to 2048 pixels come
|
||||
out exactly as before. On bigger ones a mask's edge can sit a pixel or two
|
||||
differently.
|
||||
|
||||
- **Mask by Touch no longer fills the queue.** With Image Crop + Rotate + Pad
|
||||
or Load Image + Pad wired in, every change to the loader (each step of its
|
||||
megapixels, for one) queued one more look at a picture that was already
|
||||
replaced. Now only the newest look waits. The button that pauses the
|
||||
looking also says **Pause** and **Look on**.
|
||||
|
||||
- **New node: FlashVSR Upscale 🆎.** Makes a video larger and sharper with
|
||||
the FlashVSR v1.1 model, from 1x to 4x. It needs one model file and no
|
||||
other pack: the help page has the download link. A long clip needs no more
|
||||
graphics memory than a short one, and a result too large for the free
|
||||
memory is made in tiles.
|
||||
The new example `FlashVSR Video Upscale (AusBoss)` shows it between Load
|
||||
Video and Save Video.
|
||||
|
||||
- **Two errors from outside a workflow now say what to do.** When a MiniMax H3
|
||||
run stops with `KeyError: 'latent'` because the Muse MiniMax Director pack is
|
||||
installed, a message names that pack and says to turn it off. When an older
|
||||
ComfyUI cannot read a `w6a8` model file, a message says to update ComfyUI.
|
||||
ComfyUI's own error report stays as it is.
|
||||
|
||||
- **ComfyUI starts when the pack is installed twice.** A Manager install
|
||||
(`custom_nodes/ausboss-nodes`) beside a git clone
|
||||
(`custom_nodes/ComfyUI-AusBoss`) stopped ComfyUI while it started. Now one
|
||||
copy loads and the other steps aside, and the console says which folder to
|
||||
delete. A copy from 2.23.0 or older is the one that loads when it sits
|
||||
beside a newer one, so delete the older folder.
|
||||
|
||||
- **A Workflow Note popup closes when you click outside it.** Opening an
|
||||
info tile (a Tips note, say) showed the full note, and only its Close
|
||||
button or Escape put it away. A click on the dark area around it now
|
||||
closes it too. Selecting text in the note, or a drag that starts on it,
|
||||
keeps it open, and the Edit form still asks before it drops a change.
|
||||
|
||||
## 2.23.0 - 2026-11-09
|
||||
|
||||
|
||||
+9
-3
@@ -193,7 +193,7 @@ example_workflows/ # example workflows (regular workflow JSON, not API JSON)
|
||||
drag). Messages to the user go through `showToast`, never `alert()`.
|
||||
- The mouse wheel over a node zooms the graph, on every part of its face.
|
||||
`js/canvas_passthrough/` hands the wheel on from the pack's panels, and
|
||||
keeps it only for a box that can still scroll or a field being typed in,
|
||||
keeps it only for a box that can still scroll, typed in or not,
|
||||
so a panel adds no wheel listener of its own
|
||||
(`tests/canvas_passthrough.test.mjs` lists the few that exist). After a
|
||||
change to a node face, `scripts/dev/wheel_sweep.mjs` turns the wheel at
|
||||
@@ -210,8 +210,7 @@ example_workflows/ # example workflows (regular workflow JSON, not API JSON)
|
||||
never a classic canvas number widget on a finished node face. Units
|
||||
(`px`, `MP`, `×`) ride inside the box in a fixed-width slot that every
|
||||
single-field row reserves, so the numbers down a card share one centre
|
||||
line. Known, deliberate exceptions: the Seed card's seed (typed) and Load
|
||||
Video's trim timecodes (typed).
|
||||
line. Known, deliberate exception: the Seed card's seed (typed).
|
||||
- **Node faces are widget cards.** A node whose face would be classic
|
||||
LiteGraph widgets (full-width rows with an arrow at each end) gets an
|
||||
entry in `js/widget_cards/index.js` instead: `mountWidgetCard` from
|
||||
@@ -234,6 +233,13 @@ example_workflows/ # example workflows (regular workflow JSON, not API JSON)
|
||||
widget per row. A value that is meant to be wired rather than typed is a
|
||||
backend `forceInput: True` socket (Math Expression `a`/`b`/`c`), and a
|
||||
multi-type string such as `"FLOAT,INT"` accepts either kind of link.
|
||||
- **App mode** is ComfyUI's form view of a workflow (fields and a Run
|
||||
button). A file list from the input folder declares its upload flag
|
||||
(`image_upload`, `video_upload`, `audio_upload`): the form builds the
|
||||
field's Upload button from it, and without it the field is a bare list.
|
||||
`tests/test_video_upload_flag.py` holds the video lists to it. A card's
|
||||
hidden widgets reach the form through `js/app_mode/`, so a card needs
|
||||
nothing of its own for App mode.
|
||||
- Preview-carrying nodes (Select Frame, Mask Refine, Mask by Name, LaMa
|
||||
Inpaint, Save Image) share
|
||||
`js/input_preview/`: a thin bar (node tools left, a small `preview`
|
||||
|
||||
@@ -20,6 +20,7 @@ Most of them are for image and video work. ComfyUI's core nodes already cover th
|
||||
- **[Mask by Touch](#mask-by-touch):** click a thing on the picture and get its mask. Hover shows what a click would take, and **Fine edges** follows hair and fur.
|
||||
- **[Mask by Name](#mask-by-name):** type "the red jacket" and get its mask and a cut-out.
|
||||
- **[Image Crop + Rotate + Pad](#image-crop--rotate--pad):** turn, crop and pad a picture by dragging handles on the node, for outpainting.
|
||||
- **[Edit Lock](#edit-lock):** an edit changes what you asked for, and the rest of the picture stays exactly where it was. No mask needed.
|
||||
- **[Stitch Inpaint](#stitch-inpaint):** after an inpaint or outpaint, everything you didn't ask to change is your original pixels again.
|
||||
- **[Video Crop + Rotate + Pad → Clip](#video-crop--rotate--pad--clip):** trim a clip on a timeline, and pad or crop it to a new shape.
|
||||
- **[Tiled Upscale](#tiled-upscale):** upscale a picture in tiles to the megapixels you set.
|
||||
@@ -48,6 +49,16 @@ After an update, through Manager or `git pull`, restart ComfyUI and hard-refresh
|
||||
|
||||
The pack uses the Pillow, NumPy, Torch and PyAV that come with ComfyUI. Each example lists the models it needs on its Workflow Note. LaMa Inpaint also needs [`big-lama.pt`](https://github.com/Sanster/models/releases/download/add_big_lama/big-lama.pt) in `ComfyUI/models/lama/`. The pack never downloads weights by itself.
|
||||
|
||||
### Using an AI agent
|
||||
|
||||
Your agent (Claude Code, Codex, Cursor or another one) can learn the video workflows: how to write the words for them, set them up, and run them on your computer.
|
||||
|
||||
```bash
|
||||
npx skills add ausboss/ComfyUI-AusBoss
|
||||
```
|
||||
|
||||
Or tell it: "Read https://ausboss-examples.vercel.app/skill.md and help me use these workflows." The skill is [skills/ausboss-workflows/SKILL.md](skills/ausboss-workflows/SKILL.md). For picking and wiring single nodes, point it at the [agent guide](js/docs/agents/README.md).
|
||||
|
||||
### Upgrading from 1.x
|
||||
|
||||
2.0 removed **LM Studio Chat** (use **Text** for a fixed prompt) and Align Image's `offset_x` and `offset_y` outputs. Math Expression's `a`, `b` and `c` are sockets now instead of typed boxes, so put constants in the expression itself. 1.2 removed **Drop Shadow**, **Pad Image** (use **Load Image + Pad**) and **Frame Chooser**. Save Image writes only inside ComfyUI's output folder. The [changelog](CHANGELOG.md) has the details.
|
||||
@@ -79,6 +90,12 @@ It returns the `image`, a `mask` of everything the model paints (padding, see-th
|
||||
|
||||
Krea 2 Rotate + Outpaint turns a photo, pads it and paints the new area in one pass. [Try the Krea 2 Rotate + Outpaint example →](example_workflows/Krea%202%20Rotate%20%2B%20Outpaint%20%28AusBoss%29.json)
|
||||
|
||||
### Load Image
|
||||
|
||||
Loads a picture and sets its size in one node. Pick a file, upload one, or step through the input folder with the arrows. **Size** can stay off, or bring the picture to a megapixel count, a longest side, a scale, a box it must fit inside, or a box it fills by cropping. It can also cut a piece or a shape such as 16:9 out of the picture without resizing, and **Keep** picks which part stays: the center, a side or a corner. **Round to** keeps the result on a multiple of 8, 16 or whatever the model wants. The node shows the picture going in and coming out, with both sizes, before you run, such as `912 × 1632 → 1144 × 2048`.
|
||||
|
||||
It returns the `image`, a `mask` of its see-through parts, the new `width` and `height`, the `filename` for a save prefix, and the file's `original_width` and `original_height`.
|
||||
|
||||
### Load Image + Pad
|
||||
|
||||
Makes a canvas for outpainting: drag the edges out to add room around a picture. The Source list shows the picture under the pointer. The new area can be a solid color, the average color of the nearest edge, the edge pixels smeared outward, a blurred copy of the picture, or a mirror of it. The card also has a seam feather, a canvas multiple and a megapixel budget, and **Exact padding** opens one control per side. A budget of 0 keeps the picture's size, and anything higher resizes the picture before padding. A picture wired into `source_image` is padded instead of the chosen file.
|
||||
@@ -139,6 +156,18 @@ The same watercolor edit without and with Realign to Source: 61 px off before, 0
|
||||
|
||||
Qwen drew this watercolor 5.6% taller, and Realign to Source lined it back up with the original.
|
||||
|
||||
### Edit Lock
|
||||
|
||||
Keeps every part of your picture that an edit did not change exactly where it was, with the colors it had. An edit model redraws the whole picture, so the parts you did not ask about often come back a little zoomed in, shifted or tinted, and the picture jumps when you flip between before and after. Put Edit Lock between the model and the sampler and give it the picture you are editing. It needs no mask: it sees what the edit changes while the picture is being made, and holds on to everything else.
|
||||
|
||||
When the edit changes the whole picture (a new style, new light), there is nothing to keep, so the node steps aside and the edit runs as usual. A soft effect over the whole picture, such as rain or falling snow, is let through as well: the node then only holds the picture in place. **Lock** (`lock`) set to `off` does nothing at all, `always` locks even a whole-picture edit, and an optional `mask` says where the edit may happen. **Tolerance** (`tolerance`) sets how big a change must be to count as part of the edit.
|
||||
|
||||
On MiniMax-H3-Image, 4 megapixel edits moved the picture by up to 14 px without the lock (measured at 1 megapixel) and by 0.2 px or less with it (0.08 px or less on 35 of the 36 locked runs), on 19 local edits at two seeds each. After five edits in a row on one picture the plain chain had drifted 1.65 px and turned visibly pink; the locked chain sat at 0.04 px with no tint.
|
||||
|
||||
### Edit Lock Restore
|
||||
|
||||
Puts your picture's own pixels back in every part the edit did not change. Edit Lock keeps those parts in place, but they still pass through the VAE, which softens fine detail a little. Add this right after the VAE Decode and wire in Edit Lock's `edit_lock` output. **Feather** (`feather`) sets how softly the original fades into the edit. It also returns `edit_mask`, white where the edit changed the picture, and a one-line `report` of how much was kept.
|
||||
|
||||
### Image Compare A/B
|
||||
|
||||
Compares two pictures with a slider, or flips between them whole. Nothing covers the picture, and its size shows underneath. The panel grows with the node, and images A and B pass through, so you can put it anywhere on an image wire. **Erase** mixes the two by hand: rub parts of one picture away with a brush so the other shows through, then save the mix as a PNG under a name you type, with a layered PSD for Photoshop or GIMP if you want it.
|
||||
@@ -161,13 +190,13 @@ The example keeps the clip's audio when there is any. The outpaint LoRA needs a
|
||||
|
||||
### Load Video
|
||||
|
||||
Upload a video, or pick one from the Source list, which plays the clip under the pointer. Trim it with IN and OUT handles or typed timecodes. The player plays the part you kept and can loop it. Width and height resize while decoding, and setting only one keeps the shape. `every_nth` skips frames and lowers the reported fps so the timing stays right, and `max_frames` caps how many frames load.
|
||||
Upload a video, or pick one from the Source list, which plays the clip under the pointer. Set the part you want with Start, Duration and Ends at, in seconds or frames, or drag the handles under the player. The player plays the part you kept and can loop it. Width and height resize while decoding, and setting only one keeps the shape. `every_nth` skips frames and lowers the reported fps so the timing stays right, and the Duration caps how many frames load.
|
||||
|
||||
It returns the frames, audio, frame count, fps, width, height, duration and a core `VIDEO`. Only the part you kept is decoded, and it checks memory before a large load. When you save the frames, wire its fps into the save node.
|
||||
|
||||
### Load Audio
|
||||
|
||||
Upload a sound file, or pick one from the Source list, and the node shows it as a wave. Drag the IN and OUT handles to keep only the part you want, or type the times. Press play to hear just that part. Set **Max length** to the longest sound the next step can take, such as a video's length: anything past it turns red and is cut off, so you see what comes out before you run. Click the red length next to OUT to move OUT to the cut.
|
||||
Upload a sound file, or pick one from the Source list, and the node shows it as a wave. Drag the two handles to keep only the part you want, or type a Start, a Duration or an Ends at. Press play to hear just that part. Set the Duration to the length the next step needs, such as a video's length.
|
||||
|
||||
It returns the `audio` you kept and its length in `seconds`. Only that part is decoded. [Help page](js/docs/AUSBOSS_NODES_LoadAudio.md).
|
||||
|
||||
@@ -177,7 +206,7 @@ Picks one frame from a video, then gives you the same rotate, crop and pad contr
|
||||
|
||||
### Video Crop + Rotate + Pad → Clip
|
||||
|
||||
Applies one rotate, crop and pad to every frame of a trimmed clip. It has the source picker, a timeline with a playhead and frame-accurate IN and OUT handles, frame skipping, a frame limit, the transform handles and output resizing. While you drag a trim handle, the node shows the exact first or last frame the run keeps. The bright part of the selection is what gets output.
|
||||
Applies one rotate, crop and pad to every frame of a trimmed clip. It has the source picker, a timeline with a playhead and frame-accurate start and end handles, Start, Duration and Ends at boxes, frame skipping, the transform handles and output resizing. While you drag a trim handle, the node shows the exact first or last frame the run keeps. The bright part of the selection is what gets output.
|
||||
|
||||
Fill, feather and the resize budget are in **Crop, pad & resize** below the timeline, and a new clip keeps them. That matters for video outpainting, because the black, hard-edged canvas the model needs survives a source swap. A fresh Clip node starts with that canvas, and the editor's **Reset transform** goes back to it. **Frames for** trims the end to the frame count the next video model takes: LTX (8n+1) or Wan (4n+1). Tap a ratio to pad the clip to it; a new clip is padded to the lit ratio too. The padlock keeps the shape while you drag a handle.
|
||||
|
||||
@@ -199,6 +228,12 @@ Changes a clip's frame rate by making in-between frames. You set the source and
|
||||
|
||||
It returns the new frames and their fps. It doesn't touch audio, so carry the source audio to Save Video separately. The [help page](js/docs/AUSBOSS_NODES_FrameInterpolate.md) covers the weights and timing.
|
||||
|
||||
### FlashVSR Upscale
|
||||
|
||||
Makes a video larger and sharper with the FlashVSR model. Set **Scale** from 1 to 4 and it returns every frame at the new size, with fine detail drawn back in. It reads the clip a few frames at a time, so a long clip needs no more graphics memory than a short one, and a result too large for the free memory is made in tiles.
|
||||
|
||||
It needs one model file in `ComfyUI/models/diffusion_models` and no other pack. The [help page](js/docs/AUSBOSS_NODES_FlashVSRUpscale.md) has the download link, the memory it needs and the speed. [Try the FlashVSR Video Upscale example →](example_workflows/FlashVSR%20Video%20Upscale%20%28AusBoss%29.json)
|
||||
|
||||
### Save Video
|
||||
|
||||
Saves a frame batch or a core `VIDEO` as MP4 (H.264 or H.265), WebM (VP9 or AV1), ProRes MOV, lossless FFV1 MKV, GIF or WebP. Which encoders you get depends on your FFmpeg and PyAV build, and H.264 and H.265 come in CPU and NVENC versions. The videos carry color tags, and optionally audio and the workflow.
|
||||
@@ -344,6 +379,14 @@ Puts batch `b` after `a`. When the sizes match, nothing is resized. When they di
|
||||
|
||||
A text box that grows with the node and sends its text, unchanged, down a `STRING` wire. Use it for a prompt, caption, path or other text that several nodes share. Spaces and line breaks are kept. Ctrl + Up and Ctrl + Down change the weight of the selected words, as in ComfyUI's own prompt boxes.
|
||||
|
||||
### Text Lines
|
||||
|
||||
Runs what comes after once for each line of your text. Type five prompts, one on each line, and one run makes five pictures. It also takes a list another node wrote, skips empty lines, and can take a leading `1.` or `-` off each line. `number` counts the lines from 1, to give each its own seed.
|
||||
|
||||
### Prompt Variations
|
||||
|
||||
A prompt box that changes on every run. Write choices in curly brackets, split by `|`, such as `a {red|blue|green} {sweater|jacket}`, and the node picks one from each set. Wire in the seed your sampler uses: a new seed gives a new mix, and the same seed gives the same prompt again. Set the seed to randomize and queue 50 runs to get 50 different prompts. Text without brackets comes out exactly as typed, so a JSON prompt is safe.
|
||||
|
||||
### H3 Prompt Writer
|
||||
|
||||
Type what happens, one line for each moment, and leave a blank line to cut to a new place. Connect up to four pictures and give each one a job: a person, a place, an outfit, a thing or a style. Copy the moves from a video, keep a clip going, or add a voice or a song. A timeline on the node shows when each moment plays before you run it. It needs no extra model and you never see H3's prompt format. Connect a small text model and it fills in a one-line idea.
|
||||
@@ -396,7 +439,7 @@ Turns parts of a workflow off and on from one small card. Select the nodes of on
|
||||
|
||||
## Example workflows
|
||||
|
||||
These are the workflows where the nodes do something core ComfyUI doesn't do by itself: outpainting and inpainting that put your original picture back, lining an edit back up with its source, upscaling a picture in tiles, masking a thing by its name or with a click, and removing a watermark from a clip. Open a JSON file from [`example_workflows/`](example_workflows), or find them under this pack in ComfyUI's template browser. Each one has a thumbnail, numbered groups and a Workflow Note with the setup steps and model downloads.
|
||||
These are the workflows where the nodes do something core ComfyUI doesn't do by itself: outpainting and inpainting that put your original picture back, lining an edit back up with its source, upscaling a picture in tiles, masking a thing by its name or with a click, removing a watermark from a clip, and making a clip larger and sharper. Open a JSON file from [`example_workflows/`](example_workflows), or find them under this pack in ComfyUI's template browser. Each one has a thumbnail, numbered groups and a Workflow Note with the setup steps and model downloads.
|
||||
|
||||

|
||||
|
||||
@@ -417,6 +460,7 @@ The workflows open with their image and video loaders empty, so you load your ow
|
||||
| [Mask by Name](example_workflows/Mask%20by%20Name%20%28AusBoss%29.json) | Type what to find in a picture, and get its mask and a cut-out with a see-through background | ComfyUI's own SAM 3.1 (ComfyUI 0.20 or newer) |
|
||||
| [Mask by Touch](example_workflows/Mask%20by%20Touch%20%28AusBoss%29.json) | Click a thing in a picture, or its whole subject, and get its mask and a cut-out with hair-fine edges | ComfyUI's own SAM 3.1 and its background-removal model (ComfyUI 0.21 or newer) |
|
||||
| [Simple Video Watermark Remover](example_workflows/Simple%20Video%20Watermark%20Remover%20%28AusBoss%29.json) | Find an overlay in a clip and remove it, with a one-frame comparison branch | ComfyUI's own SAM 3.1 (ComfyUI 0.20 or newer), plus `big-lama.pt` |
|
||||
| [FlashVSR Video Upscale](example_workflows/FlashVSR%20Video%20Upscale%20%28AusBoss%29.json) | Make a clip larger and sharper with one node | `flashvsr_v1.1_fp16.safetensors` |
|
||||
|
||||
Every example uses only core nodes and this pack. Speed and memory use depend on the models, the sizes, the frame count and whatever else is using the GPU.
|
||||
|
||||
@@ -452,7 +496,7 @@ People who download your workflow get the nodes through ComfyUI-Manager's missin
|
||||
- Hold Shift while you turn the green rotation handle to snap to 15° steps.
|
||||
- In the full-screen editor, the mouse wheel zooms, and the middle mouse button or Alt-drag pans. On the node itself, the wheel and middle mouse button still move the graph.
|
||||
- **Alt+E** opens the selected node's transform editor. Change the key in Settings → Keybindings.
|
||||
- The video editors have frame stepping, playback and exact timeline seeking, on the node face as well as in the editor. The Clip node's IN and OUT handles sit on the source's own frames, so the frame shown for IN is the first one decoded and the frame shown for OUT is the last.
|
||||
- The video editors have frame stepping, playback and exact timeline seeking, on the node face as well as in the editor. The Clip node's start and end handles sit on the source's own frames, so the frame shown at the start is the first one decoded and the frame shown at the end is the last.
|
||||
- Your own aspect presets go in `ausboss_presets.json` beside the pack. Copy [`ausboss_presets_example.json`](ausboss_presets_example.json) to start. Your file survives updates.
|
||||
|
||||
LoRA Loader and Latent Size have their own gear menus for per-node preferences.
|
||||
|
||||
+38
-34
@@ -7,26 +7,12 @@ name to NODE_MODULES below — nothing else to wire up.
|
||||
|
||||
A module that fails to import is skipped with a console warning instead
|
||||
of taking the whole pack down, so one broken node never hides the rest.
|
||||
|
||||
Only one copy of the pack loads. People install it from ComfyUI-Manager
|
||||
(custom_nodes/ausboss-nodes) and from GitHub (custom_nodes/ComfyUI-AusBoss),
|
||||
and two copies used to register the same server routes, which stopped
|
||||
ComfyUI at startup. A second copy now loads nothing and says which folder
|
||||
to delete.
|
||||
"""
|
||||
|
||||
import importlib
|
||||
import os
|
||||
import sys
|
||||
import tomllib
|
||||
import traceback
|
||||
import types
|
||||
|
||||
_HERE = os.path.dirname(os.path.abspath(__file__))
|
||||
_FIRST_COPY = sys.modules.setdefault(
|
||||
"_ausboss_nodes_first_copy", types.SimpleNamespace(folder=_HERE)
|
||||
)
|
||||
_IS_SECOND_COPY = _FIRST_COPY.folder != _HERE
|
||||
|
||||
# One entry per file in nodes/, kept alphabetical. The order has no
|
||||
# functional effect.
|
||||
@@ -36,6 +22,8 @@ NODE_MODULES = [
|
||||
"node_callout",
|
||||
"node_color_match",
|
||||
"node_compare",
|
||||
"node_edit_lock",
|
||||
"node_flashvsr_upscale",
|
||||
"node_frame_interpolate",
|
||||
"node_free_memory",
|
||||
"node_h3_prompt_writer",
|
||||
@@ -49,6 +37,7 @@ NODE_MODULES = [
|
||||
"node_lama_inpaint",
|
||||
"node_layout_boxes",
|
||||
"node_literals",
|
||||
"node_load_image",
|
||||
"node_load_image_pad",
|
||||
"node_load_audio",
|
||||
"node_load_video",
|
||||
@@ -56,6 +45,7 @@ NODE_MODULES = [
|
||||
"node_mask_by_name",
|
||||
"node_mask_by_touch",
|
||||
"node_math",
|
||||
"node_prompt_variations",
|
||||
"node_realign_to_source",
|
||||
"node_refine_mask",
|
||||
"node_resolution",
|
||||
@@ -66,6 +56,7 @@ NODE_MODULES = [
|
||||
"node_select_frame",
|
||||
"node_show_text",
|
||||
"node_step_preview",
|
||||
"node_text_lines",
|
||||
"node_tiled_upscale",
|
||||
"node_video_crop_rotate_pad",
|
||||
"node_video_crop_rotate_pad_clip",
|
||||
@@ -77,8 +68,18 @@ NODE_CLASS_MAPPINGS = {}
|
||||
NODE_DISPLAY_NAME_MAPPINGS = {}
|
||||
_failed_modules = []
|
||||
|
||||
# Two copies of the pack (a Manager install beside a git clone) add every
|
||||
# route twice and ComfyUI stops while starting. Only one copy loads.
|
||||
_other_copy = None
|
||||
try:
|
||||
from .nodes._single_copy import other_copy_that_loads
|
||||
|
||||
_other_copy = other_copy_that_loads(os.path.dirname(__file__))
|
||||
except Exception:
|
||||
traceback.print_exc()
|
||||
|
||||
for _module_name in NODE_MODULES:
|
||||
if _IS_SECOND_COPY:
|
||||
if _other_copy:
|
||||
break
|
||||
try:
|
||||
_module = importlib.import_module(f".nodes.{_module_name}", __name__)
|
||||
@@ -92,15 +93,15 @@ for _module_name in NODE_MODULES:
|
||||
|
||||
# Folder ComfyUI serves to the browser; every .js file in it is loaded as a
|
||||
# frontend extension. Shared modules use .mjs so they are import-only.
|
||||
WEB_DIRECTORY = None if _IS_SECOND_COPY else "./js"
|
||||
WEB_DIRECTORY = None if _other_copy else "./js"
|
||||
|
||||
__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY"]
|
||||
|
||||
|
||||
def _read_version(folder=None):
|
||||
def _read_version():
|
||||
"""Version straight from pyproject.toml so it is never duplicated."""
|
||||
try:
|
||||
pyproject = os.path.join(folder or os.path.dirname(__file__), "pyproject.toml")
|
||||
pyproject = os.path.join(os.path.dirname(__file__), "pyproject.toml")
|
||||
with open(pyproject, "rb") as f:
|
||||
version = tomllib.load(f).get("project", {}).get("version")
|
||||
return version if isinstance(version, str) and version.strip() else "unknown"
|
||||
@@ -118,17 +119,6 @@ def _print_banner():
|
||||
bar = f"{teal}{'-' * 70}{reset}"
|
||||
|
||||
print(bar)
|
||||
if _IS_SECOND_COPY:
|
||||
def where(folder):
|
||||
text = f"{os.path.basename(folder)} (v{_read_version(folder)})"
|
||||
return text.encode("ascii", "replace").decode("ascii")
|
||||
|
||||
print(f" {red}AusBoss is installed twice, so this copy is not loaded.{reset}")
|
||||
print(f" In use: custom_nodes/{where(_FIRST_COPY.folder)}")
|
||||
print(f" Skipped: custom_nodes/{where(_HERE)}")
|
||||
print(" Delete one of the two folders and restart ComfyUI.")
|
||||
print(bar)
|
||||
return
|
||||
print(
|
||||
f" {teal}AusBoss{reset} v{_read_version()} | "
|
||||
f"{len(NODE_CLASS_MAPPINGS)} nodes loaded"
|
||||
@@ -139,12 +129,26 @@ def _print_banner():
|
||||
print(bar)
|
||||
|
||||
|
||||
try:
|
||||
if not _IS_SECOND_COPY:
|
||||
def _print_second_copy():
|
||||
# ASCII only, as in the banner.
|
||||
red = "\033[0;31m"
|
||||
reset = "\033[0m"
|
||||
own = os.path.basename(os.path.dirname(os.path.realpath(__file__)))
|
||||
other = os.path.basename(_other_copy["folder"])
|
||||
print(f"{red}AusBoss: this pack is installed twice.{reset}")
|
||||
print(f" Using: custom_nodes/{other} (v{_other_copy['version']})")
|
||||
print(f" Skipping: custom_nodes/{own} (v{_read_version()})")
|
||||
print(" Delete one of the two folders, then restart ComfyUI.")
|
||||
|
||||
|
||||
if _other_copy:
|
||||
_print_second_copy()
|
||||
else:
|
||||
try:
|
||||
from .nodes._frontend_cache import register_frontend_cache_guard
|
||||
|
||||
register_frontend_cache_guard()
|
||||
except Exception:
|
||||
traceback.print_exc()
|
||||
except Exception:
|
||||
traceback.print_exc()
|
||||
|
||||
_print_banner()
|
||||
_print_banner()
|
||||
|
||||
@@ -1,9 +1,9 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" width="105" height="20" role="img" aria-label="release: v2.23.0">
|
||||
<title>release: v2.23.0</title>
|
||||
<svg xmlns="http://www.w3.org/2000/svg" width="105" height="20" role="img" aria-label="release: v2.26.3">
|
||||
<title>release: v2.26.3</title>
|
||||
<path fill="#555" d="M0 0h48v20H0z"/>
|
||||
<path fill="#00b4aa" d="M48 0h57v20H48z"/>
|
||||
<g fill="#fff" text-anchor="middle" font-family="Verdana,Geneva,DejaVu Sans,sans-serif" font-size="11">
|
||||
<text x="24" y="14">release</text>
|
||||
<text x="76.5" y="14">v2.23.0</text>
|
||||
<text x="76.5" y="14">v2.26.3</text>
|
||||
</g>
|
||||
</svg>
|
||||
|
||||
|
Before Width: | Height: | Size: 425 B After Width: | Height: | Size: 425 B |
Binary file not shown.
|
After Width: | Height: | Size: 127 KiB |
@@ -0,0 +1,526 @@
|
||||
{
|
||||
"id": "0b455c1f-6b0c-4f69-8a7e-86f8de61df5c",
|
||||
"revision": 0,
|
||||
"last_node_id": 5,
|
||||
"last_link_id": 4,
|
||||
"nodes": [
|
||||
{
|
||||
"id": 1,
|
||||
"type": "AUSBOSS_NODES_WorkflowNote",
|
||||
"pos": [
|
||||
0,
|
||||
30
|
||||
],
|
||||
"size": [
|
||||
440,
|
||||
640
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "note",
|
||||
"type": "STRING",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "note"
|
||||
}
|
||||
}
|
||||
],
|
||||
"outputs": [],
|
||||
"title": "🆎 AusBoss",
|
||||
"properties": {
|
||||
"cnr_id": "ausboss-nodes",
|
||||
"ver": "2.24.0",
|
||||
"Node name for S&R": "AUSBOSS_NODES_WorkflowNote"
|
||||
},
|
||||
"widgets_values": [
|
||||
"{\"v\": 1, \"title\": \"FlashVSR Video Upscale\", \"subtitle\": \"Make a video larger and sharper with one node and one model file.\", \"author\": \"AusBoss\", \"layout\": \"card\", \"accent\": \"\", \"body\": \"[▶ AusBoss on YouTube](https://www.youtube.com/@ausbossdev) · videos on these nodes and workflows\\n\\n---\\n\\n1. Load your video.\\n2. Set **Scale**: how many times larger each side gets. At 1 the video keeps its size and is only sharpened.\\n3. Press **Run**, then play the result in **Save Video**.\\n\\nTo try a short piece first, drag the right end of the bar under the player in Load Video.\\n\\nA long video needs no more graphics memory than a short one. A large result is made in tiles, so it fits the memory you have and takes longer.\\n\\nNeeds AusBoss nodes 2.24.0 or newer. No other pack is needed.\", \"models\": [{\"name\": \"flashvsr_v1.1_fp16.safetensors\", \"dir\": \"diffusion_models\", \"size\": \"3.5 GB\", \"url\": \"https://huggingface.co/ausboss/FlashVSR-v1.1-ComfyUI/resolve/main/flashvsr_v1.1_fp16.safetensors\", \"note\": \"FlashVSR v1.1 by Junhao Zhuang and the OpenImagingLab team, in one file\"}], \"packs\": [{\"name\": \"ComfyUI-AusBoss\", \"url\": \"https://github.com/ausboss/ComfyUI-AusBoss\", \"node\": \"AUSBOSS_NODES_FlashVSRUpscale\"}], \"links\": [{\"label\": \"YouTube · ausbossdev\", \"url\": \"https://www.youtube.com/@ausbossdev\"}, {\"label\": \"GitHub · ausboss\", \"url\": \"https://github.com/ausboss\"}, {\"label\": \"Civitai · AusBoss\", \"url\": \"https://civitai.com/user/AusBoss\"}, {\"label\": \"X · Zanzibased\", \"url\": \"https://x.com/Zanzibased\"}]}",
|
||||
""
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"note": "{\"v\": 1, \"title\": \"FlashVSR Video Upscale\", \"subtitle\": \"Make a video larger and sharper with one node and one model file.\", \"author\": \"AusBoss\", \"layout\": \"card\", \"accent\": \"\", \"body\": \"[▶ AusBoss on YouTube](https://www.youtube.com/@ausbossdev) · videos on these nodes and workflows\\n\\n---\\n\\n1. Load your video.\\n2. Set **Scale**: how many times larger each side gets. At 1 the video keeps its size and is only sharpened.\\n3. Press **Run**, then play the result in **Save Video**.\\n\\nTo try a short piece first, drag the right end of the bar under the player in Load Video.\\n\\nA long video needs no more graphics memory than a short one. A large result is made in tiles, so it fits the memory you have and takes longer.\\n\\nNeeds AusBoss nodes 2.24.0 or newer. No other pack is needed.\", \"models\": [{\"name\": \"flashvsr_v1.1_fp16.safetensors\", \"dir\": \"diffusion_models\", \"size\": \"3.5 GB\", \"url\": \"https://huggingface.co/ausboss/FlashVSR-v1.1-ComfyUI/resolve/main/flashvsr_v1.1_fp16.safetensors\", \"note\": \"FlashVSR v1.1 by Junhao Zhuang and the OpenImagingLab team, in one file\"}], \"packs\": [{\"name\": \"ComfyUI-AusBoss\", \"url\": \"https://github.com/ausboss/ComfyUI-AusBoss\", \"node\": \"AUSBOSS_NODES_FlashVSRUpscale\"}], \"links\": [{\"label\": \"YouTube · ausbossdev\", \"url\": \"https://www.youtube.com/@ausbossdev\"}, {\"label\": \"GitHub · ausboss\", \"url\": \"https://github.com/ausboss\"}, {\"label\": \"Civitai · AusBoss\", \"url\": \"https://civitai.com/user/AusBoss\"}, {\"label\": \"X · Zanzibased\", \"url\": \"https://x.com/Zanzibased\"}]}",
|
||||
"ausboss_workflow_note_panel": ""
|
||||
},
|
||||
"color": "#14424d",
|
||||
"bgcolor": "#161f21"
|
||||
},
|
||||
{
|
||||
"id": 2,
|
||||
"type": "AUSBOSS_NODES_LoadVideo",
|
||||
"pos": [
|
||||
500,
|
||||
80
|
||||
],
|
||||
"size": [
|
||||
420,
|
||||
760
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "video",
|
||||
"type": "COMBO",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "video"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "start_seconds",
|
||||
"type": "FLOAT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "start_seconds"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "end_seconds",
|
||||
"type": "FLOAT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "end_seconds"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "custom_width",
|
||||
"type": "INT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "custom_width"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "custom_height",
|
||||
"type": "INT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "custom_height"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "every_nth",
|
||||
"type": "INT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "every_nth"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "max_frames",
|
||||
"type": "INT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "max_frames"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "single_frame",
|
||||
"type": "BOOLEAN",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "single_frame"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "upload",
|
||||
"type": "IMAGEUPLOAD",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "upload"
|
||||
}
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "frames",
|
||||
"type": "IMAGE",
|
||||
"links": [
|
||||
1
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"links": [
|
||||
3
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "frame_count",
|
||||
"type": "INT",
|
||||
"links": []
|
||||
},
|
||||
{
|
||||
"name": "fps",
|
||||
"type": "FLOAT",
|
||||
"links": [
|
||||
4
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "width",
|
||||
"type": "INT",
|
||||
"links": []
|
||||
},
|
||||
{
|
||||
"name": "height",
|
||||
"type": "INT",
|
||||
"links": []
|
||||
},
|
||||
{
|
||||
"name": "duration",
|
||||
"type": "FLOAT",
|
||||
"links": []
|
||||
},
|
||||
{
|
||||
"name": "video",
|
||||
"type": "VIDEO",
|
||||
"links": []
|
||||
}
|
||||
],
|
||||
"title": "Load Video - Your video",
|
||||
"properties": {
|
||||
"cnr_id": "ausboss-nodes",
|
||||
"ver": "2.24.0",
|
||||
"Node name for S&R": "AUSBOSS_NODES_LoadVideo",
|
||||
"ausboss_load_video_loop": true
|
||||
},
|
||||
"widgets_values": {
|
||||
"video": "",
|
||||
"start_seconds": 0,
|
||||
"end_seconds": 0,
|
||||
"custom_width": 0,
|
||||
"custom_height": 0,
|
||||
"every_nth": 1,
|
||||
"max_frames": 0,
|
||||
"single_frame": false
|
||||
},
|
||||
"widgets_values_named": {
|
||||
"video": "",
|
||||
"start_seconds": 0,
|
||||
"end_seconds": 0,
|
||||
"custom_width": 0,
|
||||
"custom_height": 0,
|
||||
"every_nth": 1,
|
||||
"max_frames": 0,
|
||||
"single_frame": false,
|
||||
"upload": "image",
|
||||
"ausboss_widget_card": "",
|
||||
"ausboss_load_video_viewer": ""
|
||||
},
|
||||
"color": "#14424d",
|
||||
"bgcolor": "#161f21"
|
||||
},
|
||||
{
|
||||
"id": 3,
|
||||
"type": "AUSBOSS_NODES_FlashVSRUpscale",
|
||||
"pos": [
|
||||
980,
|
||||
80
|
||||
],
|
||||
"size": [
|
||||
360,
|
||||
190
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "frames",
|
||||
"type": "IMAGE",
|
||||
"link": 1
|
||||
},
|
||||
{
|
||||
"name": "vae",
|
||||
"type": "VAE",
|
||||
"shape": 7,
|
||||
"link": null
|
||||
},
|
||||
{
|
||||
"name": "model",
|
||||
"type": "COMBO",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "model"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "scale",
|
||||
"type": "FLOAT",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "scale"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "color_match",
|
||||
"type": "BOOLEAN",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "color_match"
|
||||
}
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "frames",
|
||||
"type": "IMAGE",
|
||||
"links": [
|
||||
2
|
||||
]
|
||||
}
|
||||
],
|
||||
"title": "FlashVSR Upscale",
|
||||
"properties": {
|
||||
"cnr_id": "ausboss-nodes",
|
||||
"ver": "2.24.0",
|
||||
"Node name for S&R": "AUSBOSS_NODES_FlashVSRUpscale",
|
||||
"models": [
|
||||
{
|
||||
"name": "flashvsr_v1.1_fp16.safetensors",
|
||||
"url": "https://huggingface.co/ausboss/FlashVSR-v1.1-ComfyUI/resolve/main/flashvsr_v1.1_fp16.safetensors",
|
||||
"directory": "diffusion_models"
|
||||
}
|
||||
]
|
||||
},
|
||||
"widgets_values": [
|
||||
"flashvsr_v1.1_fp16.safetensors",
|
||||
2,
|
||||
true,
|
||||
""
|
||||
],
|
||||
"widgets_values_named": {
|
||||
"model": "flashvsr_v1.1_fp16.safetensors",
|
||||
"scale": 2,
|
||||
"color_match": true,
|
||||
"ausboss_widget_card": ""
|
||||
},
|
||||
"color": "#14424d",
|
||||
"bgcolor": "#161f21"
|
||||
},
|
||||
{
|
||||
"id": 4,
|
||||
"type": "AUSBOSS_NODES_SaveVideo",
|
||||
"pos": [
|
||||
1400,
|
||||
80
|
||||
],
|
||||
"size": [
|
||||
440,
|
||||
640
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [
|
||||
{
|
||||
"name": "frames",
|
||||
"type": "IMAGE",
|
||||
"link": 2
|
||||
},
|
||||
{
|
||||
"name": "audio",
|
||||
"type": "AUDIO",
|
||||
"link": 3
|
||||
},
|
||||
{
|
||||
"name": "video",
|
||||
"type": "VIDEO",
|
||||
"link": null
|
||||
},
|
||||
{
|
||||
"name": "fps",
|
||||
"type": "FLOAT",
|
||||
"link": 4
|
||||
},
|
||||
{
|
||||
"name": "filename_prefix",
|
||||
"type": "STRING",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "filename_prefix"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "crf",
|
||||
"type": "INT",
|
||||
"link": null
|
||||
},
|
||||
{
|
||||
"name": "format",
|
||||
"type": "COMBO",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "format"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "pingpong",
|
||||
"type": "BOOLEAN",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "pingpong"
|
||||
}
|
||||
},
|
||||
{
|
||||
"name": "save_metadata",
|
||||
"type": "BOOLEAN",
|
||||
"link": null,
|
||||
"widget": {
|
||||
"name": "save_metadata"
|
||||
}
|
||||
}
|
||||
],
|
||||
"outputs": [
|
||||
{
|
||||
"name": "file_path",
|
||||
"type": "STRING",
|
||||
"links": []
|
||||
}
|
||||
],
|
||||
"title": "Save Video",
|
||||
"properties": {
|
||||
"cnr_id": "ausboss-nodes",
|
||||
"ver": "2.24.0",
|
||||
"Node name for S&R": "AUSBOSS_NODES_SaveVideo",
|
||||
"ausboss_save_video_loop": true
|
||||
},
|
||||
"widgets_values": {
|
||||
"fps": 24,
|
||||
"filename_prefix": "flashvsr_upscaled",
|
||||
"crf": 19,
|
||||
"format": "mp4 h264",
|
||||
"pingpong": false,
|
||||
"save_metadata": true
|
||||
},
|
||||
"widgets_values_named": {
|
||||
"fps": 24,
|
||||
"filename_prefix": "flashvsr_upscaled",
|
||||
"crf": 19,
|
||||
"format": "mp4 h264",
|
||||
"pingpong": false,
|
||||
"save_metadata": true,
|
||||
"ausboss_widget_card": "",
|
||||
"ausboss_save_video_viewer": ""
|
||||
},
|
||||
"color": "#14424d",
|
||||
"bgcolor": "#161f21"
|
||||
},
|
||||
{
|
||||
"id": 5,
|
||||
"type": "AUSBOSS_NODES_RunTimer",
|
||||
"pos": [
|
||||
1515,
|
||||
770
|
||||
],
|
||||
"size": [
|
||||
210,
|
||||
67
|
||||
],
|
||||
"flags": {},
|
||||
"order": 0,
|
||||
"mode": 0,
|
||||
"inputs": [],
|
||||
"outputs": [],
|
||||
"title": "Run Timer",
|
||||
"properties": {
|
||||
"cnr_id": "ausboss-nodes",
|
||||
"ver": "2.24.0",
|
||||
"Node name for S&R": "AUSBOSS_NODES_RunTimer"
|
||||
},
|
||||
"color": "#14424d",
|
||||
"bgcolor": "#161f21"
|
||||
}
|
||||
],
|
||||
"links": [
|
||||
[
|
||||
1,
|
||||
2,
|
||||
0,
|
||||
3,
|
||||
0,
|
||||
"IMAGE"
|
||||
],
|
||||
[
|
||||
2,
|
||||
3,
|
||||
0,
|
||||
4,
|
||||
0,
|
||||
"IMAGE"
|
||||
],
|
||||
[
|
||||
3,
|
||||
2,
|
||||
1,
|
||||
4,
|
||||
1,
|
||||
"AUDIO"
|
||||
],
|
||||
[
|
||||
4,
|
||||
2,
|
||||
3,
|
||||
4,
|
||||
3,
|
||||
"FLOAT"
|
||||
]
|
||||
],
|
||||
"groups": [
|
||||
{
|
||||
"id": 1,
|
||||
"title": "1 · Your video",
|
||||
"bounding": [
|
||||
480,
|
||||
0,
|
||||
460,
|
||||
880
|
||||
],
|
||||
"color": "#1d5560",
|
||||
"flags": {}
|
||||
},
|
||||
{
|
||||
"id": 2,
|
||||
"title": "2 · Make it larger",
|
||||
"bounding": [
|
||||
960,
|
||||
0,
|
||||
400,
|
||||
880
|
||||
],
|
||||
"color": "#1d5560",
|
||||
"flags": {}
|
||||
},
|
||||
{
|
||||
"id": 3,
|
||||
"title": "3 · Result",
|
||||
"bounding": [
|
||||
1380,
|
||||
0,
|
||||
480,
|
||||
880
|
||||
],
|
||||
"color": "#1d5560",
|
||||
"flags": {}
|
||||
}
|
||||
],
|
||||
"definitions": {
|
||||
"subgraphs": []
|
||||
},
|
||||
"config": {},
|
||||
"extra": {
|
||||
"ds": {
|
||||
"scale": 0.73,
|
||||
"offset": [
|
||||
137,
|
||||
20
|
||||
]
|
||||
}
|
||||
},
|
||||
"version": 0.4
|
||||
}
|
||||
@@ -0,0 +1,45 @@
|
||||
// App mode for nodes with cards: the simple form view shows a real control
|
||||
// for each field instead of an empty row. The logic is in
|
||||
// js/shared/app_mode.mjs; this file watches which mode the open workflow is
|
||||
// in and hands the fields over.
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { createAppModeSync, workflowMode } from "../shared/app_mode.mjs";
|
||||
|
||||
const sync = createAppModeSync({
|
||||
find(nodeId, name) {
|
||||
const graph = app.rootGraph ?? app.graph;
|
||||
const node = graph?.getNodeById?.(/^\d+$/.test(nodeId) ? Number(nodeId) : nodeId);
|
||||
const widget = node?.widgets?.find((entry) => entry.name === name);
|
||||
return widget ? { node, widget } : null;
|
||||
},
|
||||
changed(node) {
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
},
|
||||
});
|
||||
|
||||
let queued = false;
|
||||
// After the current task: a card hides its widgets in a microtask once the
|
||||
// node has loaded, and the form's fields are shown after that.
|
||||
function schedule() {
|
||||
if (queued) return;
|
||||
queued = true;
|
||||
setTimeout(() => {
|
||||
queued = false;
|
||||
try {
|
||||
const workflow = app.extensionManager?.workflow?.activeWorkflow;
|
||||
sync(workflowMode(workflow), (app.rootGraph ?? app.graph)?.extra?.linearData);
|
||||
} catch (error) {
|
||||
console.warn("[AusBoss] App mode fields:", error);
|
||||
}
|
||||
}, 0);
|
||||
}
|
||||
|
||||
app.registerExtension({
|
||||
name: "ausboss.app.mode",
|
||||
setup() {
|
||||
app.extensionManager?.workflow?.$subscribe?.(schedule);
|
||||
},
|
||||
afterConfigureGraph() {
|
||||
schedule();
|
||||
},
|
||||
});
|
||||
+20
-1
@@ -49,6 +49,25 @@ function askForReload(installed) {
|
||||
}
|
||||
}
|
||||
|
||||
// "2.26.2" -> [2, 26, 2]. Anything else, such as "unknown", is null.
|
||||
function parseVersion(version) {
|
||||
const match = /^(\d+)\.(\d+)\.(\d+)$/.exec(typeof version === "string" ? version : "");
|
||||
return match ? match.slice(1).map(Number) : null;
|
||||
}
|
||||
|
||||
// Only a server newer than the page means this tab runs old files. The page
|
||||
// is newer when the Manager switched the pack on disk and the page was
|
||||
// reloaded before ComfyUI restarted, and then a restart is what is needed.
|
||||
// Two version numbers cannot tell that apart from a downgrade with the tab
|
||||
// left open, which needs a reload, so in that case the guard says nothing.
|
||||
function serverIsNewer(installed, running) {
|
||||
const server = parseVersion(installed);
|
||||
const page = parseVersion(running);
|
||||
if (!server || !page) return false;
|
||||
const step = server.findIndex((part, index) => part !== page[index]);
|
||||
return step >= 0 && server[step] > page[step];
|
||||
}
|
||||
|
||||
// The installed version this page already asked a reload for. A server that
|
||||
// drops and comes back several times must not stack the same message.
|
||||
let asked = null;
|
||||
@@ -58,7 +77,7 @@ async function check() {
|
||||
try {
|
||||
const installed = (await (await api.fetchApi("/ausboss/pack_version")).json())?.version;
|
||||
const running = await runningVersion();
|
||||
if (!installed || installed === "unknown" || !running || installed === running || installed === asked) return;
|
||||
if (!serverIsNewer(installed, running) || installed === asked) return;
|
||||
asked = installed;
|
||||
const listing = await (await api.fetchApi("/ausboss/pack_modules")).json();
|
||||
await refreshModules(Array.isArray(listing?.modules) ? listing.modules : []);
|
||||
|
||||
+115
-23
@@ -13,10 +13,12 @@ import { BRAND, chainCallback, keepDomWidgetWidthAuto } from "../shared/index.mj
|
||||
import {
|
||||
ARROW_ANGLE,
|
||||
ARROW_CHOICES,
|
||||
CARD_TEMPLATE,
|
||||
DOUBLE_ARROWS,
|
||||
arrowText,
|
||||
fitFontSize,
|
||||
parseBlocks,
|
||||
parseCard,
|
||||
pastedLink,
|
||||
repointLinks,
|
||||
targetId,
|
||||
@@ -45,6 +47,10 @@ const LINK_TILES = {
|
||||
youtube: { label: "Watch Me on YouTube", url: "https://www.youtube.com/@ausbossdev" },
|
||||
discord: { label: "Join for Help or More Workflows", url: "https://discord.gg/RwVxY9Fuaj" },
|
||||
};
|
||||
// A new Info card opens at this size, which shows its starter text whole.
|
||||
const CARD_SIZE = [300, 230];
|
||||
// The card's text grows with the node like a note's, but a card stays a card.
|
||||
const CARD_FONT_MAX = 20;
|
||||
const FRAME_FILL = 0.8; // the share of the view a node brought into view takes, as in Workflow Switches
|
||||
const FONT = '"IBM Plex Sans", "Inter", "Segoe UI", system-ui, -apple-system, sans-serif';
|
||||
|
||||
@@ -81,6 +87,20 @@ ${TILE_FACE_CSS}
|
||||
.ausboss-callout-card[data-tile] .ausboss-callout-url svg{width:36px;height:28px;flex:none}
|
||||
.ausboss-callout-card[data-tile] .ausboss-callout-out{margin-left:auto}
|
||||
.ausboss-callout-card[data-tile] .ausboss-callout-tools{right:27px;}
|
||||
.ausboss-callout-card[data-card]{background:#111818;border:1px solid #2e3c3b;}
|
||||
.ausboss-callout-card[data-card] .ausboss-callout-inner{padding:.8em 1em;color:#d9e4e3;font-weight:400;line-height:1.45;}
|
||||
.ausboss-callout-card[data-card] b{color:#f2f7f6;font-weight:650;}
|
||||
.ausboss-card-block{margin:0 0 .85em;}
|
||||
.ausboss-card-block:last-child{margin-bottom:0;}
|
||||
.ausboss-card-block>*+*{margin-top:.3em;}
|
||||
.ausboss-card-title{display:flex;align-items:center;gap:.4em;color:#f2f7f6;font-size:1.08em;font-weight:650;}
|
||||
.ausboss-card-icon{flex:none;width:1em;height:1em;color:#c9d6d4;}
|
||||
.ausboss-card-icon svg{display:block;width:100%;height:100%;}
|
||||
.ausboss-card-row{display:flex;justify-content:space-between;align-items:baseline;gap:1em;}
|
||||
.ausboss-card-row>span:last-child{flex:none;text-align:right;white-space:nowrap;}
|
||||
.ausboss-card-row.is-mark,.ausboss-card-row.is-mark b{color:#6fe0d6;font-weight:500;}
|
||||
.ausboss-card-small,.ausboss-card-small b{color:#93a6a4;}
|
||||
.ausboss-card-small{font-size:.92em;}
|
||||
.ausboss-callout-empty{color:#6f8987;font-size:14px;font-weight:400;}
|
||||
.ausboss-callout-arrow{display:inline-block;width:1.2em;height:1.2em;margin:0 .1em;vertical-align:-.26em;filter:drop-shadow(0 0 .3em rgba(0,180,170,.55));}
|
||||
.ausboss-callout-arrow svg{display:block;width:100%;height:100%;overflow:visible;}
|
||||
@@ -154,6 +174,25 @@ function arrowIcon(kind) {
|
||||
return root;
|
||||
}
|
||||
|
||||
// The small line icons an Info card's title can start with.
|
||||
const CARD_ICON_PATHS = {
|
||||
clock: ["M12 7v5l3.2 2"],
|
||||
info: ["M12 11v6", "M12 7.6v.01"],
|
||||
warning: ["M12 3.5 2.8 19.5h18.4Z", "M12 10v4", "M12 16.9v.01"],
|
||||
tip: ["M9.5 18h5", "M10.5 21h3", "M12 3a6 6 0 0 0-3.6 10.8c.7.6 1.1 1.3 1.1 2.2h5c0-.9.4-1.6 1.1-2.2A6 6 0 0 0 12 3Z"],
|
||||
check: ["M8 12.4l2.8 2.8 5.2-5.6"],
|
||||
};
|
||||
const CARD_ICON_RING = new Set(["clock", "info", "check"]);
|
||||
|
||||
function cardIcon(kind) {
|
||||
const span = el("span", "ausboss-card-icon");
|
||||
const root = svg("svg", { viewBox: "0 0 24 24", "aria-hidden": "true", fill: "none", stroke: "currentColor", "stroke-width": "2", "stroke-linecap": "round", "stroke-linejoin": "round" });
|
||||
if (CARD_ICON_RING.has(kind)) root.append(svg("circle", { cx: 12, cy: 12, r: 9 }));
|
||||
for (const d of CARD_ICON_PATHS[kind] ?? []) root.append(svg("path", { d }));
|
||||
span.append(root);
|
||||
return span;
|
||||
}
|
||||
|
||||
// The node a link names, looked up in the graph the note sits in.
|
||||
function linkedNode(state, id) {
|
||||
const graph = state.node.graph ?? app.graph;
|
||||
@@ -262,17 +301,79 @@ function bringIntoView(target) {
|
||||
}
|
||||
}
|
||||
|
||||
// One line's pieces into an element. The pieces of one link sit in one
|
||||
// element; a link that leaves ComfyUI ends with a small ↗.
|
||||
function appendPieces(state, into, line) {
|
||||
let run = null;
|
||||
let wrap = null;
|
||||
const close = () => {
|
||||
if (wrap?.classList.contains("ausboss-callout-url")) wrap.append(el("span", "ausboss-callout-out", "↗"));
|
||||
};
|
||||
for (const piece of line) {
|
||||
if (piece.run !== run) {
|
||||
close();
|
||||
run = piece.run;
|
||||
wrap = run === undefined ? null : linkRun(state, piece);
|
||||
if (wrap) into.append(wrap);
|
||||
}
|
||||
const target = wrap ?? into;
|
||||
if (piece.arrow) target.append(arrowSpan(state, piece.arrow, piece));
|
||||
else target.append(piece.bold ? el("b", "", piece.text) : document.createTextNode(piece.text));
|
||||
}
|
||||
close();
|
||||
}
|
||||
|
||||
// An Info card: a title, rows of label and value, and small print.
|
||||
function renderCard(state, blocks) {
|
||||
for (const lines of blocks) {
|
||||
const block = el("div", "ausboss-card-block");
|
||||
for (const line of lines) {
|
||||
let row;
|
||||
if (line.kind === "title") {
|
||||
row = el("div", "ausboss-card-title");
|
||||
if (line.icon) row.append(cardIcon(line.icon));
|
||||
const words = el("span");
|
||||
appendPieces(state, words, line.pieces);
|
||||
row.append(words);
|
||||
} else if (line.kind === "row") {
|
||||
row = el("div", `ausboss-card-row${line.highlight ? " is-mark" : ""}`);
|
||||
for (const cell of line.cells) {
|
||||
const side = el("span");
|
||||
appendPieces(state, side, cell);
|
||||
row.append(side);
|
||||
}
|
||||
} else {
|
||||
row = el("div", "ausboss-card-text");
|
||||
appendPieces(state, row, line.pieces);
|
||||
}
|
||||
if (line.small) row.classList.add("ausboss-card-small");
|
||||
block.append(row);
|
||||
}
|
||||
state.inner.append(block);
|
||||
}
|
||||
}
|
||||
|
||||
function renderText(state) {
|
||||
const style = state.node.properties?.ausbossCalloutTile;
|
||||
if (["youtube","discord","title"].includes(style)) state.card.dataset.tile = style;
|
||||
else delete state.card.dataset.tile;
|
||||
setTileFace(state.node, state.root, Boolean(state.card.dataset.tile), () => startEdit(state));
|
||||
const blocks = parseBlocks(state.valueWidget.value);
|
||||
// An Info card has no title bar, like a tile, but keeps the full editor
|
||||
// and saves as a plain Callout: an older pack shows its text as a note.
|
||||
if (style === "card") state.card.dataset.card = "";
|
||||
else delete state.card.dataset.card;
|
||||
const isCard = style === "card";
|
||||
setTileFace(state.node, state.root, Boolean(state.card.dataset.tile) || isCard, () => startEdit(state));
|
||||
const blocks = isCard ? parseCard(state.valueWidget.value) : parseBlocks(state.valueWidget.value);
|
||||
state.inner.replaceChildren();
|
||||
state.empty = blocks.length === 0;
|
||||
if (state.empty) {
|
||||
state.inner.append(el("div", "ausboss-callout-empty", "Double-click to write a note"));
|
||||
}
|
||||
if (isCard) {
|
||||
renderCard(state, blocks);
|
||||
fit(state);
|
||||
return;
|
||||
}
|
||||
for (const lines of blocks) {
|
||||
const point = el("div", "ausboss-callout-point");
|
||||
lines.forEach((line, index) => {
|
||||
@@ -280,21 +381,7 @@ function renderText(state) {
|
||||
// lines are told apart by a break, as before; a block needs none.
|
||||
const into = line.small ? el("div", "ausboss-callout-small") : point;
|
||||
if (index && !line.small && !lines[index - 1].small) point.append(document.createElement("br"));
|
||||
let run = null;
|
||||
let wrap = null;
|
||||
for (const piece of line) {
|
||||
// The pieces of one link sit in one element.
|
||||
if (piece.run !== run) {
|
||||
if (wrap?.classList.contains("ausboss-callout-url")) wrap.append(el("span", "ausboss-callout-out", "↗"));
|
||||
run = piece.run;
|
||||
wrap = run === undefined ? null : linkRun(state, piece);
|
||||
if (wrap) into.append(wrap);
|
||||
}
|
||||
const target = wrap ?? into;
|
||||
if (piece.arrow) target.append(arrowSpan(state, piece.arrow, piece));
|
||||
else target.append(piece.bold ? el("b", "", piece.text) : document.createTextNode(piece.text));
|
||||
}
|
||||
if (wrap?.classList.contains("ausboss-callout-url")) wrap.append(el("span", "ausboss-callout-out", "↗"));
|
||||
appendPieces(state, into, line);
|
||||
if (line.small) point.append(into);
|
||||
});
|
||||
state.inner.append(point);
|
||||
@@ -364,7 +451,7 @@ function fit(state) {
|
||||
inner.style.fontSize = `${px}px`;
|
||||
return inner.offsetHeight <= height && inner.scrollWidth <= width + 1;
|
||||
};
|
||||
const size = fitFontSize(fits);
|
||||
const size = fitFontSize(fits, undefined, card.dataset.card === undefined ? undefined : CARD_FONT_MAX);
|
||||
const social = card.dataset.tile === "youtube" || card.dataset.tile === "discord";
|
||||
inner.style.fontSize = `${social ? Math.min(size, 16) : size}px`;
|
||||
}
|
||||
@@ -453,7 +540,9 @@ function startEdit(state) {
|
||||
const wrap = el("div", "ausboss-callout-editor");
|
||||
const area = el("textarea", "ausboss-callout-edit");
|
||||
area.value = state.valueWidget.value ?? "";
|
||||
area.placeholder = "One idea per paragraph. **Bold** a word to stress it. The buttons below add an arrow, make a line small or make words a link.";
|
||||
area.placeholder = state.card.dataset.card === undefined
|
||||
? "One idea per paragraph. **Bold** a word to stress it. The buttons below add an arrow, make a line small or make words a link."
|
||||
: "The first line is the title. Write label | value for a row. **Bold** a whole row to make it teal. Small makes a line small print.";
|
||||
area.spellcheck = true;
|
||||
const bar = el("div", "ausboss-callout-arrowbar");
|
||||
bar.append(el("span", "", "Arrow"));
|
||||
@@ -560,7 +649,6 @@ function startEdit(state) {
|
||||
if (event.target !== area) event.preventDefault();
|
||||
});
|
||||
wrap.addEventListener("pointerdown", (event) => event.stopPropagation());
|
||||
wrap.addEventListener("wheel", (event) => event.stopPropagation(), { passive: true });
|
||||
// Clicking another node on the canvas takes the keyboard from the text box.
|
||||
// That is how the link is picked, so the edit stays open then; a click on
|
||||
// empty canvas, or on the note itself, ends it.
|
||||
@@ -687,7 +775,7 @@ app.registerExtension({
|
||||
const state = buildPanel(this);
|
||||
if (!state) return;
|
||||
renderText(state);
|
||||
if (state.card.dataset.tile && Number.isFinite(saved?.size?.[1])) {
|
||||
if ((state.card.dataset.tile || state.card.dataset.card !== undefined) && Number.isFinite(saved?.size?.[1])) {
|
||||
this.setSize?.([this.size[0], saved.size[1]]);
|
||||
}
|
||||
});
|
||||
@@ -699,14 +787,18 @@ app.registerExtension({
|
||||
chainCallback(nodeType.prototype, "getExtraMenuOptions", function (_canvas, options) {
|
||||
const state = this.__ausbossCallout;
|
||||
if (!state || !Array.isArray(options)) return;
|
||||
for (const [style, label] of [["plain","Plain callout"],["title","Title tile"],["youtube","Set YouTube link tile"],["discord","Set Discord link tile"]]) {
|
||||
for (const [style, label] of [["plain","Plain callout"],["card","Info card"],["title","Title tile"],["youtube","Set YouTube link tile"],["discord","Set Discord link tile"]]) {
|
||||
options.unshift({ content: label, callback: () => {
|
||||
this.properties ??= {};
|
||||
this.properties.ausbossCalloutTile = style;
|
||||
const preset = LINK_TILES[style];
|
||||
if (preset) commitWidgetValue(this, state.valueWidget, `[${preset.label}](${preset.url})`, app.canvas);
|
||||
// An empty card starts with text that shows how one is written.
|
||||
const starter = style === "card" && !String(state.valueWidget.value ?? "").trim();
|
||||
if (starter) commitWidgetValue(this, state.valueWidget, CARD_TEMPLATE, app.canvas);
|
||||
renderText(state);
|
||||
if (style !== "plain") this.setSize?.([this.size[0], style === "title" ? 86 : 78]);
|
||||
if (starter) this.setSize?.([Math.max(this.size[0], CARD_SIZE[0]), CARD_SIZE[1]]);
|
||||
else if (style !== "plain" && style !== "card") this.setSize?.([this.size[0], style === "title" ? 86 : 78]);
|
||||
markChanged();
|
||||
}});
|
||||
}
|
||||
|
||||
@@ -5,7 +5,7 @@
|
||||
// zooming stopped over Compare, the crop previews, the video viewers and every
|
||||
// control on a card. One listener here hands those gestures back, the same way
|
||||
// ComfyUI does for its own text boxes. A panel that wants the wheel for itself
|
||||
// (a scrolling list, a focused text field, an editor) still keeps it.
|
||||
// (a scrolling list, a long text box) keeps it while it can still scroll.
|
||||
//
|
||||
// A file dragged in from the desktop, or a picture's card from the Assets
|
||||
// panel, has the same problem in the classic renderer: dropped on a node's
|
||||
@@ -31,7 +31,7 @@ function onWheel(event) {
|
||||
const root = panelRoot(target);
|
||||
const lgCanvas = canvas();
|
||||
if (!root || !lgCanvas || event.defaultPrevented) return;
|
||||
if (panelKeepsWheel(target, event, { root, host: panelHost(target), styleOf: (el) => getComputedStyle(el), activeElement: document.activeElement })) return;
|
||||
if (panelKeepsWheel(target, event, { root, host: panelHost(target), styleOf: (el) => getComputedStyle(el) })) return;
|
||||
event.preventDefault();
|
||||
lgCanvas.processMouseWheel(event);
|
||||
}
|
||||
|
||||
@@ -0,0 +1,53 @@
|
||||
// Clear VRAM 🆎 — a right-click entry on the empty canvas that unloads every
|
||||
// model and frees memory on the ComfyUI server this page is connected to.
|
||||
//
|
||||
// It sends the same request as ComfyUI's own "Unload Models and Execution
|
||||
// Cache" command (top menu, Edit): POST /free with unload_models and
|
||||
// free_memory. That command is not on the right-click menu, so this entry
|
||||
// puts it one click away. A setting hides the entry; a ComfyUI that turns
|
||||
// manual unloading off (Comfy.Memory.AllowManualUnload) hides it too.
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { showToast } from "../shared/index.mjs";
|
||||
|
||||
const SETTING_ID = "AusBoss.Memory.ClearVramMenu";
|
||||
const MENU_LABEL = "Clear VRAM 🆎";
|
||||
|
||||
function setting(id) {
|
||||
return app.ui?.settings?.getSettingValue?.(id);
|
||||
}
|
||||
|
||||
async function clearVram() {
|
||||
try {
|
||||
const response = await api.fetchApi("/free", {
|
||||
method: "POST",
|
||||
headers: { "Content-Type": "application/json" },
|
||||
body: JSON.stringify({ unload_models: true, free_memory: true }),
|
||||
});
|
||||
if (!response.ok) throw new Error(`the server answered ${response.status}`);
|
||||
showToast({ severity: "success", summary: MENU_LABEL, detail: "Models unloaded and memory freed.", life: 3000 });
|
||||
} catch (error) {
|
||||
showToast({ severity: "error", summary: MENU_LABEL, detail: `Could not clear VRAM: ${error?.message ?? error}` });
|
||||
}
|
||||
}
|
||||
|
||||
app.registerExtension({
|
||||
name: "ausboss.clear.vram",
|
||||
settings: [
|
||||
{
|
||||
id: SETTING_ID,
|
||||
name: "Clear VRAM on the right-click menu",
|
||||
type: "boolean",
|
||||
defaultValue: true,
|
||||
category: ["🆎 AusBoss", "Memory", "Clear VRAM menu"],
|
||||
tooltip:
|
||||
"Right-click the empty canvas for Clear VRAM: it unloads every model and frees "
|
||||
+ "memory on the ComfyUI server. Off hides the entry.",
|
||||
},
|
||||
],
|
||||
getCanvasMenuItems() {
|
||||
if (setting(SETTING_ID) === false) return [];
|
||||
if (setting("Comfy.Memory.AllowManualUnload") === false) return [];
|
||||
return [{ content: MENU_LABEL, callback: clearVram }];
|
||||
},
|
||||
});
|
||||
@@ -15,7 +15,38 @@ Right-click the Callout and choose **Title tile** to give your existing text a b
|
||||
- Double-click a title to edit it. **Save** or `Ctrl+Enter` keeps the change; **Cancel** or `Esc` leaves the original text.
|
||||
- Drag the visible bottom-right grip to resize. You can also focus the grip and use arrow keys; hold **Shift** for smaller changes. Resizing keeps the tile in place and supports Undo.
|
||||
|
||||
## Writing
|
||||
## Info card
|
||||
|
||||
A small card for facts people look up: run times, settings, what each choice
|
||||
does. Right-click the Callout and choose **Info card**. An empty note starts
|
||||
with sample text that shows each part.
|
||||
|
||||
Write it like this:
|
||||
|
||||
```text
|
||||
⏱️ Too slow?
|
||||
###### Redraw time on an RTX 5090. Smaller cards take longer.
|
||||
|
||||
2x, 18 steps | ~4 min
|
||||
2x, 10 steps | ~3 min
|
||||
1.5x, 18 steps | ~2.5 min
|
||||
**1.5x, 10 steps (default) | ~2 min**
|
||||
|
||||
###### 1.5x is a 4608-wide sheet, 2x is 6144.
|
||||
```
|
||||
|
||||
- The first line is the card's title. Start it with ⏱️ ℹ️ ⚠️ 💡 or ✅ and
|
||||
that emoji is drawn as a small icon.
|
||||
- A line with a `|` in it is a row: the words before the `|` on the left,
|
||||
the rest on the right.
|
||||
- Put the whole row in two stars, `**like this | value**`, and it is drawn
|
||||
in teal. Use it for the default or the one to pick.
|
||||
- Small print (the **Small** button) is a grey line, for a subtitle or a note
|
||||
at the bottom.
|
||||
- A blank line leaves a gap. Arrows and links work as in any callout.
|
||||
- Double-click the card to edit it. Choose **Plain callout** to turn it back
|
||||
into a note; the text stays.
|
||||
|
||||
|
||||
- On a plain callout, double-click the note, or press **Edit**, to type. Click outside the box to
|
||||
keep the text. **Esc** drops what you typed.
|
||||
@@ -90,5 +121,13 @@ In the text a link is written `[the words](https://example.com)`, or
|
||||
it to read the address before you click.
|
||||
- An older version of the pack shows a link as typed, brackets and address
|
||||
included, so the address can still be read and copied there.
|
||||
- An Info card is saved as an ordinary Callout with the text you typed.
|
||||
The card look is a setting on the node, so an older version of the pack
|
||||
shows the same text as a plain note, rows written out as `label | value`.
|
||||
A plain callout never reads `|` as a row, so notes written before look
|
||||
the same.
|
||||
- A row splits at its first `|` that is not inside a link. A row pasted
|
||||
from a Markdown table loses its outer pipes, and the table's `---` line
|
||||
is skipped. An Info card's text grows with the card, up to 20 px.
|
||||
- The font is IBM Plex Sans when the computer has it, otherwise the system
|
||||
font. The node loads no fonts and makes no network requests.
|
||||
|
||||
@@ -0,0 +1,142 @@
|
||||
# Edit Lock
|
||||
|
||||
Keeps every part of your picture that an edit did not change exactly where
|
||||
it was, with the colors it had.
|
||||
|
||||
An edit model redraws the whole picture, also the parts you did not ask it
|
||||
to change. Those parts often come back a little zoomed in, shifted or
|
||||
tinted. Flip between the original and the edit and the whole picture
|
||||
jumps. Edit Lock stops that: only the edit is new.
|
||||
|
||||
It needs no mask. It sees what the edit changes while the picture is being
|
||||
made, and holds on to everything else.
|
||||
|
||||
## Quick start
|
||||
|
||||
1. Put **Edit Lock 🆎** between the model and the sampler: the model (after
|
||||
any LoRA loader) goes in, and its `model` output goes to the sampler or
|
||||
guider.
|
||||
2. Wire the VAE into `vae` and the picture you are editing into `image`.
|
||||
Use the same picture the edit model gets, at the size of the result.
|
||||
3. After the VAE Decode, add **Edit Lock Restore 🆎**. Wire the decoded
|
||||
picture into `image` and Edit Lock's `edit_lock` output into
|
||||
`edit_lock`. This puts your own pixels back in the untouched parts.
|
||||
4. Wire Restore's `report` into **Show Text 🆎** to read how much was kept.
|
||||
|
||||
Run your edit as usual. Prompt, steps and seed stay as they were.
|
||||
|
||||
## What it does with each kind of edit
|
||||
|
||||
- **A local edit** (add glasses, repaint the car, change the sign): the
|
||||
edit happens, and the rest of the picture stays put.
|
||||
- **An edit of the whole picture** (a new style, new light, fog):
|
||||
there is nothing to keep, so Edit Lock steps aside and the edit runs as
|
||||
if the node were not there. The report says so.
|
||||
- **A soft effect over the whole picture** (rain, falling snow, haze):
|
||||
it touches everything a little, so there is no untouched part to put
|
||||
back. Edit Lock lets the effect through and only holds the picture in
|
||||
place, so it does not come back zoomed or shifted. The report says so.
|
||||
- **A soft effect and a clear change in one edit** (rain and an umbrella):
|
||||
the node may lock on the umbrella and hold the rain back everywhere
|
||||
else. If the effect is missing outside the changed part, set **Lock** to
|
||||
`off`.
|
||||
- **Something in between** (a new hair color on a close-up, a new
|
||||
background behind a person): it may lock or step aside. The report says
|
||||
which. These edits often change the light on the face too. `auto` steps
|
||||
aside for most of them, so the new light comes with the new hair.
|
||||
`always` keeps the old light on everything it keeps, which can look
|
||||
wrong on a face. A mask settles it: paint the hair and the face, and
|
||||
the rest is locked.
|
||||
- **The model ignores your picture and draws a new one:** that reads as a
|
||||
whole-picture edit, so `auto` steps aside. `always` holds the model to
|
||||
your picture.
|
||||
|
||||
## Controls
|
||||
|
||||
- **model**: the edit model. Patch last, after any LoRAs.
|
||||
- **vae**: the same VAE the workflow decodes with.
|
||||
- **image**: the picture you are editing, at the size of the result. If
|
||||
the sizes differ by more than a few percent the lock does nothing, and
|
||||
the report says so.
|
||||
- **Lock** (`lock`): `auto` (the default) locks the untouched parts and
|
||||
steps aside when the whole picture changes. `always` locks even then.
|
||||
`off` does nothing, which makes a quick before and after.
|
||||
- **Tolerance** (`tolerance`): how big a change must be to count as part of
|
||||
the edit. 1 is the default. Raise it if small unwanted changes get
|
||||
through. Lower it if the lock holds back part of your edit.
|
||||
- **mask** (optional): white marks where the edit may happen; everything
|
||||
else is locked. With a mask the node does not look for the change
|
||||
itself. A mask from **Mask by Touch 🆎** or **Mask by Name 🆎** works.
|
||||
|
||||
## Outputs
|
||||
|
||||
- **model**: the model with the lock on.
|
||||
- **edit_lock**: what the lock kept, for Edit Lock Restore 🆎.
|
||||
|
||||
## Good to know
|
||||
|
||||
- The lock belongs to one picture. For an edit with two pictures, give it
|
||||
the one the result should line up with.
|
||||
- It works on single pictures, not video.
|
||||
- The node watches the first steps of a run and locks for the rest, so
|
||||
the model finishes the edit on top of your real picture. Shadows and
|
||||
reflections the edit adds next to the changed part still come through.
|
||||
- If your edit is missing and the report says nearly all of the picture
|
||||
was kept, set **Lock** to `off`.
|
||||
- Tested on MiniMax-H3-Image. The lock itself is not tied to one model, but
|
||||
the line between "local edit" and "whole picture" was measured there.
|
||||
|
||||
## Technical details
|
||||
|
||||
At every step an edit model predicts the finished picture. Edit Lock
|
||||
compares that prediction with your picture in latent space.
|
||||
|
||||
**Watching.** The first prediction of a run is a blur, and rain, snow or
|
||||
the light a new hair color throws on a face only show a few steps in. So
|
||||
the node watches until the noise is down to 0.8 of where it began: step 4
|
||||
of a 6-step MiniMax-H3-Image run. A run that never gets that low decides
|
||||
at its last step. While it watches, the model draws its own picture and
|
||||
only one thing is corrected. The model often draws the whole picture about
|
||||
one percent larger, or a little shifted. The node measures that zoom and
|
||||
shift at every step, on the calmer half of the picture, and moves the
|
||||
prediction back so it sits where your picture does.
|
||||
|
||||
**Deciding.** Change is measured against the quiet part of the picture:
|
||||
the tenth where the prediction differs least. An even tint over the whole
|
||||
picture is taken out first, so a color cast is not read as a change. A
|
||||
spot counts as changed when it differs by 3 times the quiet level, or by 2
|
||||
times while it touches a spot that does. So a faint change beside a clear
|
||||
one belongs to the edit, and a faint change alone is redraw noise.
|
||||
**Tolerance** multiplies both numbers.
|
||||
|
||||
**Locking.** From the decision on, wherever the picture did not change,
|
||||
the original latent is put back before the sampler takes its next step. A
|
||||
part once read as changed stays changed for the rest of the run. The last
|
||||
step ends on the original there, so what was kept is the original itself,
|
||||
not a redrawn copy.
|
||||
|
||||
**Stepping aside.** `auto` does not lock in three cases:
|
||||
|
||||
- At the first step the whole picture is already changing (a new style,
|
||||
new light). The node reads the calmest fifth of a smoothed copy, where a
|
||||
slightly larger redraw nearly vanishes and a new style does not. On
|
||||
MiniMax-H3-Image, local edits at 1, 4 and 16 megapixels read 0.22 or
|
||||
less there and whole-picture edits 0.27 or more; the line sits at 0.243.
|
||||
The model then runs untouched.
|
||||
- At the decision no place has changed clearly: the strongest spot of the
|
||||
picture is under 4 times the quiet level. That is light rain or fog.
|
||||
Local edits read 4.7 or more there, glasses and makeup included.
|
||||
- At the decision the lock would hand a tenth of the picture or more to
|
||||
the edit, yet the strongest hundredth of the picture is under 4 times
|
||||
the quiet level. That is heavier rain or snow, and a lock would let it
|
||||
through in patches. Local edits that wide read 4.4 or more; rain and
|
||||
snow read 3.6 or less.
|
||||
|
||||
In the last two cases the node keeps undoing the zoom and shift to the end
|
||||
of the run, and puts nothing back. An edit close to one of these lines (a
|
||||
hair color change on a close-up, a new background, smoke) can fall on
|
||||
either side. The report says which, and **Lock** overrides it.
|
||||
|
||||
This is the target the TailSteer paper trains a model toward (the source
|
||||
where nothing changed, the model's own edit elsewhere), applied while
|
||||
sampling. It needs no LoRA and no training.
|
||||
@@ -0,0 +1,73 @@
|
||||
# Edit Lock Restore
|
||||
|
||||
Puts your picture's own pixels back in every part the edit did not change,
|
||||
and takes out the faint pink tint some edit models add to the whole picture.
|
||||
|
||||
**Edit Lock 🆎** keeps the untouched parts in place while the picture is
|
||||
made. They still pass through the VAE, which softens fine detail a little.
|
||||
This node swaps your original pixels in for those parts, so they match
|
||||
your picture pixel for pixel. Only the edit is new.
|
||||
|
||||
## Quick start
|
||||
|
||||
1. Add it right after the VAE Decode, before any upscaler.
|
||||
2. Wire the decoded picture into `image`.
|
||||
3. Wire the `edit_lock` output of Edit Lock 🆎 into `edit_lock`.
|
||||
4. Wire `report` into **Show Text 🆎** to read what was kept.
|
||||
|
||||
## Controls
|
||||
|
||||
- **image**: the edited picture, straight from the VAE Decode.
|
||||
- **edit_lock**: from Edit Lock 🆎. It carries your original picture and
|
||||
the map of what the edit changed, so there is nothing else to wire.
|
||||
- **Feather** (`feather`): how softly your original fades into the edit,
|
||||
in pixels. 16 is the default. The fade lies outside the edit, so the
|
||||
edit itself is never mixed with the old picture.
|
||||
- **Remove tint** (`remove_tint`): on by default. MiniMax H3 Image makes
|
||||
every edit a little pinker and lighter all over. This reads that tint
|
||||
from the part of the picture the edit left alone and takes it out of the
|
||||
whole picture, the edit included. It matters most when Edit Lock stepped
|
||||
aside, such as for a new hair color, a new background or rain, because
|
||||
then nothing else removes it. Edits meant to recolor everything (golden
|
||||
hour, red light, a new art style) are left alone. Turn it off if a color
|
||||
change you asked for comes out weaker.
|
||||
|
||||
## Outputs
|
||||
|
||||
- **image**: the edit, with your original pixels in the untouched parts.
|
||||
- **edit_mask**: white where the edit changed the picture. Use it to see
|
||||
what the lock decided, or for a later step that should only touch the
|
||||
edit.
|
||||
- **report**: one line, such as "Edit Lock kept 82% of your picture as it
|
||||
was and left 18% to the edit". When the tint was taken out it adds a
|
||||
sentence that says so.
|
||||
|
||||
## When nothing was locked
|
||||
|
||||
Your original pixels are not pasted back, the mask is all white, and the
|
||||
report says why: the edit changed the whole picture, the lock was off, or
|
||||
the picture and the result were different sizes. Remove tint still works
|
||||
here, so the picture can come out with the tint taken out.
|
||||
|
||||
## Technical details
|
||||
|
||||
The map of what changed is made while sampling, on the latent grid. Here
|
||||
it is enlarged to the picture's size, widened by the feather and blurred
|
||||
by the same amount, and used to blend the decoded picture with your
|
||||
original. If the decoded picture is not the size of your original (after
|
||||
an upscale), the original is resized to fit first, which costs it some
|
||||
sharpness. That is why this node goes before an upscaler.
|
||||
|
||||
**Remove tint.** The edit is lined up with your original (edits often come
|
||||
back about one percent larger or a little shifted), both are averaged
|
||||
into coarse cells of about 8 px at 1 megapixel, and their difference is
|
||||
taken in Lab color. The tint is the even shift most cells share: it starts
|
||||
at the median and keeps only cells within 4 Lab units of it, five times
|
||||
over, so a new background or new hair drops out of the count. It is taken
|
||||
out only when it looks like the model's own tint: a between +0.5 and +10,
|
||||
b between -8 and 0, L between -1 and +6, read on at least 15% of the
|
||||
picture. On 21 test edits at 16 megapixels it took out the tint on the 5
|
||||
it should (a +2.6 to +6.5 down to +0.4 or less) and left the other 16
|
||||
alone, golden hour, red light and an anime restyle included. It cannot
|
||||
tell this tint from an edit that asks for a slightly pinker, lighter look
|
||||
over the whole picture; turn it off for one of those.
|
||||
@@ -0,0 +1,69 @@
|
||||
# FlashVSR Upscale
|
||||
|
||||
Makes a video larger and sharper. Give it the frames of a clip, set how much
|
||||
larger, and it returns the same clip with more detail. Use it as the last
|
||||
step before Save Video, for a generated clip or one you loaded.
|
||||
|
||||
It needs one model file and no other pack.
|
||||
|
||||
## The model file
|
||||
|
||||
Download `flashvsr_v1.1_fp16.safetensors` (3.5 GB) from
|
||||
[ausboss/FlashVSR-v1.1-ComfyUI](https://huggingface.co/ausboss/FlashVSR-v1.1-ComfyUI)
|
||||
and put it in `ComfyUI/models/diffusion_models`. Then refresh the page, or
|
||||
press `R`, and pick it under **Model**.
|
||||
|
||||
The file holds the official FlashVSR v1.1 weights in one piece. Another
|
||||
FlashVSR file will not load here: the node stops and says which part is
|
||||
missing.
|
||||
|
||||
## Controls
|
||||
|
||||
- **frames**: The clip, as a batch of frames. Load Video's frames output
|
||||
goes here.
|
||||
- **Model** (`model`): The model file above. The list shows the files in
|
||||
`diffusion_models` that have FlashVSR in their name.
|
||||
- **Scale** (`scale`, starts at `2`): How many times larger each side gets,
|
||||
from 1 to 4. At `1` the clip keeps its size and is only sharpened.
|
||||
- **Keep colors** (`color_match`, starts on): Brings each frame's colours
|
||||
back to those of your clip. Leave it on unless you want the model's own
|
||||
colours.
|
||||
- **vae** (optional): Connect the Wan 2.1 VAE for a slower, finer finish.
|
||||
With nothing connected, the fast decoder inside the model file makes the
|
||||
picture.
|
||||
|
||||
## Outputs
|
||||
|
||||
- **frames**: The larger frames, as many as went in, in the same order.
|
||||
The frame rate and the sound are not touched, so wire those from Load
|
||||
Video to Save Video as before.
|
||||
|
||||
## What to expect
|
||||
|
||||
FlashVSR redraws fine detail: skin, hair, cloth and edges come back
|
||||
sharper than the clip had them. It keeps the scene, the motion and the
|
||||
timing.
|
||||
|
||||
A long clip needs no more graphics memory than a short one. The node reads
|
||||
the clip a few frames at a time and keeps only the last few in memory.
|
||||
|
||||
## Technical details
|
||||
|
||||
- **Size.** The result is exactly your size times **Scale**, rounded to
|
||||
even numbers. Inside, the model works at a multiple of 128 pixels each
|
||||
way; the node pads the picture for that and cuts the padding off again,
|
||||
so nothing of your picture is cropped.
|
||||
- **Graphics memory.** About 10 GB for a 576 x 1024 result. A result too
|
||||
large for the free memory is made in overlapping tiles and blended, and
|
||||
the tile size is chosen from the memory that is free when the node runs.
|
||||
1152 x 2048 ran in two tiles at about 16 GB. On an 8 GB card (RTX 3070
|
||||
Laptop) the same result ran in smaller tiles at about 6.5 GB, at about 3
|
||||
seconds a frame.
|
||||
- **Speed.** On an RTX 5090: about 7 frames a second for a 576 x 1024
|
||||
result and about 2 frames a second for 1152 x 2048.
|
||||
- **One step.** The model runs once per group of frames. There are no
|
||||
steps, sampler, prompt or seed to set.
|
||||
- **The model.** FlashVSR v1.1 by Junhao Zhuang and the OpenImagingLab
|
||||
team, released under Apache-2.0. It is built on the Wan 2.1 1.3B video
|
||||
model. This node is the pack's own code in plain PyTorch.
|
||||
- The node makes no network requests and writes no files.
|
||||
@@ -84,31 +84,42 @@ from.
|
||||
|
||||
Connect a video, then pick what it is for:
|
||||
|
||||
- **copy its moves**: your person does the moves in the video. It needs a
|
||||
picture of your person. Say where in your lines:
|
||||
`@1 dances on a rooftop at sunset.` The clothes can come from words or from an
|
||||
outfit picture.
|
||||
- **copy its moves**: someone does the moves in the video. With a picture of
|
||||
your person, your person does them. Say where in your lines:
|
||||
`@1 dances on a rooftop at sunset.` With no person picture, the person in
|
||||
the video keeps their own moves and goes somewhere new: give a picture the
|
||||
job **place** and type `He dances in @1.` The clothes can come from words or
|
||||
from an outfit picture. It doesn't have to be a person: give a cartoon hand,
|
||||
a toy or a mascot the job **thing** and type `@1 waves.` It does the moves in
|
||||
its own look, and nothing of the video's subject is kept.
|
||||
- **swap the person**: your person takes over from the one in the video. The
|
||||
place, the moves and the camera stay. With two person pictures, the first
|
||||
replaces the person on the left and the second replaces the person on the
|
||||
right. To choose who goes, type one line per picture, for example
|
||||
`@1 replaces the man in the hat`. Put spoken words in quotes and name
|
||||
the speaker. If a voice comes from off screen, say so in the line.
|
||||
A line like `@1 dances.` is not read as who goes; only a description such
|
||||
as `the man in the hat` is. With two people, name who speaks.
|
||||
- **change something**: keep the video and describe the change, for example
|
||||
`She wears a red raincoat`. Be specific about what should change.
|
||||
`She wears a red raincoat`. Be specific about what should change. To change
|
||||
the whole look, connect a picture with the job **style** (or **place**, for
|
||||
the same room looking different) and type `It looks like @1.` A clip with
|
||||
nobody in it, such as an empty hallway, stays empty.
|
||||
- **keep going**: more of the same clip, from where it ends. Say what happens
|
||||
next, or leave the box empty and it just carries on. Connect the whole clip:
|
||||
the node sends H3 only the last 2 seconds.
|
||||
- **new voice**: the same clip with a new voice. Connect the voice, and type
|
||||
what they say in quotes.
|
||||
This keeps the person in the clip, so a person picture is not allowed here;
|
||||
use **swap the person** to change who is in it.
|
||||
|
||||
If your swap keeps the video's audio, connect it to **video_sound** and turn on
|
||||
**More → Clip's own sound**. The writer uses the clip to estimate when each
|
||||
line is spoken and keeps only the incoming people's words in the swap prompt.
|
||||
Type a time such as `3.2s:` to set it yourself. This switch starts off.
|
||||
|
||||
With an active video, **swap the person**, **change something** and **new voice**
|
||||
use the incoming clip to set the frame count. **Length** says **From the clip**;
|
||||
With an active video, **copy its moves**, **swap the person**, **change
|
||||
something** and **new voice** use the incoming clip to set the frame count. **Length** says **From the clip**;
|
||||
set the clip's length upstream. The bar does not show a guessed duration or offer
|
||||
retiming controls for these jobs. H3 rounds the frame count to a supported length.
|
||||
Other video jobs, or turning the video off, use the editable **Length** again.
|
||||
@@ -137,6 +148,9 @@ workflows and scripts keep their previous behavior when it is absent.
|
||||
`ends on her sitting by the window`.
|
||||
- **Speech**: put the words in quotes: `@1 says "we are late."` Each person
|
||||
keeps their own voice.
|
||||
- **Talking over a cut**: end the words with `...` and the line keeps going
|
||||
into the next shot, said once. On the last line, `...` means the clip ends
|
||||
mid-sentence.
|
||||
- **Words on a sign**: name the sign: `a neon sign reads "OPEN"`.
|
||||
- **Sound and music**: use the **Sound** and **Music** boxes, or type
|
||||
`sound: rain on a tin roof` and `music: slow sad piano`. With no music you get
|
||||
@@ -195,7 +209,7 @@ as not plugged in.
|
||||
- **Sound** (`sound`): what the clip sounds like, for the whole clip. Same as a
|
||||
`sound:` line.
|
||||
- **Music** (`music`): the music for the whole clip. Same as a `music:` line.
|
||||
- **Length** (`seconds`): how long the clip is, 5 to 15.
|
||||
- **Length** (`seconds`): how long the clip is, 4 to 15.
|
||||
- **Camera** (`camera`): how the camera moves the whole way through.
|
||||
- **Look** (`look`): from the picture keeps your photo's look. The others redraw
|
||||
the clip. Clay and 3D don't work from words, so change your picture first with
|
||||
|
||||
@@ -24,6 +24,15 @@ pan. **Save & close** keeps your edits. **Cancel** or Escape puts everything
|
||||
back, and asks first if anything changed. You do not need the editor: a run
|
||||
uses what the node shows.
|
||||
|
||||
Right-click the picture and choose **Copy image** to copy the original picture,
|
||||
without added padding, rotation, crop, resize, handles or guides. To include
|
||||
your edits, open the gear and turn on **Copy with crop, rotation and padding**.
|
||||
That also includes resize, and is remembered for this browser. **Open image**
|
||||
follows the same choice. This works on the node and in the editor.
|
||||
Direct copying needs localhost or HTTPS. On a plain HTTP address, choose
|
||||
**Open image**, then right-click that
|
||||
picture and use the browser's **Copy image**.
|
||||
|
||||
## On the node
|
||||
|
||||
From top to bottom:
|
||||
|
||||
@@ -5,37 +5,31 @@ clip, a song or any sound a video model or a save node takes, when you want
|
||||
one part of the file and not the whole thing.
|
||||
|
||||
The node shows the whole file as a wave, so you can see where the loud parts
|
||||
are. The part that comes out is bright. Drag the white **IN** and **OUT**
|
||||
handles to set where it starts and stops, drag the bright part to slide it,
|
||||
or click somewhere else on the wave to move it there. You can also type a time
|
||||
under the wave, in seconds (`12.5`) or minutes and seconds (`1:02.5`).
|
||||
are. The part that comes out is bright. Under the wave are three boxes:
|
||||
**Start**, **Duration** and **Ends at**, in seconds. Drag the left handle to
|
||||
move the start, drag the right handle or type a Duration to change how long
|
||||
the sound is, or drag the bright part to slide it. Click somewhere else on
|
||||
the wave to move it there. Typing **Ends at** sets the Duration for you.
|
||||
|
||||
Press **▶** to hear only the part that comes out. **LOOP** plays it again and
|
||||
again.
|
||||
|
||||
## Max length
|
||||
Set the Duration to the length the next step needs, such as the length of
|
||||
the video you are making. An empty Duration runs to the end of the file.
|
||||
|
||||
Set **Max length** to the longest sound the next step can take, such as the
|
||||
length of the video you are making. The start stays where it is, and anything
|
||||
past the limit turns red with a dashed line where it is cut. Next to OUT,
|
||||
the node shows how long the sound that comes out is, such as `15.0 s`. It
|
||||
turns red when part of your pick is cut; hold the mouse over it to see how
|
||||
much, and click it to move OUT to the cut. Play stops at the cut, so what you
|
||||
hear is what comes out. `0` means no limit.
|
||||
|
||||
Max length can be wired. When the wire comes from a number node such as
|
||||
Float 🆎, the wave shows the cut for that number; otherwise the length's tooltip says
|
||||
`max wired` and the run still cuts the sound to the wired length.
|
||||
A box set by another node, such as Float 🆎, shows a **linked** button that
|
||||
jumps to that node, and the wave shows the part that number picks.
|
||||
|
||||
## Controls
|
||||
|
||||
- **Source** (`audio`): a sound file in ComfyUI's input folder. **Upload**
|
||||
adds one. Videos are listed as well, for their sound.
|
||||
- **Max length** (`max_seconds`): the longest the sound may be, in seconds.
|
||||
`0` means no limit.
|
||||
- **IN** (`start_seconds`): where the sound starts.
|
||||
- **OUT** (`end_seconds`): where the sound stops. `0` means the end of the
|
||||
file.
|
||||
- **Start** (`start_seconds`): where the sound starts.
|
||||
- **Duration** (`max_seconds`): how long the sound is, in seconds. Empty, or
|
||||
`0`, runs to the end of the file.
|
||||
- **Ends at**: where the sound stops. Typing it sets the Duration.
|
||||
A workflow saved with an end time (`end_seconds`) keeps it and runs the
|
||||
same.
|
||||
|
||||
## Outputs
|
||||
|
||||
@@ -49,4 +43,4 @@ Float 🆎, the wave shows the cut for that number; otherwise the length's toolt
|
||||
- The sound keeps its own sample rate and channels.
|
||||
- A start past the end of the file stops the run with a message that says so.
|
||||
- The wave is drawn in the browser from the file itself; nothing extra is
|
||||
stored with the workflow besides the four values above.
|
||||
stored with the workflow besides the node's own values.
|
||||
|
||||
@@ -0,0 +1,108 @@
|
||||
# Load Image
|
||||
|
||||
Loads a picture and sets its size in the same node, so a workflow does not
|
||||
need a resize node and a size node after its loader.
|
||||
|
||||
Pick a file in **Source**, press **Load** to upload one, or use the arrows
|
||||
under it to step through the pictures in ComfyUI's input folder. The line
|
||||
beside the arrows shows the size of the file and, when the size changes, the
|
||||
size that comes out:
|
||||
|
||||
`912 × 1632 → 1144 × 2048`
|
||||
|
||||
That is the real output size, worked out before you run.
|
||||
|
||||
Under that line are two small pictures: the file as it goes in, and the
|
||||
picture as it comes out. On the first one the part that is kept is lit and
|
||||
the part that is cut off is dark. Between them the node says the shape
|
||||
(`3:4 → 1:1`) and what happens (`77% size`, `cuts 449 px of height`).
|
||||
|
||||
## Size
|
||||
|
||||
**Size** (`size`) picks how the picture is resized. Each choice shows only
|
||||
the numbers it uses.
|
||||
|
||||
| Choice | What it does | Its number |
|
||||
|---|---|---|
|
||||
| **off** | The picture comes out exactly as the file has it. | none |
|
||||
| **megapixels** | Brings the whole picture to a total size and keeps its shape. | **Megapixels** (`megapixels`): 1.0 is about 1024 × 1024 |
|
||||
| **longest side** | Sets the longer edge; the other edge follows. | **Longest side** (`longest_side`), in pixels |
|
||||
| **scale** | Multiplies both edges. | **Scale** (`scale`): 0.5 is half, 2 is double |
|
||||
| **fit inside** | Makes the picture as big as it can be inside a box, with nothing cut off. | **Width** and **Height** (`width`, `height`) |
|
||||
| **crop to fill** | Covers the box completely and cuts off what sticks out. | **Width** and **Height** |
|
||||
| **crop** | Cuts a piece of that size out of the picture. Nothing is resized. | **Width** and **Height** |
|
||||
| **crop to ratio** | Cuts the largest piece of a shape, such as 16:9, out of the picture. Nothing is resized. | **Ratio** (`ratio`) |
|
||||
|
||||
**Round to** (`multiple`) rounds the new width and height to a multiple of
|
||||
that number. Most models want 8 or 16. Set it to 1 to keep the numbers as
|
||||
they are. It does nothing while Size is off.
|
||||
|
||||
## Which part is kept
|
||||
|
||||
The three crop choices show **Keep** (`crop_from`). It picks the part of the
|
||||
picture that stays: `center` cuts evenly from both sides, `top` keeps the top
|
||||
and cuts from the bottom, `bottom left` keeps that corner, and so on.
|
||||
|
||||
You can also click the first small picture: click its top to keep the top,
|
||||
a corner to keep that corner. The lit part moves so you see what you get.
|
||||
|
||||
In **Ratio**, `width : height` takes the shape from the Width and Height
|
||||
boxes, for a shape that is not in the list (1600 and 900 give 16:9).
|
||||
|
||||
## More
|
||||
|
||||
- **Method** (`method`): how pixels are resampled. `lanczos` is the sharpest
|
||||
for photos. `nearest` keeps pixel art crisp.
|
||||
- **Enlarge** (`enlarge`): on, a small picture is made bigger to reach the
|
||||
size you asked for. Off, a picture is only ever made smaller: one that is
|
||||
already below the target comes out at its own size, and crop to fill cuts
|
||||
the box's shape out of it without growing it.
|
||||
|
||||
## The picture on the node
|
||||
|
||||
The picture under the settings is ComfyUI's own preview, the same one the
|
||||
stock Load Image shows. Drop a file on the node or paste one to upload it,
|
||||
and right-click the node for the mask editor and for copy and paste.
|
||||
|
||||
## Outputs
|
||||
|
||||
- `image`: the picture at its new size.
|
||||
- `mask`: white where the picture is see-through and black where it is
|
||||
solid, as the stock loader gives it. A picture with no transparency gives
|
||||
an all black mask. It is always the same size as `image`.
|
||||
- `width`, `height`: the size that comes out.
|
||||
- `filename`: the file's name without its folder or its ending
|
||||
(`shots/hero.png` gives `hero`), ready for a save node's prefix.
|
||||
- `original_width`, `original_height`: the size of the file before any
|
||||
resize. Use them to bring a result back to the size you started from.
|
||||
|
||||
## Which loader to use
|
||||
|
||||
- **Load Image** (this node): load a picture and set its size.
|
||||
- **Load Image + Pad**: add empty space around a picture for outpainting.
|
||||
- **Image Crop + Rotate + Pad**: turn, crop and pad a picture by hand.
|
||||
|
||||
## Technical details
|
||||
|
||||
- Every frame of an animated file comes through as a batch, all at the same
|
||||
new size. A phone photo is turned the way its EXIF data says.
|
||||
- With Size off the pixels are the file's own, bit for bit, and Round to is
|
||||
not applied.
|
||||
- Rounding moves each edge to the nearest multiple, so the shape can change
|
||||
by up to half a step. Fit inside rounds down instead, so the result never
|
||||
leaves its box.
|
||||
- Crop to fill resizes the picture until it covers the box, then cuts the
|
||||
box out at the side Keep names. The mask goes through the same resize and
|
||||
the same cut.
|
||||
- Crop and crop to ratio only cut. A box larger than the picture is cut down
|
||||
to the picture, and Round to rounds down so the piece stays inside it.
|
||||
Enlarge does not apply to them.
|
||||
- The small pictures are drawn on one scale, so a picture made half as large
|
||||
is drawn half as large.
|
||||
- With Enlarge off, a rounded size is rounded down so it never passes the
|
||||
file's own size.
|
||||
- The size line on the node uses the same arithmetic as the run. The pack's
|
||||
tests hold the two to the same answers over a table of cases.
|
||||
- Files are read only from ComfyUI's input, output and temp folders. A name
|
||||
such as `photo.png [output]` loads a picture ComfyUI made.
|
||||
- The node runs again when the file changes on disk or a setting changes.
|
||||
@@ -1,37 +1,33 @@
|
||||
# Load Video
|
||||
|
||||
Loads a video from ComfyUI's input folder as an `IMAGE` frame batch plus its
|
||||
audio. One responsive player shows the source and previews the selected trim.
|
||||
Drag the high-contrast **IN** and **OUT** handles below it, drag the selected
|
||||
range to move the whole window, or type a time into the two fields, in seconds
|
||||
(`95.5`) or as a timecode (`1:35.5`). Click the strip outside the selection to
|
||||
move the window to that spot.
|
||||
The playhead follows each edit, so the frame at either boundary is visible
|
||||
while trimming. In **Source**, the list plays the clip under the pointer, muted, and filters as you type.
|
||||
audio. The player shows the video; the timing card under it picks the part
|
||||
that comes out.
|
||||
|
||||
Playback is constrained to the selected window. **LOOP** repeats it; turn loop
|
||||
off and playback pauses at OUT. The reload button refreshes the source without
|
||||
creating a second preview. The player takes the height you give the node:
|
||||
drag the node taller for a bigger picture. A tall video is shown whole, never
|
||||
cropped.
|
||||
The timing card has three boxes: **Start**, **Duration** and **Ends at**.
|
||||
Drag the left handle on the bar to move the start, drag the right handle or
|
||||
type a Duration to change how long the clip is, or drag the middle of the
|
||||
clip to slide it. **s | fr** switches every box between seconds and frames;
|
||||
it only changes what you see, never a saved value. **Use as start** starts
|
||||
the clip at the frame in the picture, and **Whole video** puts the clip back
|
||||
to the full file. In **Source**, the list plays the clip under the pointer,
|
||||
muted, and filters as you type.
|
||||
|
||||
**FRAME** switches the loader into a frame picker: the trim strip becomes a
|
||||
single scrub rail, and only the frame at the marker loads — a one-image batch,
|
||||
ready for image workflows without a separate frame-select node. Click or drag
|
||||
anywhere on the rail to choose the frame, or type an exact time into the AT
|
||||
field. Playback runs the whole source freely in this mode so you can hunt for
|
||||
the right moment. Picking a frame moves IN to that time. When you toggle back,
|
||||
OUT stays where it was, unless the frame is at or after OUT: then OUT moves to
|
||||
the end of the video.
|
||||
The player takes the height you give the node: drag the node taller for a
|
||||
bigger picture. A tall video is shown whole, never cropped. **LOOP** plays
|
||||
the clip again and again; with it off, playback pauses at the clip's end.
|
||||
The reload button refreshes the source.
|
||||
|
||||
The label between IN and OUT reports what one Run will actually load — e.g.
|
||||
`0:04.0 of 0:10.0 · 48 frames @ 12 fps`, or `1 frame at 0:05.2` while FRAME
|
||||
is on. The preview cannot re-render `every_nth` frame drops or a `max_frames`
|
||||
cap, so it counts them instead: the frame total and playback rate come from
|
||||
the source's real frame rate (probed from the file, since the browser's
|
||||
player never exposes it) with `every_nth` and `max_frames` applied. When a
|
||||
deciding value arrives over a link, or the rate cannot be probed, the count
|
||||
drops out rather than guessing.
|
||||
**FRAME** turns the loader into a frame picker: only the frame at the marker
|
||||
loads, as a one-image batch, ready for image workflows without a separate
|
||||
frame-select node. Drag the bar to choose the frame, or type it into the
|
||||
**FRAME** box. Playback runs the whole video in this mode so you can hunt for
|
||||
the right moment. Picking a frame moves the clip's start there.
|
||||
|
||||
The boxes count real frames: the frame rate and frame count come from the
|
||||
file itself, so a 4.1 s clip at 60 fps reads 247 frames, the same number
|
||||
the run loads. When a box is set by another node it shows a **linked**
|
||||
button that jumps to that node.
|
||||
|
||||
## Drop a video onto the node
|
||||
|
||||
@@ -48,21 +44,21 @@ embedded workflow replaces the whole graph.
|
||||
|
||||
- **Source** (`video`): A file in the input folder; **Upload** adds new ones.
|
||||
A video ComfyUI made, named like `clip.mp4 [output]`, loads too.
|
||||
- **IN** (`start_seconds`): Skip everything before this time.
|
||||
- **OUT** (`end_seconds`): Stop at this time; `0` means the source end until a
|
||||
handle is moved.
|
||||
- **Every nth** (`every_nth`): Keep one frame in this many — `2` halves the frame count.
|
||||
The `fps` output divides to match, so the clip still plays at real speed
|
||||
downstream.
|
||||
- **Limit** (`max_frames`): Stop after this many kept frames; `0` loads the
|
||||
whole trim window. Caps memory on long clips — the decode ends early
|
||||
instead of loading and discarding.
|
||||
- **LOOP**: Repeat the selected window during preview. It starts on and is
|
||||
saved with the workflow.
|
||||
- **FRAME** (`single_frame`): Load only the frame at the IN time (the field
|
||||
reads **AT** while FRAME is on) as a one-image batch; `end_seconds`,
|
||||
`every_nth`, and `max_frames` are ignored while it is on.
|
||||
- **↻**: Reload the selected source preview.
|
||||
- **Start** (`start_seconds`): Skip everything before this time.
|
||||
- **Duration** (`max_frames`): How many frames the clip keeps. **Whole
|
||||
video** removes the limit, so the clip runs to the end (the widget holds
|
||||
`0`). The decode stops early, so a short clip from a long video stays
|
||||
quick.
|
||||
- **Ends at**: Where the clip stops. Typing it sets the Duration. A
|
||||
workflow saved with an end time (`end_seconds`) keeps it and runs the same.
|
||||
- **Every nth** (`every_nth`): Keep one frame in this many; `2` halves the
|
||||
frame count. The `fps` output divides to match, so the clip still plays at
|
||||
real speed downstream.
|
||||
- **LOOP**: Repeat the clip during preview. It starts on and is saved with
|
||||
the workflow.
|
||||
- **FRAME** (`single_frame`): Load only the frame at the marker as a
|
||||
one-image batch; the Duration and Every nth are ignored while it is on.
|
||||
- **↻**: Reload the source preview.
|
||||
- **Width** / **Height** (`custom_width` / `custom_height`): `0` keeps the
|
||||
source size. Set one side only and the other follows the aspect ratio,
|
||||
rounded to an even number.
|
||||
|
||||
@@ -266,6 +266,14 @@ red with advice to re-pick. Thumbnails, the info card, and range lookups
|
||||
all use the resolved name, so a moved file keeps its previews and Civitai
|
||||
data.
|
||||
|
||||
A missing LoRA is also listed in ComfyUI's own missing-models list (the
|
||||
Issues panel, and the "Download all" button) next to the other models the
|
||||
workflow needs, with a Download button when the workflow saved where the
|
||||
file comes from. Rows that are switched off, at zero strength, or whose
|
||||
file moved folders are left off that list, because the run does not need
|
||||
them or finds them anyway. Rows that come into the node over a link are
|
||||
listed by the node that sets them, not here.
|
||||
|
||||
The run side matches: a stale saved path resolves by name at load time
|
||||
(exact → unique case-insensitive path → unique basename, with a one-line
|
||||
console note when it remaps). A LoRA that is genuinely missing **stops the
|
||||
|
||||
@@ -19,8 +19,8 @@ What is lit in teal is what the mask will be. When it is right, press Run
|
||||
for the whole workflow.
|
||||
|
||||
The first time the pointer rests on a picture, the node looks at it. A
|
||||
small button under the picture shows how far along it is, **❚❚ 40/104**,
|
||||
filling up as it goes. The first things light up after a few seconds, and
|
||||
button under the picture shows how far along it is, **❚❚ Pause · looking
|
||||
40/104**, filling up as it goes. The first things light up after a few seconds, and
|
||||
it has seen the whole picture after about 20.
|
||||
|
||||
You do not have to wait for that. Whatever is lit under the pointer can be
|
||||
@@ -28,7 +28,7 @@ picked: a click takes it at once, also while the node is still looking.
|
||||
|
||||
Click that button to pause the looking when what you need is already lit:
|
||||
what is lit stays, and you go on picking and painting. It turns into
|
||||
**▶**; click it to look on from where it stopped. A click on a spot where
|
||||
**▶ Look on**; click it to look on from where it stopped. A click on a spot where
|
||||
nothing is lit yet looks on as well.
|
||||
|
||||
A click on a spot where nothing is lit yet is not lost: a dot marks it, the
|
||||
@@ -50,6 +50,9 @@ at that spot.
|
||||
A stroke across two things, like both shoes, takes both.
|
||||
- Did the thing not light up at all? Type its name into the field under the
|
||||
picture, like "screen" or "earring", and press Enter.
|
||||
- Is something in the mask that should not be, like the face or the hair of
|
||||
a person you picked? Click the **+** in front of that field so it shows
|
||||
**−**, type "face" and press Enter. Or hold Ctrl and click the face.
|
||||
- When nothing fits, switch to **Paint** and paint the area yourself.
|
||||
|
||||
## The whole subject, or the background
|
||||
@@ -59,6 +62,10 @@ click: the person, the animal, the product. **Background** picks everything
|
||||
else. They are picks like any other: click more things to add them, or
|
||||
Ctrl-click a thing to take it away again.
|
||||
|
||||
Ctrl-click takes away the thing under the pointer, where it is in the mask.
|
||||
On a spot where nothing is picked it changes nothing, and the picture says
|
||||
so.
|
||||
|
||||
Both need ComfyUI's background-removal model, see "The model for Fine
|
||||
edges" below.
|
||||
|
||||
@@ -115,8 +122,11 @@ or an area you painted.
|
||||
it.
|
||||
- **all 3** on a chip takes every thing of that name instead of the one you
|
||||
clicked: all five birds, both shoes.
|
||||
- The field at the end, **+ type a name**, picks every thing of the name
|
||||
you type. With Erase chosen it takes them away.
|
||||
- The field at the end picks every thing of the name you type. The sign in
|
||||
front of it says what happens to the name: **+** adds it, **−** takes it
|
||||
away. Click the sign to change it. A name typed with a minus in front,
|
||||
like "-face", is taken away whatever the sign shows. With Erase chosen the
|
||||
sign starts on **−**.
|
||||
|
||||
## The tools on the node
|
||||
|
||||
@@ -273,6 +283,15 @@ then.
|
||||
|
||||
### How SAM 3 is asked
|
||||
|
||||
- A picture whose long side is over 2048 pixels is asked about on a copy
|
||||
with its long side at 2048. SAM 3 reads every picture at 1008 pixels
|
||||
anyway, and returns each answer at the size it was given, on the graphics
|
||||
card: at 8192 pixels one word with many matches needed over 27 GB. The
|
||||
answer is brought back to the picture's own size, so the mask, the cut-out
|
||||
and your touches stay in pixels of the picture. Paint is never scaled.
|
||||
Measured on a 3840 pixel picture, five words: the new masks share 95 to
|
||||
99.6 % of their area with the old ones, the edge sits a median 0 to 2
|
||||
pixels differently, and a look takes 31 seconds in place of 53.
|
||||
- The search is ComfyUI's own SAM 3: SAM3 Detect, asked about points, or
|
||||
about a word that CLIP Text Encode has read. Both use ComfyUI's own
|
||||
settings (threshold 0.5, two edge passes). The node needs ComfyUI 0.20 or
|
||||
|
||||
@@ -0,0 +1,54 @@
|
||||
# Prompt Variations
|
||||
|
||||
A prompt box that changes on every run. Write choices in curly brackets and
|
||||
the node picks one from each set:
|
||||
|
||||
```text
|
||||
A woman in a {plain|striped|patterned} {tank top|sweater|button-up shirt}
|
||||
and {jeans|a pleated skirt|cargo shorts}, {standing|sitting on a stool|leaning on the wall}.
|
||||
```
|
||||
|
||||
One run might give "A woman in a striped sweater and jeans, sitting on a
|
||||
stool." The next gives another mix.
|
||||
|
||||
Use it to make many different pictures in one go. Wire in the seed, set the
|
||||
Seed node to **randomize**, and set the number of runs next to the Run
|
||||
button to 50. Each run gets a new seed, so each picture gets new choices.
|
||||
|
||||
## Controls
|
||||
|
||||
- The text box (`text`): your prompt. Put each set of choices in curly
|
||||
brackets and split the choices with `|`. Write as many sets as you like,
|
||||
each with as many choices as you like.
|
||||
- `seed` (a socket): the seed your sampler uses. A new seed picks new
|
||||
choices. The same seed picks the same ones again, so a picture you like
|
||||
can be made again from its seed.
|
||||
|
||||
## Output
|
||||
|
||||
- **prompt**: your text with one choice picked from each set. Wire it to
|
||||
Show Text to see what was picked.
|
||||
|
||||
## Tips
|
||||
|
||||
- An empty choice lets a part come and go: `a {|striped} shirt` gives "a
|
||||
shirt" or "a striped shirt".
|
||||
- Sets can go inside sets: `{a {red|blue} hat|a scarf}`.
|
||||
- Text without curly brackets comes out exactly as typed.
|
||||
- To make one choice come up more often, write it more than once:
|
||||
`{jeans|jeans|skirt}`.
|
||||
|
||||
## Technical details
|
||||
|
||||
Only brackets that hold a `|` are choices. Brackets without one are left as
|
||||
they are, so a JSON prompt such as `{"light": "soft"}` passes through
|
||||
unchanged. A `|` inside JSON brackets still counts as a choice.
|
||||
|
||||
The innermost sets are picked first, from left to right, with Python's
|
||||
`random.Random(seed)`. The same text and seed always give the same prompt,
|
||||
on any machine. Spaces around each choice are trimmed, and an empty pick
|
||||
also drops the one space in front of it.
|
||||
|
||||
The picks happen when the prompt runs, on the server, not in the browser.
|
||||
ComfyUI's own `{a|b}` choices in some text boxes are picked by the browser
|
||||
when you queue, with no seed; this node does not use them.
|
||||
@@ -50,8 +50,22 @@ file's name. A caption on the `caption_text` input writes the paired `.txt`.
|
||||
shorter; on again, the node returns to its height with the last saved
|
||||
image. It is display only: the file is saved either way and the queue
|
||||
still lists it. When a run saves more than one image (a batch, or a list
|
||||
such as a folder of images), arrows beside the switch (‹ 2 / 5 ›) step
|
||||
through them.
|
||||
such as a folder of images), they all show side by side. The arrows
|
||||
beside the switch (‹ all 5 ›) go on to each one by itself (‹ 2 / 5 ›) and
|
||||
back round to all of them. A run of more than 48 shows one at a time.
|
||||
The node shows a small copy of each picture (up to 2048 pixels on the
|
||||
long side), so a big file shows at once on a slow connection too. The
|
||||
saved file is always full size. The copies are in ComfyUI's temp folder.
|
||||
- **Auto save** (`autosave`, on by default): on, every run saves its
|
||||
pictures. Off, a run only shows them here and saves nothing. A row over
|
||||
the picture then says **Not saved yet**, with a **Save picture** button.
|
||||
With several pictures it reads **Save all 4**, and with the arrows on one
|
||||
picture, **Save this one**. After a save the row says what is saved.
|
||||
Use it to make a batch, look, and keep only the good ones. The pictures
|
||||
are saved under the name and folder set above, in the format picked,
|
||||
with the workflow inside if **Embed workflow** is on. They wait in
|
||||
ComfyUI's temp folder, which is emptied when ComfyUI restarts, so save
|
||||
what you want before that. The Save button works for the last 256 runs.
|
||||
- **Preview steps** (`step_preview`, off by default): the switch left of
|
||||
**preview** shows the picture while it is being made, one step at a time,
|
||||
where the saved image goes, with a **STOP** button on the bar. Stop a run
|
||||
|
||||
@@ -155,9 +155,25 @@ correction turns the corner without a crease.
|
||||
The shift is subtracted from the pasted region before the blend. In the
|
||||
padding that is the whole shift; inside the feathered band the paste
|
||||
alpha scales it, which is exactly the share of each band pixel the
|
||||
sampler painted. Source pixels stay bit-identical. A crop stitcher, which
|
||||
has no source rectangle, gets a single global shift measured in its
|
||||
blend band instead.
|
||||
sampler painted. Source pixels stay bit-identical.
|
||||
|
||||
A Crop For Inpaint crop has no source rectangle, so it gets one shift for
|
||||
the whole painted area. That shift is read inside the painted area, in the
|
||||
32 pixels next to its edge: the same pixels before and after, taking the
|
||||
middle value. Those 32 pixels are counted at the size the model painted, so
|
||||
a large picture is read the same way and takes no longer. Most inpaint workflows let the model paint a little wider
|
||||
than the paste, so the soft edge itself comes back close to your picture
|
||||
and shows only a fraction of the drift. When the painted area holds
|
||||
something new (less than half of that rim agrees, or the change is larger
|
||||
than 20 LAB units, such as a black jacket over skin), it is not drift: the
|
||||
old pixels in that place show something else. The new part is then
|
||||
compared with your untouched picture right beside it, each side averaged
|
||||
within 48 pixels of the seam. Where the two sides look alike (within 20
|
||||
LAB units, on at least a quarter of the seam, such as new skin next to
|
||||
skin), the gap between them is taken off. Where they do not, such as a
|
||||
jacket next to skin, Tone match leaves that picture as painted. Before,
|
||||
the soft edge was read there instead, and a model that repaints the edge
|
||||
too could tint the whole patch: a yellow swimsuit over skin came back blue.
|
||||
|
||||
On a Crop + Rotate + Pad canvas the source rectangle is the crop, so it
|
||||
can also hold the rotation's empty corners and any transparency. The
|
||||
|
||||
@@ -0,0 +1,60 @@
|
||||
# Text Lines
|
||||
|
||||
Runs what comes after once for each line of your text. Type five prompts,
|
||||
one on each line, press Run once, and you get five pictures.
|
||||
|
||||
Use it for a list of prompts you typed, or for a list another node wrote,
|
||||
such as a text model asked for "five different poses".
|
||||
|
||||
## How to use it
|
||||
|
||||
1. Type your lines in the box, or wire text into **text**.
|
||||
2. Wire **line** to where one prompt would go, such as a text encoder's
|
||||
prompt.
|
||||
3. Run. Every node after it runs once for each line, and a Save Image at the
|
||||
end saves one picture for each.
|
||||
|
||||
Empty lines are skipped. Models and pictures that do not depend on the line
|
||||
are loaded once, not once for each line.
|
||||
|
||||
## Controls
|
||||
|
||||
- The text box (`text`): One line for each run.
|
||||
- **Split at** (`split`): **each line** makes every line its own run.
|
||||
**blank line** keeps lines together and starts the next run at an empty
|
||||
line. Pick it when one prompt is several lines long.
|
||||
- **Numbering** (`remove_numbering`): **remove** takes a number or bullet
|
||||
off the start of each line, such as `1.`, `2)`, `-` or `[prompt 3]`.
|
||||
Lists a text model writes usually have them. **keep** leaves every line
|
||||
as written.
|
||||
- **Limit** (`limit`): The most lines to run. 0 runs them all. Set it when
|
||||
a text model might write more lines than you asked for.
|
||||
|
||||
## Outputs
|
||||
|
||||
- **line**: Each line, one at a time.
|
||||
- **number**: The line's place in the list, starting at 1. Add it to a seed
|
||||
with Math Expression (`a + b`) to give each line its own seed.
|
||||
- **count**: How many lines run.
|
||||
|
||||
## Notes
|
||||
|
||||
- With no lines at all, the run stops at this node and says so.
|
||||
- Every run uses the same seed unless you add **number** to it. Different
|
||||
prompts still give different pictures.
|
||||
- A line such as "2 women on a beach" or "1.5x zoom" keeps its number: only
|
||||
a number followed by a period, bracket or colon counts as numbering.
|
||||
|
||||
## Technical details
|
||||
|
||||
**line** and **number** are ComfyUI lists (`OUTPUT_IS_LIST`), the same way
|
||||
Image Folder hands out pictures. ComfyUI runs each later node once for each
|
||||
item and keeps the results in order. A node that takes a whole list at once,
|
||||
such as core Rebatch Images, gets all of them together.
|
||||
|
||||
Numbering that is removed: a number in square, round or CJK brackets with an
|
||||
optional word before it (`[prompt 1]`, `(2)`), a word and a number followed by
|
||||
a colon, period, bracket or dash (`Shot 3:`), a bare number followed by a
|
||||
period, bracket or colon that has no digit right after it, and the bullets
|
||||
`-`, `*` and `•` followed by a space. Only one marker is removed from each
|
||||
line.
|
||||
@@ -14,28 +14,43 @@ model that paints black regions.
|
||||
- **Inputs**: Pick an existing input video (the list plays the clip under the pointer, muted, and filters as you type), click **Load**, or drop a video file
|
||||
onto the node. The old source widgets remain compatible with saved workflows, but
|
||||
are now driven by this compact card. Choosing another video keeps the canvas
|
||||
settings - Fill, Feather, **Align**, Resize, Every nth, Frames for, Length - and
|
||||
settings - Fill, Feather, **Align**, Resize, Every nth, Frames for, Duration - and
|
||||
fits the new clip to the lit ratio button, the way **Fit** is set. Rotation, crop,
|
||||
padding, IN/OUT and the playhead start over.
|
||||
padding, the clip's timing and the playhead start over.
|
||||
- **Outputs** (`source_mode` / `local_path`): Pick a video from ComfyUI's output folder,
|
||||
such as one you just rendered. The list shows the newest first and previews the video
|
||||
under the pointer. The video is read in place, without an upload copy. In an API prompt
|
||||
it is named the way ComfyUI names a file there: `clip.mp4 [output]`. A whole path such
|
||||
as `/home/me/ComfyUI/output/clip.mp4` is not read, for queued runs and the editor's live
|
||||
preview alike.
|
||||
- **Play / Pause** beside the scrubber previews the selected frames and stops at OUT. Press Play at OUT to start from IN again.
|
||||
- **Timeline** (right under the preview and size line): The rail on the node and in the editor is one timeline. Press or
|
||||
drag anywhere on it to scrub the **playhead**, with Length on or off; the stage shows
|
||||
that frame. Drag the **IN** or **OUT** handle to trim: the playhead rides on the handle, so what you see
|
||||
is the first (IN) or last (OUT) frame the run keeps, and it stays there when you let
|
||||
go. IN and OUT are frames, snapped to the source's frame grid; the boxes under the
|
||||
rail take a frame number, and `start_seconds` / `end_seconds` are derived from them
|
||||
(OUT is exclusive, 0 means the end of the source). Arrow keys on a handle move one
|
||||
second, Shift one frame. **OUT always sits on the last frame the run outputs**, so
|
||||
the handle and the bright bar end together; with Every nth or Frames for it steps to the
|
||||
frames the run keeps, and a dim tail marks frames the window holds but the run
|
||||
drops. In the editor, **Set IN** / **Set OUT** (or the **I** / **O** keys) put a
|
||||
trim point at the playhead, and **Full clip** resets the window. The editor also
|
||||
- **Play / Pause** beside the timeline plays the clip from its start and stops at its end. Pause and Play again to carry on from where you stopped.
|
||||
- **Timeline** (right under the picture): Press or drag anywhere on it to look at
|
||||
another frame; the picture shows it. Point at the bar to see where you are, the
|
||||
clip in words and the source's frame rate. Drag the **start** handle to move the
|
||||
clip (its Duration stays), and the **end** handle to change the Duration. The
|
||||
start stops where the whole Duration still fits, and a short note says so.
|
||||
**Use as start** starts the clip at the frame on the picture.
|
||||
- **Start / Duration / Ends at**: the clip in three boxes. Start and Ends at are
|
||||
places in the source video; Duration is how long the clip is. Type any of them,
|
||||
or drag over a box to scrub it. The small **s | fr** switch shows them in seconds
|
||||
or frames: Start and Ends at count source frames, Duration counts the frames the
|
||||
clip sends on, at the output rate. Point at a box for the other unit and what it
|
||||
covers. Typing a Start that leaves too little video is allowed: an amber note says
|
||||
how many frames still fit and offers to shorten the clip or move the start.
|
||||
- **Linked timing**: when another node feeds `start_frame`, `end_frame`,
|
||||
`frame_load_cap` or `fixed_frames`, that box is locked and its label shows a small
|
||||
**linked** button that jumps to the node. `fixed_frames` is an exact count;
|
||||
`frame_load_cap` is only a limit, so when the video ends first the note says how
|
||||
many frames you get. A count only worked out when the graph runs (a Math
|
||||
Expression, say) shows as **at run time**, and the timeline shows striped room
|
||||
after the start instead of a clip. A rate typed inside a subgraph counts as known.
|
||||
- **Technical details**: `start_seconds` / `end_seconds` are derived from the frames
|
||||
(the end is exclusive, 0 means the end of the source), and the Duration is stored
|
||||
as `max_frames`. Arrow keys on a handle move one second, Shift one frame. With
|
||||
Every nth or Frames for the end steps to the frames the run keeps, and a dim tail
|
||||
marks frames the window holds but the run drops.
|
||||
In the editor, **Set IN** / **Set OUT** (or the **I** / **O** keys) put a
|
||||
trim point at the playhead, and **Whole video** resets the clip. The editor also
|
||||
has buttons for the first and last frame, steps of 1, 25, 50 and 100 frames, and
|
||||
**Play**. Arrow keys step one frame (10 with Shift), Home and End jump to the
|
||||
ends, and Space plays or pauses. **Save & close** keeps your edits; **Cancel** or
|
||||
@@ -55,22 +70,16 @@ model that paints black regions.
|
||||
fill colour with a faint hatch, so black bands show on the black stage. A fresh clip node starts
|
||||
with black and feather 0 and the editor's **Reset transform** returns to them;
|
||||
the image nodes keep their grey fill and feathered mask.
|
||||
- **Length** (`max_frames`): the one other way to set how long the clip is. **Off**
|
||||
(the default): OUT ends the clip. **On**: a number of output frames from IN; OUT
|
||||
sits that many frames after IN and follows it when you drag IN, and dragging OUT or
|
||||
typing the number changes the Length. To move the whole clip, drag IN or the small
|
||||
grip in the middle of the kept part (the grip also slides a trimmed window with
|
||||
Length off). When the Length is the whole clip there is nowhere to move it, and the
|
||||
row says so: "= the whole clip. Shorten it to move IN." The Length stays when you
|
||||
swap the video, so a workflow built for 97 frames keeps taking 97. Turning it on or
|
||||
off never moves anything: the count becomes the OUT, or the OUT the count. If the
|
||||
source ends first, the footer says how many frames are left. **To start** moves IN
|
||||
to the start of the source.
|
||||
- **Whole video / To start** (bottom right of the timing card): **Whole video**
|
||||
uses the whole source, with no Duration limit. When the length is linked, the
|
||||
button reads **To start** and moves the start to the beginning instead. The
|
||||
Duration stays when you swap the video, so a workflow built for 97 frames keeps
|
||||
taking 97.
|
||||
- **Every nth** (`every_nth`): Thin the batch. The `fps` output divides to match
|
||||
every_nth, so the clip keeps real-time downstream; a Length counts the frames kept.
|
||||
every_nth, so the clip keeps real-time downstream; a Duration counts the frames kept.
|
||||
- **Frames for** (`frame_snap`): Keep a frame count a video model takes:
|
||||
**LTX (8n+1)** (49, 97, 121) or **Wan (4n+1)**. **any** keeps every frame in the window.
|
||||
OUT drags and the Length step through those counts.
|
||||
Dragging the end and typing a Duration step through those counts.
|
||||
With it on, `frame_count`, `duration`, the audio window and the stitcher all match the
|
||||
clip the sampler hands back, so wiring `frame_count` into the empty latent's length
|
||||
never leaves the stitch with more source frames than generated ones.
|
||||
@@ -102,7 +111,7 @@ with the padlock on it also takes away the bands the lock added.
|
||||
**Reset** on the node clears rotation, crop and padding and turns the padlock off;
|
||||
fill, feather, **Align** and the timeline stay.
|
||||
**Reset transform** in the editor resets rotation, crop, padding, fill, feather and
|
||||
**Align**; it keeps the source, current frame, trim, Length, resize and stitch
|
||||
**Align**; it keeps the source, current frame, trim, Duration, resize and stitch
|
||||
settings. **Align** makes the width and height divide evenly by the number
|
||||
you pick: padding grows by a few pixels, and a side with no padding is trimmed by
|
||||
a few instead. Some models need sizes divisible by 8, 16 or 32; 1 turns it off.
|
||||
@@ -162,8 +171,8 @@ Five optional sockets sit on the left, above the editor:
|
||||
| `force_rate` | Sample at this fps before Every nth. 0 keeps the source rate. Drops or repeats frames to keep playback speed. Accepts FLOAT or INT. |
|
||||
| `start_frame` | Zero-based source start frame. Overrides timeline IN. |
|
||||
| `end_frame` | Exclusive source end frame. 0 means the end of the source. Overrides timeline OUT. |
|
||||
| `fixed_frames` | Exact output frames from IN; the source must be long enough. Connected, it sets the length: OUT, Length and Frames for step aside. |
|
||||
| `frame_load_cap` | Maximum frames after rate conversion and Every nth, before Frames for. 0 means unlimited. Connected, it sets the length in place of Length. When the count is known, dragging IN stops at the last start that fits the kept frames. |
|
||||
| `fixed_frames` | Exact output frames from IN; the source must be long enough. Connected, it sets the length: Duration, Ends at and Frames for step aside. |
|
||||
| `frame_load_cap` | Maximum frames after rate conversion and Every nth, before Frames for. 0 means unlimited. Connected, it sets the length in place of Duration. When the count is known, dragging the start stops at the last start that fits the kept frames. |
|
||||
| `last_seconds` | Load only this many seconds from the end of the clip, or of the part before OUT. 0 loads all of it. Use it to carry on from a clip: a long clip then costs no more memory than a short one. |
|
||||
|
||||
Frame bounds use the source's frame-rate grid, independent of `force_rate`.
|
||||
@@ -177,17 +186,17 @@ For example, start 24 and end 72 select two seconds from a 24 fps source.
|
||||
A 12 fps forced rate returns 24 frames; a cap of 10 then keeps only the first
|
||||
10 of those. Audio and duration follow the returned batch.
|
||||
|
||||
A connected start or end locks that handle and its IN/OUT field on both
|
||||
A connected start or end locks that handle and its box on both
|
||||
the node and fullscreen editor. Hover to see why. Disconnect to restore
|
||||
local control. While timing inputs are connected, the preview shows the local
|
||||
trim as a reference and avoids claiming an output count before the linked
|
||||
values are evaluated at run time.
|
||||
|
||||
**A connected length input takes over the length.** Connect a count to
|
||||
`fixed_frames` (exact) or `frame_load_cap` (at most) and the OUT handle goes
|
||||
away, OUT and Length grey out, and only IN is left to set: the clip starts at
|
||||
IN and the input decides where it ends. Literal Integer/Float nodes (including
|
||||
`fixed_frames` (exact) or `frame_load_cap` (at most) and the end handle goes
|
||||
away, Duration and Ends at lock, and only Start is left to set: the clip starts
|
||||
there and the input decides where it ends. Literal Integer/Float nodes (including
|
||||
AusBoss cards and reroutes) show the resulting frames on the rail before
|
||||
running; calculated counts show "at run time". An older workflow that saved a
|
||||
Fixed frames number without a connection reads as a Length; the first change
|
||||
Fixed frames number without a connection reads as a Duration; the first change
|
||||
here turns it into one.
|
||||
|
||||
@@ -60,10 +60,12 @@ answered on the same line as the button.
|
||||
- A found file swaps its Download button for an **✓ installed** pill; hovering
|
||||
the pill shows where it was found. A file that lives in a subfolder still
|
||||
counts. When a loader in the workflow still asks for the bare file name,
|
||||
which ComfyUI cannot find there (it marks that loader red), the row says
|
||||
where the file is instead (**in LTXV 2.3/**) and **Use it** points those
|
||||
loaders at the copy it found. The check only asks the ComfyUI server for its
|
||||
own model lists.
|
||||
which ComfyUI cannot find there (it marks that loader red), the note points
|
||||
that loader at your copy by itself and says so in a short message. If you
|
||||
have two copies in different subfolders, the row says where one is
|
||||
(**in LTXV 2.3/**) and **Use it** picks it. To always choose yourself, turn
|
||||
off Settings, 🆎 AusBoss, Models, **Pick found files**. The check only asks
|
||||
the ComfyUI server for its own model lists.
|
||||
- **↻ check files** re-runs the check after you move files in. ComfyUI's own
|
||||
**R** refresh triggers it as well.
|
||||
- Folder `input` is for a sample image or clip shipped with the workflow: it
|
||||
|
||||
@@ -17,6 +17,7 @@ help panel.
|
||||
| Pick a picture size by shape and megapixels, in place of Empty Latent | Latent Size (`AUSBOSS_NODES_Resolution`) |
|
||||
| Resize a picture, or snap its size to a multiple | Image Resize (`AUSBOSS_NODES_ImageResize`), Align Image (`AUSBOSS_NODES_AlignImage`) |
|
||||
| Read a picture's width and height | Image Size (`AUSBOSS_NODES_ImageSize`) |
|
||||
| Load a picture and set its size or crop it (megapixels, longest side, a box, a ratio) in one node | Load Image (`AUSBOSS_NODES_LoadImage`) |
|
||||
| Outpaint: load a picture and add empty space around it | Load Image + Pad (`AUSBOSS_NODES_LoadImagePad`) |
|
||||
| Turn, crop and pad a picture in one node | Image Crop + Rotate + Pad (`AUSBOSS_NODES_ImageCropRotatePad`) |
|
||||
| Inpaint only the masked part, then put it back | Crop For Inpaint (`AUSBOSS_NODES_CropForInpaint`) then Stitch Inpaint (`AUSBOSS_NODES_StitchInpaint`) |
|
||||
@@ -24,6 +25,7 @@ help panel.
|
||||
| Get a mask by typing what to find | Mask by Name (`AUSBOSS_NODES_MaskByName`) |
|
||||
| Get a mask by clicking the thing | Mask by Touch (`AUSBOSS_NODES_MaskByTouch`) |
|
||||
| Grow, shrink, fill or smooth a mask | Mask Refine (`AUSBOSS_NODES_RefineMask`) |
|
||||
| Keep the untouched parts of a picture exactly as they were while an edit model edits it | Edit Lock (`AUSBOSS_NODES_EditLock`) then Edit Lock Restore (`AUSBOSS_NODES_EditLockRestore`) |
|
||||
| Line an edited picture back up with the original | Realign to Source (`AUSBOSS_NODES_RealignToSource`, experimental) |
|
||||
| Match one picture's colors to another | Color Match (`AUSBOSS_NODES_ColorMatch`) |
|
||||
| Upscale big pictures in tiles | Tiled Upscale (`AUSBOSS_NODES_TiledUpscale`) then Tiled Upscale Stitch (`AUSBOSS_NODES_TiledUpscaleStitch`) |
|
||||
@@ -37,8 +39,11 @@ help panel.
|
||||
| Load a sound file and keep only part of it, up to a set length | Load Audio (`AUSBOSS_NODES_LoadAudio`) |
|
||||
| Turn, crop and pad a video: one frame, or the whole clip | Video Crop + Rotate + Pad → Frame (`AUSBOSS_NODES_VideoCropRotatePad`), → Clip (`AUSBOSS_NODES_VideoCropRotatePadClip`) |
|
||||
| Change a clip's frame rate | Frame Interpolate (`AUSBOSS_NODES_FrameInterpolate`) |
|
||||
| Make a clip larger and sharper | FlashVSR Upscale (`AUSBOSS_NODES_FlashVSRUpscale`) |
|
||||
| Work with frame batches | Select Frame, Select Every Nth, Split Batch, Merge Batches (`AUSBOSS_NODES_SelectFrame`, `_SelectEveryNth`, `_SplitBatch`, `_MergeBatches`) |
|
||||
| Run a workflow once for each line of text, such as a list of prompts | Text Lines (`AUSBOSS_NODES_TextLines`) |
|
||||
| One shared seed, number or text | Seed, Integer, Float, Text (`AUSBOSS_NODES_Seed`, `_Integer`, `_Float`, `_Text`) |
|
||||
| A prompt that picks from `{a\|b\|c}` choices on each run, by seed | Prompt Variations (`AUSBOSS_NODES_PromptVariations`) |
|
||||
| Do maths on numbers | Math Expression (`AUSBOSS_NODES_MathExpression`) |
|
||||
| Show text on the node, or the picture while it is made | Show Text (`AUSBOSS_NODES_ShowText`), Step Preview (`AUSBOSS_NODES_StepPreview`) |
|
||||
| Free graphics memory between steps | Free Memory (`AUSBOSS_NODES_FreeMemory`) |
|
||||
|
||||
+306
-20
@@ -14,7 +14,7 @@
|
||||
// switch lays each sampling step over the stage, with a Stop button on the
|
||||
// bar (shared/step_preview.mjs).
|
||||
|
||||
import { panelSwitch } from "../shared/panel_controls.mjs";
|
||||
import { panelSwitch, primaryAction } from "../shared/panel_controls.mjs";
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { isForeignRun } from "../shared/prompt_scope.mjs";
|
||||
@@ -26,14 +26,16 @@ import { suppressCoreImagePreview } from "../shared/core_preview.mjs";
|
||||
import { autoMaskValues } from "../shared/mask_auto.mjs";
|
||||
import { hideInputsInDef, hideWidget, setWidgetVisible } from "../shared/widget_visibility.mjs";
|
||||
import { makePictureStepper } from "../shared/picture_stepper.mjs";
|
||||
import { keepIndex, stepIndex } from "../shared/picture_steps.mjs";
|
||||
import { ALL, allLabel, gridShape, heldPlaces, keepIndex, keepWords, stepIndex, stepWithAll } from "../shared/picture_steps.mjs";
|
||||
import { addStepConsumer, mountStepOverlay, showsStepsOn, stepRun, stopStepRun, watchSteps } from "../shared/step_preview.mjs";
|
||||
import {
|
||||
describeNodePreview,
|
||||
describeSourcePreview,
|
||||
lightPreviewQuery,
|
||||
outputLocatorId,
|
||||
outputRecordCount,
|
||||
outputRecordQuery,
|
||||
ownResultIsCurrent,
|
||||
placeholderText,
|
||||
staleText,
|
||||
sourceFileWidget,
|
||||
@@ -54,7 +56,15 @@ const BAR_GAP = 6;
|
||||
const BAR_PADDING = 2;
|
||||
const PANEL_HEIGHT = STAGE_HEIGHT + BAR_HEIGHT + 4 + 16;
|
||||
const OFF_HEIGHT = BAR_HEIGHT + 16;
|
||||
// The row over pictures that wait for a Save, and the gap under it.
|
||||
const KEEP_HEIGHT = 30;
|
||||
const KEEP_SPACE = KEEP_HEIGHT + 4;
|
||||
const INPUT_SIDE = 1; // LiteGraph.INPUT
|
||||
// A run that left this many pictures or fewer opens with all of them side
|
||||
// by side. More than that and each would be a speck, so the arrows go
|
||||
// through them one at a time, as before.
|
||||
const GRID_MAX = 48;
|
||||
const GRID_GAP = 4;
|
||||
// The node's height from before its preview was switched off, so switching
|
||||
// it back on returns the node to it (saved with the workflow).
|
||||
const HEIGHT_PROPERTY = "ausboss_preview_height";
|
||||
@@ -114,7 +124,17 @@ const NODE_CONFIG = {
|
||||
},
|
||||
AUSBOSS_NODES_SelectFrame: { inputName: "frames", noun: "frames" },
|
||||
// Its result is the file it just saved; the card sits above the panel.
|
||||
AUSBOSS_NODES_SaveImage: { inputName: "images", noun: "an image", switchTitle: "Show the saved image here. The file is saved either way.", steps: true },
|
||||
// SAVE shows only for a run made with Auto save off, which holds its
|
||||
// pictures until this button keeps them.
|
||||
AUSBOSS_NODES_SaveImage: {
|
||||
inputName: "images",
|
||||
noun: "an image",
|
||||
switchTitle: "Show the saved image here. The file is saved either way.",
|
||||
steps: true,
|
||||
// Auto save off: a row over the picture says it is not saved yet and
|
||||
// holds the button that saves it.
|
||||
keep: "Save what is showing: every picture, or the one the arrows picked.",
|
||||
},
|
||||
};
|
||||
|
||||
// The picture sits out of flow, centred in the stage, so it can never set
|
||||
@@ -137,11 +157,22 @@ function ensureCss() {
|
||||
.ausboss-input-preview.preview-off .ausboss-stepper{display:none;}
|
||||
.ausboss-input-preview-stage img,.ausboss-input-preview-stage video{display:none;position:absolute;inset:0;margin:auto;max-width:100%;max-height:100%;object-fit:contain;}
|
||||
.ausboss-input-preview-stage.show-image img{display:block;}
|
||||
.ausboss-input-preview-stage.loading-image img{opacity:.4;transition:opacity .3s ease 1.2s;}
|
||||
.ausboss-input-preview-stage.show-video video{display:block;}
|
||||
.ausboss-input-preview-grid{display:none;position:absolute;inset:0;box-sizing:border-box;gap:${GRID_GAP}px;padding:${GRID_GAP}px;justify-content:center;align-content:center;}
|
||||
.ausboss-input-preview-stage.show-grid .ausboss-input-preview-grid{display:grid;}
|
||||
.ausboss-input-preview-stage .ausboss-input-preview-grid img{display:block;position:static;margin:0;width:100%;height:100%;min-width:0;min-height:0;max-width:none;max-height:none;object-fit:contain;}
|
||||
.ausboss-input-preview-hint{display:none;max-width:86%;color:${T.faint};font:11px/1.35 ${FONT_UI};text-align:center;}
|
||||
.ausboss-input-preview-stage:not(.show-image):not(.show-video) .ausboss-input-preview-hint{display:block;}
|
||||
.ausboss-input-preview-stage:not(.show-image):not(.show-video):not(.show-grid) .ausboss-input-preview-hint{display:block;}
|
||||
.ausboss-input-preview-keep{box-sizing:border-box;flex:none;display:flex;align-items:center;justify-content:space-between;gap:10px;height:${KEEP_HEIGHT}px;padding:0 4px 0 10px;border-radius:${RADIUS_CONTROL}px;background:${T.brandSoft};pointer-events:none;}
|
||||
.ausboss-input-preview-keep[hidden]{display:none;}
|
||||
.ausboss-input-preview-keep span{min-width:0;overflow:hidden;text-overflow:ellipsis;white-space:nowrap;color:${T.brandInk};font:600 11px/1.2 ${FONT_UI};}
|
||||
.ausboss-input-preview-keep.done{background:${T.fill};}
|
||||
.ausboss-input-preview-keep.done span{color:${T.label};}
|
||||
.ausboss-input-preview-tools{display:flex;gap:4px;pointer-events:auto;}
|
||||
.ausboss-input-preview-tool{box-sizing:border-box;height:${BAR_HEIGHT}px;min-width:24px;padding:0 6px;border:1px solid transparent;border-radius:${RADIUS_XS}px;background:${T.fill};color:${T.label};font:600 9.5px/18px ${FONT_UI};letter-spacing:.02em;cursor:pointer;}
|
||||
.ausboss-input-preview-tool[hidden]{display:none;}
|
||||
.ausboss-input-preview-stage .ausboss-input-preview-grid img.kept{outline:2px solid ${T.brand};outline-offset:-2px;}
|
||||
.ausboss-input-preview-tool:hover{color:${T.ink};background:${T.fillHover};}
|
||||
.ausboss-input-preview-tool.active{color:${T.brandInk};background:${T.brandSoft};}
|
||||
.ausboss-input-preview-tool:focus-visible{outline:2px solid ${T.brand};outline-offset:1px;}
|
||||
@@ -451,7 +482,7 @@ function syncPreviewMode(state, resize) {
|
||||
if (resize) {
|
||||
const node = state.node;
|
||||
node.properties ??= {};
|
||||
const floor = node.computeSize?.()[1] ?? (enabled ? PANEL_HEIGHT : OFF_HEIGHT);
|
||||
const floor = node.computeSize?.()[1] ?? panelFloor(state);
|
||||
if (!enabled) {
|
||||
// Off: remember the height, then shrink to what is left.
|
||||
const height = Number(node.size?.[1]);
|
||||
@@ -494,7 +525,7 @@ function buildTools(state, tools, signal) {
|
||||
}
|
||||
|
||||
function showHint(state, text) {
|
||||
state.stage.classList.remove("show-image", "show-video");
|
||||
state.stage.classList.remove("show-image", "show-video", "show-grid");
|
||||
state.hint.textContent = text;
|
||||
state.placeholder = null;
|
||||
releaseSteps(state);
|
||||
@@ -509,14 +540,64 @@ function showPlaceholder(state, connected) {
|
||||
function showMedia(state, kind) {
|
||||
state.stage.classList.toggle("show-image", kind === "image");
|
||||
state.stage.classList.toggle("show-video", kind === "video");
|
||||
state.stage.classList.toggle("show-grid", kind === "grid");
|
||||
state.placeholder = null;
|
||||
releaseSteps(state);
|
||||
}
|
||||
|
||||
function clearMedia(state) {
|
||||
state.img.removeAttribute("src");
|
||||
state.stage.classList.remove("loading-image");
|
||||
state.video.pause?.();
|
||||
state.video.removeAttribute("src");
|
||||
clearGrid(state);
|
||||
}
|
||||
|
||||
function clearGrid(state) {
|
||||
if (!state.gridKey) return;
|
||||
state.gridKey = "";
|
||||
state.grid?.replaceChildren();
|
||||
}
|
||||
|
||||
// Every picture of the run at once, each as big as the stage allows.
|
||||
function layoutGrid(state) {
|
||||
const cells = [...(state.grid?.children ?? [])];
|
||||
if (!cells.length) return;
|
||||
const loaded = cells.find((cell) => cell.naturalWidth > 0 && cell.naturalHeight > 0);
|
||||
const aspect = loaded ? loaded.naturalWidth / loaded.naturalHeight : 1;
|
||||
const { columns, rows } = gridShape(cells.length, state.stage.clientWidth, state.stage.clientHeight, aspect);
|
||||
// Tracks the size of one picture, so the pictures sit together in the
|
||||
// middle of the stage with no bands between the rows.
|
||||
const wide = (state.stage.clientWidth - GRID_GAP * (columns + 1)) / columns;
|
||||
const tall = (state.stage.clientHeight - GRID_GAP * (rows + 1)) / rows;
|
||||
if (!(wide > 0) || !(tall > 0)) return;
|
||||
const width = Math.floor(Math.min(wide, tall * aspect));
|
||||
const height = Math.floor(width / aspect);
|
||||
state.grid.style.gridTemplateColumns = `repeat(${columns}, ${width}px)`;
|
||||
state.grid.style.gridTemplateRows = `repeat(${rows}, ${height}px)`;
|
||||
}
|
||||
|
||||
function showGrid(state, count) {
|
||||
const urls = Array.from({ length: count }, (_, index) => storedResultUrl(state.node, index) ?? "");
|
||||
const key = urls.join("\n");
|
||||
if (state.gridKey !== key) {
|
||||
state.gridKey = key;
|
||||
state.grid.replaceChildren(...urls.map((url, index) => {
|
||||
const cell = document.createElement("img");
|
||||
cell.alt = `Picture ${index + 1} of ${count}`;
|
||||
cell.decoding = "async";
|
||||
// The first picture to load tells the grid the shape to lay out for.
|
||||
cell.addEventListener("load", () => layoutGrid(state), { once: true });
|
||||
cell.addEventListener("error", () => dropLightPreview(state), { once: true });
|
||||
cell.src = url;
|
||||
return cell;
|
||||
}));
|
||||
}
|
||||
state.img.removeAttribute("src");
|
||||
state.video.pause?.();
|
||||
state.video.removeAttribute("src");
|
||||
layoutGrid(state);
|
||||
showMedia(state, "grid");
|
||||
}
|
||||
|
||||
// Each source has one hook, with subscribers that can leave in any order.
|
||||
@@ -560,17 +641,36 @@ function rewatchSource(state, source) {
|
||||
// record: an overwritten file reloads after the next run, and a plain refresh
|
||||
// never reloads the picture.
|
||||
const storedRand = new WeakMap();
|
||||
const outputRecord = (node) => app.nodeOutputs?.[outputLocatorId(node)];
|
||||
// Save Image also writes a small copy of each picture for its face, and the
|
||||
// panel shows that one: a few hundred kilobytes where the file can be tens
|
||||
// of megabytes. The copies are in the temp folder, which ComfyUI empties
|
||||
// when it starts; a record whose copy is gone goes back to the file itself.
|
||||
const lightGone = new WeakSet();
|
||||
function lightQuery(record, index = null) {
|
||||
return record && !lightGone.has(record) ? lightPreviewQuery(record, index) : null;
|
||||
}
|
||||
function storedResultUrl(node, index = null) {
|
||||
const record = app.nodeOutputs?.[outputLocatorId(node)];
|
||||
const query = outputRecordQuery(record, index);
|
||||
const record = outputRecord(node);
|
||||
const light = lightQuery(record, index);
|
||||
const query = light ?? outputRecordQuery(record, index);
|
||||
if (!query) return null;
|
||||
if (!storedRand.has(record)) storedRand.set(record, app.getRandParam?.() ?? "");
|
||||
return api.apiURL(`/view?${query}${app.getPreviewFormatParam?.() ?? ""}${storedRand.get(record)}`);
|
||||
return api.apiURL(`/view?${query}${light ? "" : app.getPreviewFormatParam?.() ?? ""}${storedRand.get(record)}`);
|
||||
}
|
||||
// A small copy that would not load: show the file itself from now on.
|
||||
function dropLightPreview(state) {
|
||||
const record = outputRecord(state.node);
|
||||
if (!lightQuery(record)) return false;
|
||||
lightGone.add(record);
|
||||
scheduleRefresh(state);
|
||||
return true;
|
||||
}
|
||||
|
||||
// Several pictures from one run (a list, or a batch Save Image saved one by
|
||||
// one): the arrows in the bar pick which one shows. One picture keeps the
|
||||
// panel as it was, with no arrows.
|
||||
// one): they all show side by side, and the arrows in the bar go on to each
|
||||
// one by itself and back round to all of them. One picture keeps the panel
|
||||
// as it was, with no arrows.
|
||||
function resultCount(state) {
|
||||
if (state.stale || state.resultGone) return 0;
|
||||
return outputRecordCount(app.nodeOutputs?.[outputLocatorId(state.node)]);
|
||||
@@ -586,10 +686,156 @@ function previewMinWidth(state) {
|
||||
return Math.max(PANEL_MIN_WIDTH, Math.ceil(state.node.size[0] - state.bar.clientWidth + needed));
|
||||
}
|
||||
|
||||
// --- pictures held until Save (Save Image with Auto save off) ---------------
|
||||
|
||||
// The pictures this node is holding, in the order it shows them. Empty when
|
||||
// the last run saved its pictures itself.
|
||||
function heldPictures(state) {
|
||||
if (state.stale || state.resultGone) return [];
|
||||
return heldPlaces(app.nodeOutputs?.[outputLocatorId(state.node)]);
|
||||
}
|
||||
|
||||
// Which of the shown pictures are kept, by their place on the node.
|
||||
function keptOf(state, held) {
|
||||
const key = held.map((picture) => `${picture.token}_${picture.index}`).join(" ");
|
||||
if (state.kept?.key !== key) state.kept = { key, indexes: new Set() };
|
||||
return state.kept.indexes;
|
||||
}
|
||||
|
||||
// A click keeps every picture, or only the one the arrows picked. With the
|
||||
// preview off there is no picture to pick, so it keeps them all.
|
||||
function keepsAll(state, count) {
|
||||
return count < 2 || showsAll(state, count) || !previewEnabled(state);
|
||||
}
|
||||
|
||||
// The row shows while the node holds pictures. It says what is saved, and
|
||||
// its button what a click will do (keepWords).
|
||||
function syncHeld(state, count = resultCount(state)) {
|
||||
const row = state.keepRow;
|
||||
if (!row) return;
|
||||
const held = heldPictures(state);
|
||||
const hidden = !held.length || count < 1;
|
||||
if (row.hidden !== hidden) {
|
||||
row.hidden = hidden;
|
||||
fitKeepRow(state);
|
||||
}
|
||||
if (hidden) return;
|
||||
const kept = keptOf(state, held);
|
||||
const all = keepsAll(state, count);
|
||||
const done = all ? kept.size >= held.length : kept.has(state.index);
|
||||
const words = keepWords({
|
||||
held: held.length, kept: kept.size, all, done, saving: Boolean(state.keeping), name: state.kept?.name ?? "",
|
||||
});
|
||||
state.keepNote.textContent = words.note;
|
||||
state.keepNote.title = state.kept?.path ? `ComfyUI/output/${state.kept.path}` : "";
|
||||
state.keepButton.textContent = words.button;
|
||||
state.keepButton.classList.toggle("quiet", words.quiet);
|
||||
state.keepButton.disabled = Boolean(state.keeping);
|
||||
row.classList.toggle("done", words.quiet && kept.size >= held.length);
|
||||
[...(state.grid?.children ?? [])].forEach((cell, index) => cell.classList.toggle("kept", kept.has(index)));
|
||||
}
|
||||
|
||||
// The row takes room of its own, so the picture keeps its floor: the node
|
||||
// grows to hold it, and a node with the preview off gives the room back.
|
||||
function fitKeepRow(state) {
|
||||
const node = state.node;
|
||||
if (previewEnabled(state)) {
|
||||
ensureNodeMinHeight(node);
|
||||
} else {
|
||||
const floor = Number(node.computeSize?.()?.[1]);
|
||||
if (Number.isFinite(floor)) node.setSize?.([node.size?.[0] ?? PANEL_MIN_WIDTH, floor]);
|
||||
}
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
|
||||
function buildKeepRow(state, title, signal) {
|
||||
const row = document.createElement("div");
|
||||
row.className = "ausboss-input-preview-keep";
|
||||
row.hidden = true;
|
||||
const note = document.createElement("span");
|
||||
const button = primaryAction("Save picture", title);
|
||||
// Stop the graph from seeing the click as a canvas drag on the node.
|
||||
button.addEventListener("pointerdown", (event) => event.stopPropagation(), { signal });
|
||||
button.addEventListener("click", (event) => {
|
||||
event.preventDefault();
|
||||
event.stopPropagation();
|
||||
keepHeld(state);
|
||||
}, { signal });
|
||||
row.append(note, button);
|
||||
state.keepRow = row;
|
||||
state.keepNote = note;
|
||||
state.keepButton = button;
|
||||
return row;
|
||||
}
|
||||
|
||||
async function keepHeld(state) {
|
||||
const held = heldPictures(state);
|
||||
if (!held.length || state.keeping) return;
|
||||
const count = resultCount(state);
|
||||
const wanted = keepsAll(state, count) ? held.map((_, place) => place) : [state.index];
|
||||
// One request for each run, in the order the pictures show.
|
||||
const runs = new Map();
|
||||
for (const place of wanted) {
|
||||
const picture = held[place];
|
||||
if (!picture) continue;
|
||||
if (!runs.has(picture.token)) runs.set(picture.token, []);
|
||||
runs.get(picture.token).push(picture.index);
|
||||
}
|
||||
state.keeping = true;
|
||||
syncHeld(state);
|
||||
const kept = keptOf(state, held);
|
||||
const saved = [];
|
||||
let skipped = 0;
|
||||
try {
|
||||
for (const [token, picks] of runs) {
|
||||
const response = await api.fetchApi("/ausboss/save_image/keep", {
|
||||
method: "POST",
|
||||
headers: { "Content-Type": "application/json" },
|
||||
body: JSON.stringify({ token, picks }),
|
||||
});
|
||||
const answer = await response.json().catch(() => ({}));
|
||||
if (!response.ok) throw new Error(answer.error || "Could not save.");
|
||||
if (!state.alive) return;
|
||||
const done = new Set(answer.kept ?? []);
|
||||
held.forEach((picture, place) => {
|
||||
if (picture.token === token && done.has(picture.index)) kept.add(place);
|
||||
});
|
||||
saved.push(...(answer.saved ?? []));
|
||||
// One picture: the row says the name it was saved under.
|
||||
const path = held.length === 1 ? answer.saved?.[0]?.path : null;
|
||||
if (typeof path === "string" && path && state.kept) {
|
||||
state.kept.path = path;
|
||||
state.kept.name = path.slice(path.lastIndexOf("/") + 1);
|
||||
}
|
||||
skipped += answer.skipped?.length ?? 0;
|
||||
}
|
||||
toast(state, saved.length === 1
|
||||
? `Saved ${saved[0].path}`
|
||||
: saved.length
|
||||
? `Saved ${saved.length} pictures`
|
||||
: skipped ? "Not saved: the file already exists" : "Already saved");
|
||||
} catch (error) {
|
||||
if (state.alive) toast(state, error?.message || "Could not save.");
|
||||
} finally {
|
||||
state.keeping = false;
|
||||
if (state.alive) syncHeld(state);
|
||||
}
|
||||
}
|
||||
|
||||
function showsAll(state, count) {
|
||||
return count > 1 && count <= GRID_MAX && state.all !== false;
|
||||
}
|
||||
|
||||
function stepPicture(state, step) {
|
||||
const count = resultCount(state);
|
||||
if (count < 2) return;
|
||||
state.index = stepIndex(state.index, count, step);
|
||||
if (count > GRID_MAX) {
|
||||
state.index = stepIndex(state.index, count, step);
|
||||
} else {
|
||||
const at = stepWithAll(showsAll(state, count) ? ALL : state.index, count, step);
|
||||
state.all = at === ALL;
|
||||
if (!state.all) state.index = at;
|
||||
}
|
||||
refresh(state);
|
||||
}
|
||||
|
||||
@@ -597,6 +843,7 @@ function refresh(state) {
|
||||
if (!state.alive) return;
|
||||
if (!previewEnabled(state)) {
|
||||
clearMedia(state);
|
||||
syncHeld(state);
|
||||
return;
|
||||
}
|
||||
const source = upstreamNode(state.node, state.inputName);
|
||||
@@ -604,7 +851,9 @@ function refresh(state) {
|
||||
const count = resultCount(state);
|
||||
const hidden = state.stepper.root.hidden;
|
||||
state.index = keepIndex(state.index, count);
|
||||
state.stepper.set(state.index, count);
|
||||
const all = showsAll(state, count);
|
||||
state.stepper.set(state.index, count, all ? allLabel(count) : undefined);
|
||||
syncHeld(state, count);
|
||||
if (state.stepper.root.hidden !== hidden) growToNodeFloor(state.widget);
|
||||
if (state.stale) {
|
||||
// The node ran while its preview was off, so the picture it still has
|
||||
@@ -616,11 +865,20 @@ function refresh(state) {
|
||||
// A node that just stopped the run has no result for this picture: the
|
||||
// one it still holds is from an earlier run (Mask by Name's last find,
|
||||
// tinted), so show what feeds it instead.
|
||||
if (all) {
|
||||
showGrid(state, count);
|
||||
// The grid's cells are new: mark the ones already kept.
|
||||
syncHeld(state, count);
|
||||
return;
|
||||
}
|
||||
clearGrid(state);
|
||||
const described = state.resultGone
|
||||
? describeSourcePreview(source)
|
||||
: count > 1
|
||||
? { kind: "url", url: storedResultUrl(state.node, state.index) }
|
||||
: describeNodePreview(state.node, state.inputName, storedResultUrl(state.node));
|
||||
: describeNodePreview(state.node, state.inputName, storedResultUrl(state.node),
|
||||
// The small copy goes first: node.imgs is the full file.
|
||||
!lightQuery(outputRecord(state.node)) && ownResultIsCurrent(state.node, outputRecord(state.node), state.imgsFor));
|
||||
if (!described) {
|
||||
clearMedia(state);
|
||||
showPlaceholder(state, !!source);
|
||||
@@ -642,6 +900,9 @@ function refresh(state) {
|
||||
state.video.removeAttribute("src");
|
||||
if (state.img.getAttribute("src") !== url) {
|
||||
state.img.src = url;
|
||||
// The picture before stays up while this one loads. A slow load
|
||||
// dims it, so an old picture is never taken for the new one.
|
||||
state.stage.classList.toggle("loading-image", true);
|
||||
} else if (state.img.complete && state.img.naturalWidth > 0) {
|
||||
showMedia(state, "image");
|
||||
}
|
||||
@@ -664,8 +925,19 @@ function scheduleRefresh(state) {
|
||||
// Watching the property rather than an api event catches both routes that set
|
||||
// it - the progress frames streamed during a run and the ui payload delivered
|
||||
// when the node finishes - without this file having to know either event name.
|
||||
// The panel's floor: the bar and the picture, the bar alone with the
|
||||
// preview off, and the Save row on top while it shows.
|
||||
function panelFloor(state, enabled = state ? previewEnabled(state) : true) {
|
||||
const keep = state?.keepRow && !state.keepRow.hidden ? KEEP_SPACE : 0;
|
||||
return (enabled ? PANEL_HEIGHT : OFF_HEIGHT) + keep;
|
||||
}
|
||||
|
||||
function watchOwnResult(node, state) {
|
||||
let images = node.imgs;
|
||||
// Which output the pictures in node.imgs were loaded for. The frontend
|
||||
// gives up on a picture that loads slowly and keeps the run before, so
|
||||
// the panel checks this before it trusts node.imgs (ownResultIsCurrent).
|
||||
state.imgsFor = outputRecord(node);
|
||||
try {
|
||||
Object.defineProperty(node, "imgs", {
|
||||
configurable: true,
|
||||
@@ -673,6 +945,7 @@ function watchOwnResult(node, state) {
|
||||
get: () => images,
|
||||
set: (value) => {
|
||||
images = value;
|
||||
state.imgsFor = outputRecord(node);
|
||||
if (state.alive) scheduleRefresh(state);
|
||||
},
|
||||
});
|
||||
@@ -701,7 +974,9 @@ function buildPanel(node, config) {
|
||||
const toastEl = document.createElement("div");
|
||||
toastEl.className = "ausboss-input-preview-toast";
|
||||
toastEl.style.display = "none";
|
||||
stage.append(img, video, hint, toastEl);
|
||||
const grid = document.createElement("div");
|
||||
grid.className = "ausboss-input-preview-grid";
|
||||
stage.append(img, video, grid, hint, toastEl);
|
||||
const bar = document.createElement("div");
|
||||
bar.className = "ausboss-input-preview-bar";
|
||||
root.append(bar, stage);
|
||||
@@ -711,7 +986,7 @@ function buildPanel(node, config) {
|
||||
const widget = node.addDOMWidget(WIDGET_NAME, "ausboss_input_preview", root, {
|
||||
serialize: false,
|
||||
hideOnZoom: false,
|
||||
getMinHeight: () => (enabled() ? PANEL_HEIGHT : OFF_HEIGHT),
|
||||
getMinHeight: () => panelFloor(node.__ausbossInputPreview, enabled()),
|
||||
});
|
||||
keepDomWidgetWidthAuto(widget);
|
||||
// Not saved with the workflow either: options.serialize only keeps it out
|
||||
@@ -723,7 +998,7 @@ function buildPanel(node, config) {
|
||||
// dragged taller. With the preview off the floor is the bar alone.
|
||||
fillNodeHeight(widget, {
|
||||
minWidth: () => previewMinWidth(node.__ausbossInputPreview),
|
||||
minHeight: () => (enabled() ? PANEL_HEIGHT : OFF_HEIGHT),
|
||||
minHeight: () => panelFloor(node.__ausbossInputPreview, enabled()),
|
||||
minNodeSize: [PANEL_MIN_WIDTH, 90],
|
||||
exactMinWidth: true,
|
||||
});
|
||||
@@ -735,7 +1010,10 @@ function buildPanel(node, config) {
|
||||
noun: config.noun,
|
||||
switchTitle: config.switchTitle,
|
||||
advanced: config.advanced,
|
||||
root, bar, stage, img, video, hint, widget, abort,
|
||||
root, bar, stage, img, video, grid, hint, widget, abort,
|
||||
// Several pictures open side by side; the arrows pick one (false).
|
||||
all: true,
|
||||
gridKey: "",
|
||||
toast: toastEl,
|
||||
toastTimer: 0,
|
||||
toolButtons: null,
|
||||
@@ -751,6 +1029,7 @@ function buildPanel(node, config) {
|
||||
};
|
||||
livePanels.add(state);
|
||||
if (config.tools?.length) bar.append(buildTools(state, config.tools, abort.signal));
|
||||
if (config.keep) bar.after(buildKeepRow(state, config.keep, abort.signal));
|
||||
if (config.readout) {
|
||||
state.readout = document.createElement("span");
|
||||
state.readout.className = "ausboss-input-preview-readout";
|
||||
@@ -769,9 +1048,16 @@ function buildPanel(node, config) {
|
||||
installPreviewSwitch(state, abort.signal);
|
||||
if (config.steps) installStepsSwitch(state, abort.signal);
|
||||
|
||||
img.addEventListener("load", () => showMedia(state, "image"));
|
||||
// The grid is laid out for the stage's shape, so a resized node lays it out again.
|
||||
if (typeof ResizeObserver === "function") {
|
||||
const watcher = new ResizeObserver(() => { if (state.alive && state.gridKey) layoutGrid(state); });
|
||||
watcher.observe(stage);
|
||||
abort.signal.addEventListener("abort", () => watcher.disconnect(), { once: true });
|
||||
}
|
||||
img.addEventListener("load", () => { stage.classList.remove("loading-image"); showMedia(state, "image"); });
|
||||
img.addEventListener("error", () => {
|
||||
if (img.getAttribute("src")) showPlaceholder(state, true);
|
||||
stage.classList.remove("loading-image");
|
||||
if (img.getAttribute("src") && !dropLightPreview(state)) showPlaceholder(state, true);
|
||||
});
|
||||
video.addEventListener("loadeddata", () => showMedia(state, "video"));
|
||||
video.addEventListener("error", () => {
|
||||
|
||||
@@ -20,21 +20,36 @@ export function outputWindow(durationValue, startValue, endValue, maxValue = 0)
|
||||
return { start, end, keepEnd, length: keepEnd - start, cut: Math.max(0, end - keepEnd) };
|
||||
}
|
||||
|
||||
// "0:12.0" style, the way the IN and OUT fields write a time.
|
||||
export function secondsLabel(value) {
|
||||
const seconds = Math.max(0, finite(value));
|
||||
if (seconds < 60) return `${seconds.toFixed(1)} s`;
|
||||
const minutes = Math.floor(seconds / 60);
|
||||
return `${minutes}:${(seconds - minutes * 60).toFixed(1).padStart(4, "0")}`;
|
||||
const hundredth = (value) => Math.round(value * 100) / 100;
|
||||
|
||||
// What the face stores for a pick from `start` to `end`: a Start and a
|
||||
// Duration (max_seconds, with end_seconds at 0, "to the end"). A pick that
|
||||
// reaches the end of the file needs no Duration. When another node sets the
|
||||
// Duration or the end, the pick is stored as a start and an end instead.
|
||||
// A key left out is not written.
|
||||
export function pickWrites(durationValue, startValue, endValue, { maxLinked = false, endLinked = false } = {}) {
|
||||
const duration = Math.max(0, finite(durationValue));
|
||||
const start = hundredth(Math.min(Math.max(0, finite(startValue)), duration));
|
||||
const end = hundredth(Math.max(start, Math.min(finite(endValue), duration)));
|
||||
const toEnd = end >= duration - 0.005;
|
||||
if (endLinked) return { start };
|
||||
if (maxLinked) return { start, end: toEnd ? 0 : end };
|
||||
return { start, end: 0, max: toEnd ? 0 : Math.max(0.01, hundredth(end - start)) };
|
||||
}
|
||||
|
||||
// The line between IN and OUT: what comes out, and what the max cuts off.
|
||||
export function outputSummary(window, { maxFromLink = false } = {}) {
|
||||
if (!window || window.length <= 0) return "";
|
||||
const parts = [secondsLabel(window.length)];
|
||||
if (window.cut > 0.005) parts.push(`cut ${secondsLabel(window.cut)}`);
|
||||
if (maxFromLink) parts.push("max wired");
|
||||
return parts.join(" · ");
|
||||
// A value typed into Start, Duration or Ends at, as the pick it makes.
|
||||
// Start keeps the Duration; Duration and Ends at keep the start. An empty
|
||||
// or 0 Duration, and an empty Ends at, run to the end of the file.
|
||||
export function typedPick(key, value, window, durationValue, minimum = 0.1) {
|
||||
const duration = Math.max(0, finite(durationValue));
|
||||
const clampTime = (time) => Math.min(duration, Math.max(0, finite(time)));
|
||||
if (key === "start") {
|
||||
const start = clampTime(value);
|
||||
return { start, end: Math.min(duration, start + window.length) };
|
||||
}
|
||||
if (value == null || (key === "duration" && !(finite(value) > 0))) return { start: window.start, end: duration };
|
||||
const end = key === "duration" ? window.start + Math.max(minimum, finite(value)) : Math.max(window.start + minimum, finite(value));
|
||||
return { start: window.start, end: clampTime(end) };
|
||||
}
|
||||
|
||||
// Loud and quiet across the file: for each of `buckets` slices, the lowest
|
||||
|
||||
+118
-110
@@ -1,30 +1,35 @@
|
||||
// Load Audio 🆎 - the wave, the trim and the player.
|
||||
// Load Audio 🆎 - the wave, the timing and the player.
|
||||
//
|
||||
// The card above (js/widget_cards) holds Source and Max length. This panel
|
||||
// draws the whole file as a wave: the part that comes out is bright, the
|
||||
// part a Max length cuts off is red, and the rest is dim. Drag a handle to
|
||||
// move IN or OUT, drag the middle to slide the pick, or click elsewhere to
|
||||
// move it there. Play plays only what comes out. The widgets underneath
|
||||
// (start_seconds, end_seconds, max_seconds) stay the source of truth.
|
||||
// The card above (js/widget_cards) holds Source. This panel draws the whole
|
||||
// file as a wave: the part that comes out is bright and the rest is dim.
|
||||
// Under it, Start, Duration and Ends at, as on the video nodes. Drag a
|
||||
// handle on the wave to move the start or the end, drag the middle to slide
|
||||
// the sound, or click elsewhere to move it there. Play plays only what comes
|
||||
// out. The widgets underneath stay the source of truth: the Duration is
|
||||
// max_seconds with end_seconds at 0. A saved end (or a linked Max length)
|
||||
// still works: the part a Max length cuts off shows red.
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { BRAND, chainCallback, keepDomWidgetWidthAuto, notifyAusbossChange, showToast } from "../shared/index.mjs";
|
||||
import { keepKeyInField } from "../shared/canvas_passthrough.mjs";
|
||||
import { WIDGET_FRAME, fillNodeHeight } from "../shared/panel_layout.mjs";
|
||||
import { formatTimecode, parseTimecode } from "../shared/timecode.mjs";
|
||||
import { formatTimecode } from "../shared/timecode.mjs";
|
||||
import { inputNumber } from "../shared/clip_rate.mjs";
|
||||
import { makeScrubInput } from "../shared/scrub_input.mjs";
|
||||
import { linkedNodeInfo } from "../shared/timing_links.mjs";
|
||||
import { parseImageReference } from "../shared/pad_canvas.mjs";
|
||||
import { mediaViewQuery, splitMediaName } from "../shared/video_preview.mjs";
|
||||
import { ensureVideoCss, hideCanvasWidget, makeToolButton } from "../shared/video_ui.mjs";
|
||||
import { MIN_TRIM_SECONDS, dragTrimHandle, slideTrimWindow } from "../load_video/trim_preview.mjs";
|
||||
import { grabZone, outputSummary, outputWindow, secondsLabel, wavePeaks } from "./audio_window.mjs";
|
||||
import { grabZone, outputWindow, pickWrites, typedPick, wavePeaks } from "./audio_window.mjs";
|
||||
|
||||
const NODE_NAME = "AUSBOSS_NODES_LoadAudio";
|
||||
const PANEL_WIDGET = "ausboss_load_audio_wave";
|
||||
// The frontend's own player for an audio_upload input. The wave replaces it.
|
||||
const CORE_AUDIO_WIDGET = "audioUI";
|
||||
const MIN_WIDTH = 280;
|
||||
// 8px root padding + the wave's 28px floor + 6px gap + the 22px play, IN/OUT and loop row.
|
||||
const PANEL_MIN_HEIGHT = 64;
|
||||
// 8px root padding + the wave's 28px floor + 6px gap + the 38px Start,
|
||||
// Duration and Ends at row (a label over a box).
|
||||
const PANEL_MIN_HEIGHT = 80;
|
||||
// Opens small and neat: 300 px wide, with a wave about as tall as a field
|
||||
// (32 px). Drag the corner down for a taller wave.
|
||||
const DEFAULT_WIDTH = 300;
|
||||
@@ -50,26 +55,32 @@ function hideCoreAudio(node) {
|
||||
}
|
||||
}
|
||||
|
||||
// A wired Max length: the number the wire carries, when the far end is a
|
||||
// node that shows it (Float, Integer, a Primitive). null when it can't be known.
|
||||
function wiredMax(node) {
|
||||
const input = node.inputs?.find((slot) => slot?.widget?.name === "max_seconds");
|
||||
if (!input || input.link == null) return { wired: false, value: null };
|
||||
const graph = node.graph;
|
||||
const link = graph?.getLink?.(input.link) ?? graph?.links?.get?.(input.link) ?? graph?.links?.[input.link];
|
||||
const origin = link ? graph?.getNodeById?.(link.origin_id) : null;
|
||||
const shown = origin?.widgets?.find((widget) => typeof widget?.value === "number");
|
||||
return { wired: true, value: shown ? Number(shown.value) : null };
|
||||
function linked(node, name) {
|
||||
return !!node.inputs?.some((slot) => slot?.widget?.name === name && slot.link != null);
|
||||
}
|
||||
|
||||
// A value as the run will see it: the widget, or the number a linked input
|
||||
// carries when the far end is a typed number (through Reroutes and
|
||||
// subgraphs). null when only the run knows it.
|
||||
function valueOf(state, name, widget) {
|
||||
if (!linked(state.node, name)) return Number(widget?.value) || 0;
|
||||
return inputNumber(state.node, name, null);
|
||||
}
|
||||
|
||||
function maxSeconds(state) {
|
||||
const wired = wiredMax(state.node);
|
||||
if (wired.wired) return { value: wired.value ?? 0, fromLink: true, known: wired.value != null };
|
||||
return { value: Number(state.maxWidget?.value) || 0, fromLink: false, known: true };
|
||||
const fromLink = linked(state.node, "max_seconds");
|
||||
const value = valueOf(state, "max_seconds", state.maxWidget);
|
||||
return { value: value ?? 0, fromLink, known: value != null };
|
||||
}
|
||||
|
||||
function currentWindow(state) {
|
||||
return outputWindow(state.duration, state.startWidget.value, state.endWidget.value, maxSeconds(state).value);
|
||||
const start = valueOf(state, "start_seconds", state.startWidget) ?? 0;
|
||||
const max = maxSeconds(state).value;
|
||||
let end = valueOf(state, "end_seconds", state.endWidget) ?? 0;
|
||||
// No end and a Duration: the Duration is where the sound stops, so
|
||||
// nothing past it shows as cut off. Only an end past the Duration does.
|
||||
if (!(end > 0) && max > 0 && state.duration > 0) end = Math.min(state.duration, Math.max(0, start) + max);
|
||||
return outputWindow(state.duration, start, end, max);
|
||||
}
|
||||
|
||||
async function loadPeaks(name, route) {
|
||||
@@ -105,16 +116,15 @@ function ensureAudioCss() {
|
||||
.ausboss-audio-error{pointer-events:auto;border:0;background:transparent;color:#e6bf87;font:11px system-ui;cursor:pointer;white-space:nowrap}
|
||||
.ausboss-audio-root canvas.ausboss-audio-wave{position:absolute;inset:0;width:100%;height:100%;display:block;}
|
||||
.ausboss-audio-root .ausboss-video-stage:not(.is-empty) .ausboss-video-status{display:none;}
|
||||
.ausboss-audio-root .ausboss-audio-row{display:flex;align-items:center;gap:6px;flex:none;padding:0 2px;}
|
||||
.ausboss-audio-root .ausboss-audio-row .ausboss-video-values{flex:1 1 auto;min-width:0;}
|
||||
.ausboss-audio-root .ausboss-audio-play{min-width:30px;font-size:10px;}
|
||||
.ausboss-audio-root .ausboss-audio-length{flex:none;padding:0 2px;color:#7fd9d2;font-size:10px;font-variant-numeric:tabular-nums;white-space:nowrap;cursor:default;}
|
||||
.ausboss-audio-root .ausboss-audio-length:empty{display:none;}
|
||||
.ausboss-audio-root .ausboss-audio-length.has-cut{color:#ff9aa0;cursor:pointer;text-decoration:underline dotted rgba(224,82,90,.8);text-underline-offset:2px;}
|
||||
.ausboss-audio-root .ausboss-audio-row .ausboss-video-tool{flex:none;}
|
||||
.ausboss-audio-root .ausboss-video-values{display:flex;gap:8px;}
|
||||
.ausboss-audio-root .ausboss-video-value input{width:48px;}
|
||||
.ausboss-audio-root .ausboss-audio-length{margin-left:auto;}
|
||||
.ausboss-audio-root .ausboss-audio-wave-row{display:flex;align-items:stretch;gap:6px;flex:1 1 auto;min-height:28px;}
|
||||
.ausboss-audio-root .ausboss-audio-wave-row .ausboss-video-stage{flex:1 1 auto;min-width:0;}
|
||||
.ausboss-audio-root .ausboss-audio-play{flex:none;align-self:center;min-width:30px;font-size:10px;}
|
||||
.ausboss-audio-fields{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:6px;flex:none;padding:0 2px;color:#ddd;font:11px system-ui;}
|
||||
.ausboss-audio-field{display:flex;flex-direction:column;gap:2px;min-width:0;}
|
||||
.ausboss-audio-field .ausboss-scrub{width:100%;box-sizing:border-box;}
|
||||
.ausboss-audio-field-label{display:flex;align-items:center;gap:6px;min-width:0;color:#cadddb;}
|
||||
.ausboss-audio-go{border:0;background:transparent;padding:0;color:#45d6cb;font:600 10px system-ui;text-decoration:underline;cursor:pointer;}
|
||||
.ausboss-audio-go[hidden]{display:none;}
|
||||
`;
|
||||
document.head.appendChild(style);
|
||||
}
|
||||
@@ -192,21 +202,31 @@ function draw(state) {
|
||||
}
|
||||
}
|
||||
|
||||
// One box's look: its value, whether it can be typed, and the "linked"
|
||||
// button when another node sets it.
|
||||
function showField(state, key, value, lockedBy, tip) {
|
||||
const field = state.fields[key];
|
||||
field.control.set(value);
|
||||
field.control.root.style.opacity = lockedBy ? ".45" : "";
|
||||
field.control.root.inert = Boolean(lockedBy);
|
||||
field.go.hidden = !lockedBy;
|
||||
field.cell.title = lockedBy ? `Set by ${lockedBy.title}. Click "linked" to go to that node.` : tip;
|
||||
if (lockedBy) field.go.onclick = () => lockedBy.go?.();
|
||||
}
|
||||
|
||||
function updateFace(state) {
|
||||
const win = currentWindow(state);
|
||||
const max = maxSeconds(state);
|
||||
if (document.activeElement !== state.startInput) state.startInput.value = formatTimecode(win.start);
|
||||
if (document.activeElement !== state.endInput) state.endInput.value = formatTimecode(win.end);
|
||||
if (state.duration > 0) {
|
||||
const summary = outputSummary(win, { maxFromLink: max.fromLink && !max.known });
|
||||
state.summary.textContent = win.length > 0 ? secondsLabel(win.length) : "";
|
||||
state.summary.classList.toggle("has-cut", win.cut > 0.005);
|
||||
state.summary.title = win.cut > 0.005
|
||||
? `${summary}. Max length is ${max.value} s, so ${formatTimecode(win.keepEnd)} to ${formatTimecode(win.end)} is cut off (red). Click to move OUT to ${formatTimecode(win.keepEnd)}.`
|
||||
: `${summary}: ${formatTimecode(win.start)} to ${formatTimecode(win.keepEnd)} comes out.`;
|
||||
} else {
|
||||
state.summary.textContent = "";
|
||||
}
|
||||
const link = (name) => (linked(state.node, name) ? linkedNodeInfo(state.node, name) : null);
|
||||
// A linked value the page cannot read is only known when the graph runs:
|
||||
// its box, and the boxes that depend on it, stay empty.
|
||||
const known = (name) => !linked(state.node, name) || inputNumber(state.node, name, null) != null;
|
||||
const has = state.duration > 0;
|
||||
const startKnown = has && known("start_seconds");
|
||||
const lengthKnown = startKnown && known("max_seconds") && known("end_seconds");
|
||||
showField(state, "start", startKnown ? win.start : null, link("start_seconds"), `Where the sound starts in the file, of ${formatTimecode(state.duration)}.`);
|
||||
showField(state, "duration", lengthKnown ? win.length : null, link("max_seconds"),
|
||||
win.cut > 0.005 ? `How long the sound is. It stops ${formatTimecode(win.cut)} before the saved end (red).` : "How long the sound is. Clear it, or type 0, to run to the end of the file.");
|
||||
showField(state, "end", lengthKnown ? win.keepEnd : null, link("end_seconds"), "Where the sound stops in the file. Typing it changes the Duration.");
|
||||
draw(state);
|
||||
}
|
||||
|
||||
@@ -217,12 +237,19 @@ function writeWidget(widget, value) {
|
||||
widget.callback?.(next);
|
||||
}
|
||||
|
||||
// The pick, stored the way the face shows it (pickWrites): a Start and a
|
||||
// Duration, or a start and an end when another node sets the Duration.
|
||||
function writeTrim(state, bounds) {
|
||||
state.writing = true;
|
||||
try {
|
||||
writeWidget(state.startWidget, bounds.start);
|
||||
// OUT at the very end stays 0, the widget's own "to the end".
|
||||
writeWidget(state.endWidget, bounds.end >= state.duration - 0.005 ? 0 : bounds.end);
|
||||
const startLinked = linked(state.node, "start_seconds");
|
||||
const writes = pickWrites(state.duration, startLinked ? currentWindow(state).start : bounds.start, bounds.end, {
|
||||
maxLinked: linked(state.node, "max_seconds") || !state.maxWidget,
|
||||
endLinked: linked(state.node, "end_seconds"),
|
||||
});
|
||||
if (!startLinked) writeWidget(state.startWidget, writes.start);
|
||||
if (writes.end !== undefined) writeWidget(state.endWidget, writes.end);
|
||||
if (writes.max !== undefined) writeWidget(state.maxWidget, writes.max);
|
||||
} finally {
|
||||
state.writing = false;
|
||||
}
|
||||
@@ -248,7 +275,7 @@ function installDrag(state) {
|
||||
event.preventDefault();
|
||||
event.stopPropagation();
|
||||
const initial = infoAt(event.clientX);
|
||||
const bounds = { start: initial.win.start, end: initial.win.end };
|
||||
const bounds = { start: initial.win.start, end: initial.win.keepEnd };
|
||||
const anchor = initial.fraction * state.duration;
|
||||
const downX = event.clientX;
|
||||
let moved = false;
|
||||
@@ -262,7 +289,7 @@ function installDrag(state) {
|
||||
const next = infoAt(clientX);
|
||||
if (initial.zone === "start" || initial.zone === "end") {
|
||||
const now = currentWindow(state);
|
||||
writeTrim(state, dragTrimHandle(state.duration, { start: now.start, end: now.end }, initial.zone, next.fraction, MIN_TRIM_SECONDS));
|
||||
writeTrim(state, dragTrimHandle(state.duration, { start: now.start, end: now.keepEnd }, initial.zone, next.fraction, MIN_TRIM_SECONDS));
|
||||
} else if (initial.zone === "move") {
|
||||
writeTrim(state, slideTrimWindow(state.duration, bounds, next.fraction * state.duration - anchor));
|
||||
} else {
|
||||
@@ -336,19 +363,6 @@ function installPlayer(state) {
|
||||
state.stopPlayer = () => { audio.pause(); stop(); };
|
||||
}
|
||||
|
||||
function commitTimeInput(state, edge, input) {
|
||||
if (state.duration <= 0) return;
|
||||
const seconds = parseTimecode(input.value);
|
||||
const win = currentWindow(state);
|
||||
if (seconds === null) {
|
||||
input.value = formatTimecode(edge === "start" ? win.start : win.end);
|
||||
return;
|
||||
}
|
||||
writeTrim(state, dragTrimHandle(state.duration, { start: win.start, end: win.end }, edge, seconds / state.duration, MIN_TRIM_SECONDS));
|
||||
notifyAusbossChange();
|
||||
input.blur();
|
||||
}
|
||||
|
||||
function watchWidget(state, widget) {
|
||||
if (!widget || widget.__ausbossAudioWatch) return;
|
||||
widget.__ausbossAudioWatch = true;
|
||||
@@ -385,37 +399,44 @@ function buildPanel(node) {
|
||||
playButton.classList.add("ausboss-audio-play");
|
||||
const loopButton = makeToolButton("LOOP", "Play it again and again");
|
||||
stage.append(canvas, status);
|
||||
// Play and Loop sit either side of the wave, so neither covers a handle.
|
||||
loopButton.classList.add("ausboss-audio-play");
|
||||
const waveRow = document.createElement("div");
|
||||
waveRow.className = "ausboss-audio-wave-row";
|
||||
waveRow.append(playButton, stage, loopButton);
|
||||
|
||||
const values = document.createElement("div");
|
||||
values.className = "ausboss-video-values";
|
||||
const makeValue = (labelText, title) => {
|
||||
const wrap = document.createElement("div");
|
||||
wrap.className = "ausboss-video-value";
|
||||
const label = document.createElement("label");
|
||||
label.textContent = labelText;
|
||||
const input = document.createElement("input");
|
||||
input.type = "text";
|
||||
input.inputMode = "decimal";
|
||||
input.placeholder = "m:ss.s";
|
||||
input.title = title;
|
||||
input.addEventListener("keydown", (event) => {
|
||||
keepKeyInField(event);
|
||||
if (event.key === "Enter") input.dispatchEvent(new Event("change"));
|
||||
});
|
||||
wrap.append(label, input);
|
||||
return { wrap, input };
|
||||
// Start, Duration and Ends at, as on the video nodes, in seconds.
|
||||
const fieldsRow = document.createElement("div");
|
||||
fieldsRow.className = "ausboss-audio-fields";
|
||||
const fields = {};
|
||||
const commit = (key, value) => {
|
||||
const state = node.__ausbossLoadAudio;
|
||||
if (!state || state.duration <= 0) return;
|
||||
writeTrim(state, typedPick(key, value, currentWindow(state), state.duration, MIN_TRIM_SECONDS));
|
||||
};
|
||||
const startValue = makeValue("IN", "Where the sound starts. Type seconds (12.5) or a time (0:12.5).");
|
||||
const endValue = makeValue("OUT", "Where the sound stops. Type seconds (24) or a time (0:24.0).");
|
||||
// How long the output is, beside OUT and off the wave. Red when Max
|
||||
// length cuts the pick short; the tooltip says how much.
|
||||
const summary = document.createElement("div");
|
||||
summary.className = "ausboss-audio-length";
|
||||
values.append(startValue.wrap, endValue.wrap);
|
||||
const row = document.createElement("div");
|
||||
row.className = "ausboss-audio-row";
|
||||
row.append(playButton, values, summary, loopButton);
|
||||
root.append(stage, row);
|
||||
for (const [key, caption] of [["start", "Start"], ["duration", "Duration"], ["end", "Ends at"]]) {
|
||||
const cell = document.createElement("div");
|
||||
cell.className = "ausboss-audio-field";
|
||||
const label = document.createElement("span");
|
||||
label.className = "ausboss-audio-field-label";
|
||||
label.textContent = caption;
|
||||
const go = document.createElement("button");
|
||||
go.type = "button";
|
||||
go.className = "ausboss-audio-go";
|
||||
go.textContent = "linked";
|
||||
go.hidden = true;
|
||||
label.append(go);
|
||||
const control = makeScrubInput({
|
||||
value: 0, min: 0, max: 86400, step: 0.1, fineStep: 0.01, decimals: 2, minDecimals: 0, unit: "s", unitWidth: 14,
|
||||
allowEmpty: true,
|
||||
onChange: (value) => commit(key, value),
|
||||
onSettle: () => notifyAusbossChange(),
|
||||
});
|
||||
cell.append(label, control.root);
|
||||
fieldsRow.append(cell);
|
||||
fields[key] = { cell, control, go };
|
||||
}
|
||||
root.append(waveRow, fieldsRow);
|
||||
|
||||
const widget = node.addDOMWidget(PANEL_WIDGET, "ausboss_audio", root, {
|
||||
serialize: false,
|
||||
@@ -435,22 +456,11 @@ function buildPanel(node) {
|
||||
audio.preload = "auto";
|
||||
const abort = new AbortController();
|
||||
const state = node.__ausbossLoadAudio = {
|
||||
node, root, stage, canvas, status, playButton, loopButton, summary, audio,
|
||||
startInput: startValue.input, endInput: endValue.input,
|
||||
node, root, stage, canvas, status, playButton, loopButton, audio, fields,
|
||||
audioWidget, startWidget, endWidget, maxWidget: findWidget(node, "max_seconds"),
|
||||
duration: 0, peaks: null, abort, writing: false, dragging: false, raf: 0, loading: 0,
|
||||
};
|
||||
|
||||
// Clicking the red length moves OUT to where the sound really ends.
|
||||
summary.addEventListener("click", (event) => {
|
||||
const win = currentWindow(state);
|
||||
if (win.cut <= 0.005) return;
|
||||
event.preventDefault();
|
||||
event.stopPropagation();
|
||||
writeTrim(state, { start: win.start, end: win.keepEnd });
|
||||
notifyAusbossChange();
|
||||
});
|
||||
|
||||
const updateLoop = () => loopButton.classList.toggle("active", loopEnabled(node));
|
||||
updateLoop();
|
||||
loopButton.addEventListener("click", (event) => {
|
||||
@@ -509,8 +519,6 @@ function buildPanel(node) {
|
||||
watchWidget(state, startWidget);
|
||||
watchWidget(state, endWidget);
|
||||
watchWidget(state, state.maxWidget);
|
||||
startValue.input.addEventListener("change", () => commitTimeInput(state, "start", startValue.input));
|
||||
endValue.input.addEventListener("change", () => commitTimeInput(state, "end", endValue.input));
|
||||
const observer = new ResizeObserver(() => draw(state));
|
||||
observer.observe(stage);
|
||||
state.observer = observer;
|
||||
|
||||
+145
-355
@@ -1,9 +1,11 @@
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { chainCallback, keepDomWidgetWidthAuto, notifyAusbossChange } from "../shared/index.mjs";
|
||||
import { keepKeyInField } from "../shared/canvas_passthrough.mjs";
|
||||
import { WIDGET_FRAME, fillNodeHeight } from "../shared/panel_layout.mjs";
|
||||
import { formatTimecode, parseTimecode } from "../shared/timecode.mjs";
|
||||
import { inputNumber } from "../shared/clip_rate.mjs";
|
||||
import { clipInfo, windowSeconds } from "../shared/timeline_math.mjs";
|
||||
import { mountTransformTrim } from "../shared/transform_trim.mjs";
|
||||
import { linkedNodeInfo, timingUnits } from "../shared/timing_links.mjs";
|
||||
import { parseImageReference } from "../shared/pad_canvas.mjs";
|
||||
import {
|
||||
mediaInfo,
|
||||
@@ -19,42 +21,31 @@ import {
|
||||
suppressCoreVideoPreview,
|
||||
} from "../shared/video_ui.mjs";
|
||||
import {
|
||||
MIN_TRIM_SECONDS,
|
||||
clampTrimSeek,
|
||||
closestTrimHandle,
|
||||
dragTrimHandle,
|
||||
loadSummary,
|
||||
playbackBoundaryAction,
|
||||
singleFrameFraction,
|
||||
singleFrameTime,
|
||||
slideTrimWindow,
|
||||
trimAfterFramePick,
|
||||
trimBounds,
|
||||
trimFractions,
|
||||
} from "./trim_preview.mjs";
|
||||
|
||||
const NODE_NAME = "AUSBOSS_NODES_LoadVideo";
|
||||
const PREVIEW_WIDGET = "ausboss_load_video_viewer";
|
||||
const PREVIEW_CHROME = 88;
|
||||
// The shared video root's CSS minimum: 8px root padding + the stage's 112px
|
||||
// floor + 6px gap + the 76px trim strip. The declared floor adds
|
||||
// WIDGET_FRAME on top, or the element comes up short and clips the trim
|
||||
// strip's bottom border at the node's minimum height.
|
||||
const PANEL_MIN_HEIGHT = 202;
|
||||
// 8px root padding + the stage's 112px floor + 6px gap + the 152px timing
|
||||
// card (js/shared/transform_trim.mjs). The declared floor adds WIDGET_FRAME
|
||||
// on top, or the element comes up short and clips the card's bottom border
|
||||
// at the node's minimum height.
|
||||
const PANEL_MIN_HEIGHT = 278;
|
||||
// Height the node opens at. It used to fall out of the panel's computeSize;
|
||||
// with the panel free to follow the node, the default has to be stated. The
|
||||
// chrome covers the transport row and the trim strip under the player.
|
||||
const DEFAULT_NODE_SIZE = [380, responsivePreviewHeight(380) + PREVIEW_CHROME + 190];
|
||||
// extra covers the tools over the picture and the timing card under it.
|
||||
const DEFAULT_NODE_SIZE = [380, responsivePreviewHeight(380) + PREVIEW_CHROME + 210];
|
||||
const TRIM_EPSILON = 0.04;
|
||||
|
||||
function findWidget(node, name) {
|
||||
return node.widgets?.find((widget) => widget.name === name);
|
||||
}
|
||||
|
||||
function roundHundredth(value) {
|
||||
return Math.round(Number(value || 0) * 100) / 100;
|
||||
}
|
||||
|
||||
function getLoopEnabled(node) {
|
||||
node.properties ??= {};
|
||||
if (node.properties.ausboss_load_video_loop === undefined) {
|
||||
@@ -66,14 +57,25 @@ function getLoopEnabled(node) {
|
||||
function setEmpty(state, text) {
|
||||
state.stage.classList.add("is-empty");
|
||||
state.status.textContent = text;
|
||||
state.status.hidden = false;
|
||||
}
|
||||
|
||||
function setReady(state, text) {
|
||||
state.stage.classList.remove("is-empty");
|
||||
state.status.textContent = text;
|
||||
state.status.hidden = !text;
|
||||
}
|
||||
|
||||
// The run's window in seconds, as the timing card works it out (Start, and
|
||||
// the Duration or the end), or the plain start/end before the video's frame
|
||||
// rate is known.
|
||||
function currentBounds(state) {
|
||||
const info = clipInfo(state.metadata);
|
||||
const window = state.card?.window?.();
|
||||
if (info.fps > 0 && window && Number.isFinite(window.first)) {
|
||||
const last = Number.isFinite(window.last) ? window.last : info.count - 1;
|
||||
return { start: window.first / info.fps, end: Math.min(info.duration || Infinity, (last + 1) / info.fps) };
|
||||
}
|
||||
return trimBounds(state.video.duration, state.startWidget.value, state.endWidget.value);
|
||||
}
|
||||
|
||||
@@ -82,271 +84,116 @@ function isSingleFrame(state) {
|
||||
}
|
||||
|
||||
// Where the preview belongs for the current mode: the chosen frame, or the
|
||||
// trim window's IN point.
|
||||
// clip's start.
|
||||
function homeTime(state) {
|
||||
return isSingleFrame(state)
|
||||
? singleFrameTime(state.video.duration, state.startWidget.value)
|
||||
: currentBounds(state).start;
|
||||
}
|
||||
|
||||
function applyMode(state) {
|
||||
const single = isSingleFrame(state);
|
||||
state.root.classList.toggle("is-single-frame", single);
|
||||
state.frameButton?.classList.toggle("active", single);
|
||||
state.startLabel.textContent = single ? "AT" : "IN";
|
||||
state.range.title = single
|
||||
? "Drag or click to choose the frame that loads"
|
||||
: "Drag either handle to set the trim; drag the selected range to move it";
|
||||
updateTrimFace(state);
|
||||
}
|
||||
|
||||
function widgetLinkDriven(node, name) {
|
||||
return !!node.inputs?.some((input) => input?.widget?.name === name && input.link != null);
|
||||
}
|
||||
|
||||
// The honest "what will actually load" text for the label under the trim
|
||||
// strip, or "" when there is nothing honest to say. Any of the deciding
|
||||
// widgets driven by a link means its face value is stale, so the count
|
||||
// drops out rather than guessing.
|
||||
function trimLoadSummary(state, duration, bounds) {
|
||||
const deciders = ["start_seconds", "end_seconds", "every_nth", "max_frames", "single_frame"];
|
||||
if (deciders.some((name) => widgetLinkDriven(state.node, name))) return "";
|
||||
return loadSummary(
|
||||
duration,
|
||||
bounds,
|
||||
state.sourceFps,
|
||||
state.everyNthWidget?.value,
|
||||
state.maxFramesWidget?.value,
|
||||
isSingleFrame(state),
|
||||
);
|
||||
}
|
||||
|
||||
// The <video> element never exposes the source frame rate, so the frame
|
||||
// count in the label needs a probe of the pack's metadata route. Fail-soft:
|
||||
// on any error the label simply omits the count and the timecodes stand.
|
||||
async function probeSourceFps(state, name) {
|
||||
state.sourceFps = 0;
|
||||
// Frame rate and the real frame count come from the pack's metadata route:
|
||||
// the <video> element knows neither, and its length rounds up a frame.
|
||||
// Fail-soft: without it the card waits for a source.
|
||||
async function probeMetadata(state, name) {
|
||||
state.metadata = null;
|
||||
try {
|
||||
const query = new URLSearchParams({ source_mode: "input folder", video: name });
|
||||
const type = parseImageReference(name)?.type;
|
||||
if (type && type !== "input") {
|
||||
query.set("source_mode", "local path");
|
||||
query.set("local_path", name);
|
||||
}
|
||||
const response = await api.fetchApi(`/ausboss/transform/video/metadata?${query}`);
|
||||
if (!response.ok) return;
|
||||
const metadata = await response.json();
|
||||
if (state.videoWidget.value !== name) return; // a newer selection won
|
||||
state.sourceFps = Math.max(0, Number(metadata?.fps) || 0);
|
||||
updateTrimFace(state);
|
||||
state.metadata = metadata;
|
||||
state.card?.sync();
|
||||
} catch (_error) {
|
||||
/* no fps, no frame count */
|
||||
/* no metadata, no timing card numbers */
|
||||
}
|
||||
}
|
||||
|
||||
function writeWidget(widget, value) {
|
||||
const next = roundHundredth(value);
|
||||
if (Math.abs(Number(widget.value || 0) - next) < 0.0001) return;
|
||||
widget.value = next;
|
||||
widget.callback?.(next);
|
||||
function setWidget(widget, value) {
|
||||
if (!widget) return;
|
||||
if (widget.value === value) return;
|
||||
widget.value = value;
|
||||
widget.callback?.(value);
|
||||
}
|
||||
|
||||
function writeTrim(state, bounds, seekEdge = null) {
|
||||
state.writing = true;
|
||||
try {
|
||||
writeWidget(state.startWidget, bounds.start);
|
||||
writeWidget(state.endWidget, bounds.end);
|
||||
} finally {
|
||||
state.writing = false;
|
||||
}
|
||||
updateTrimFace(state);
|
||||
if (state.video.readyState >= 1 && seekEdge) {
|
||||
state.video.currentTime = seekEdge === "start"
|
||||
? bounds.start
|
||||
: Math.max(bounds.start, bounds.end - TRIM_EPSILON);
|
||||
}
|
||||
state.node.setDirtyCanvas?.(true, true);
|
||||
state.node.graph?.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
|
||||
function writeFrameTime(state, timeValue) {
|
||||
const duration = Number.isFinite(state.video.duration) ? state.video.duration : 0;
|
||||
const time = singleFrameTime(duration, timeValue);
|
||||
state.writing = true;
|
||||
try {
|
||||
writeWidget(state.startWidget, time);
|
||||
} finally {
|
||||
state.writing = false;
|
||||
}
|
||||
updateTrimFace(state);
|
||||
if (state.video.readyState >= 1) state.video.currentTime = time;
|
||||
state.node.setDirtyCanvas?.(true, true);
|
||||
state.node.graph?.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
|
||||
function updateTrimFace(state) {
|
||||
const duration = Number.isFinite(state.video.duration) ? state.video.duration : 0;
|
||||
if (isSingleFrame(state)) {
|
||||
const time = singleFrameTime(duration, state.startWidget.value);
|
||||
const percent = singleFrameFraction(duration, time) * 100;
|
||||
state.selection.style.width = "0%";
|
||||
state.startHandle.style.left = `${percent}%`;
|
||||
if (document.activeElement !== state.startInput) state.startInput.value = formatTimecode(time);
|
||||
// "1 frame at 0:05.2" - the count comes from the same pure summary the
|
||||
// trim label uses, so both modes report what one Run will load.
|
||||
const summary = trimLoadSummary(state, duration, null);
|
||||
state.durationLabel.textContent = duration > 0
|
||||
? `${summary || "frame"} at ${formatTimecode(time)}`
|
||||
: "load a video to pick a frame";
|
||||
if (duration > 0) state.status.textContent = `frame at ${formatTimecode(time)}`;
|
||||
return;
|
||||
}
|
||||
const bounds = trimBounds(duration, state.startWidget.value, state.endWidget.value);
|
||||
const fractions = trimFractions(duration, bounds);
|
||||
const startPercent = fractions.start * 100;
|
||||
const endPercent = fractions.end * 100;
|
||||
state.selection.style.left = `${startPercent}%`;
|
||||
state.selection.style.width = `${Math.max(0, endPercent - startPercent)}%`;
|
||||
state.startHandle.style.left = `${startPercent}%`;
|
||||
state.endHandle.style.left = `${endPercent}%`;
|
||||
if (document.activeElement !== state.startInput) state.startInput.value = formatTimecode(bounds.start);
|
||||
if (document.activeElement !== state.endInput) state.endInput.value = formatTimecode(bounds.end);
|
||||
const selected = Math.max(0, bounds.end - bounds.start);
|
||||
// The preview cannot re-render every_nth or max_frames, so the label
|
||||
// reports them instead: "0:04.0 of 0:10.0 · 48 frames @ 12 fps".
|
||||
const summary = trimLoadSummary(state, duration, bounds);
|
||||
state.durationLabel.textContent = duration > 0
|
||||
? `${formatTimecode(selected)} of ${formatTimecode(duration)}${summary ? ` · ${summary}` : ""}`
|
||||
: "load a video to trim";
|
||||
state.status.textContent = duration > 0
|
||||
? `${formatTimecode(bounds.start)} → ${formatTimecode(bounds.end)}`
|
||||
: state.status.textContent;
|
||||
}
|
||||
|
||||
function installTrimDrag(state) {
|
||||
const { range } = state;
|
||||
const infoAt = (clientX) => {
|
||||
const rect = range.getBoundingClientRect();
|
||||
const duration = Number.isFinite(state.video.duration) ? state.video.duration : 0;
|
||||
const bounds = currentBounds(state);
|
||||
const fractions = trimFractions(duration, bounds);
|
||||
const x = clientX - rect.left;
|
||||
const hit = 10;
|
||||
let zone = "jump";
|
||||
const closest = closestTrimHandle(
|
||||
fractions.start,
|
||||
fractions.end,
|
||||
rect.width ? x / rect.width : 0,
|
||||
rect.width,
|
||||
hit,
|
||||
);
|
||||
if (closest) zone = closest;
|
||||
else if (x > fractions.start * rect.width && x < fractions.end * rect.width) zone = "move";
|
||||
return { rect, duration, bounds, fraction: rect.width ? x / rect.width : 0, zone };
|
||||
};
|
||||
|
||||
range.addEventListener("pointermove", (event) => {
|
||||
if (state.dragging) return;
|
||||
if (isSingleFrame(state)) {
|
||||
range.style.cursor = "ew-resize";
|
||||
return;
|
||||
}
|
||||
const { zone } = infoAt(event.clientX);
|
||||
range.style.cursor = zone === "start" || zone === "end"
|
||||
? "ew-resize"
|
||||
: zone === "move" ? "grab" : "pointer";
|
||||
});
|
||||
|
||||
range.addEventListener("pointerdown", (event) => {
|
||||
if (event.button !== 0 || state.dragging) return;
|
||||
const initial = infoAt(event.clientX);
|
||||
if (initial.duration <= 0) return;
|
||||
event.preventDefault();
|
||||
event.stopPropagation();
|
||||
if (isSingleFrame(state)) {
|
||||
// Frame-picker gesture: the strip is one scrub rail. A click already
|
||||
// sets the frame; a drag keeps refining it until release.
|
||||
state.dragging = true;
|
||||
try { range.setPointerCapture(event.pointerId); } catch (_error) { /* mouse fallback */ }
|
||||
const scrub = (clientX) => {
|
||||
const next = infoAt(clientX);
|
||||
writeFrameTime(state, next.fraction * next.duration);
|
||||
};
|
||||
scrub(event.clientX);
|
||||
const move = (moveEvent) => {
|
||||
if (!(moveEvent.buttons & 1)) { finish(); return; }
|
||||
scrub(moveEvent.clientX);
|
||||
};
|
||||
const finish = () => {
|
||||
if (!state.dragging) return;
|
||||
state.dragging = false;
|
||||
range.removeEventListener("pointermove", move);
|
||||
try { range.releasePointerCapture(event.pointerId); } catch (_error) { /* already released */ }
|
||||
// One notification per gesture: the frame changed on pointerdown.
|
||||
notifyAusbossChange();
|
||||
};
|
||||
range.addEventListener("pointermove", move);
|
||||
range.addEventListener("pointerup", finish, { once: true });
|
||||
range.addEventListener("pointercancel", finish, { once: true });
|
||||
range.addEventListener("lostpointercapture", finish, { once: true });
|
||||
return;
|
||||
}
|
||||
state.dragging = true;
|
||||
try { range.setPointerCapture(event.pointerId); } catch (_error) { /* mouse fallback */ }
|
||||
const downX = event.clientX;
|
||||
const anchorTime = initial.fraction * initial.duration;
|
||||
let moved = false;
|
||||
|
||||
const apply = (clientX) => {
|
||||
const next = infoAt(clientX);
|
||||
if (initial.zone === "start" || initial.zone === "end") {
|
||||
writeTrim(
|
||||
state,
|
||||
dragTrimHandle(next.duration, currentBounds(state), initial.zone, next.fraction),
|
||||
initial.zone,
|
||||
);
|
||||
} else if (initial.zone === "move") {
|
||||
const pointerTime = next.fraction * next.duration;
|
||||
writeTrim(
|
||||
state,
|
||||
slideTrimWindow(next.duration, initial.bounds, pointerTime - anchorTime),
|
||||
"start",
|
||||
);
|
||||
} else {
|
||||
const span = initial.bounds.end - initial.bounds.start;
|
||||
const centeredStart = next.fraction * next.duration - span / 2;
|
||||
writeTrim(
|
||||
state,
|
||||
slideTrimWindow(next.duration, initial.bounds, centeredStart - initial.bounds.start),
|
||||
"start",
|
||||
);
|
||||
// The timing card: the clip node's own, with Start, Duration and Ends at,
|
||||
// or, while FRAME is on, its frame picker. The player above it is the
|
||||
// picture, so the card leaves out its own Play button.
|
||||
function mountCard(state) {
|
||||
const { node } = state;
|
||||
const widgetFor = (name) => node.widgets?.find((widget) => widget.name === name);
|
||||
const fps = () => clipInfo(state.metadata).fps;
|
||||
const single = isSingleFrame(state);
|
||||
const card = mountTransformTrim({
|
||||
get: (name, fallback) => {
|
||||
if (name === "frame_index") return state.head ?? 0;
|
||||
const widget = widgetFor(name);
|
||||
return widget ? widget.value : fallback;
|
||||
},
|
||||
set: (name, value) => {
|
||||
state.writing = true;
|
||||
try { setWidget(widgetFor(name), value); } finally { state.writing = false; }
|
||||
if (state.video.readyState >= 1 && !state.video.paused) return;
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
},
|
||||
has: (name) => Boolean(widgetFor(name)),
|
||||
driven: (name) => widgetLinkDriven(node, name),
|
||||
number: (name, fallback) => inputNumber(node, name, fallback),
|
||||
metadata: () => state.metadata,
|
||||
trim: !single,
|
||||
play: false,
|
||||
isPlaying: () => !state.video.paused,
|
||||
onTogglePlayback: () => {},
|
||||
onSeek: (frame, settled) => {
|
||||
state.head = frame;
|
||||
const rate = fps();
|
||||
if (single && rate > 0) {
|
||||
state.writing = true;
|
||||
try { setWidget(state.startWidget, windowSeconds(clipInfo(state.metadata), frame, frame).start_seconds); } finally { state.writing = false; }
|
||||
}
|
||||
};
|
||||
|
||||
const move = (moveEvent) => {
|
||||
if (!(moveEvent.buttons & 1)) { finish(moveEvent); return; }
|
||||
if (!moved && Math.abs(moveEvent.clientX - downX) < 3) return;
|
||||
moved = true;
|
||||
apply(moveEvent.clientX);
|
||||
};
|
||||
const finish = (finishEvent) => {
|
||||
if (!state.dragging) return;
|
||||
state.dragging = false;
|
||||
range.removeEventListener("pointermove", move);
|
||||
try { range.releasePointerCapture(event.pointerId); } catch (_error) { /* already released */ }
|
||||
if (!moved && initial.zone === "jump") {
|
||||
const next = infoAt(finishEvent?.clientX ?? downX);
|
||||
const span = initial.bounds.end - initial.bounds.start;
|
||||
const centeredStart = next.fraction * next.duration - span / 2;
|
||||
writeTrim(
|
||||
state,
|
||||
slideTrimWindow(next.duration, initial.bounds, centeredStart - initial.bounds.start),
|
||||
"start",
|
||||
);
|
||||
}
|
||||
// One notification per gesture, on release — never per pointermove.
|
||||
if (moved || initial.zone === "jump") notifyAusbossChange();
|
||||
};
|
||||
range.addEventListener("pointermove", move);
|
||||
range.addEventListener("pointerup", finish, { once: true });
|
||||
range.addEventListener("pointercancel", finish, { once: true });
|
||||
range.addEventListener("lostpointercapture", finish, { once: true });
|
||||
if (state.video.readyState >= 1 && rate > 0) state.video.currentTime = Math.min(state.video.duration || Infinity, (frame + 0.5) / rate);
|
||||
if (settled) notifyAusbossChange();
|
||||
},
|
||||
onCommit: notifyAusbossChange,
|
||||
linkInfo: (name) => linkedNodeInfo(node, name),
|
||||
units: () => timingUnits(),
|
||||
setUnits: (next) => timingUnits(next),
|
||||
});
|
||||
state.card?.root.remove();
|
||||
state.card = card;
|
||||
state.cardSlot.append(card.root);
|
||||
return card;
|
||||
}
|
||||
|
||||
function applyMode(state) {
|
||||
const single = isSingleFrame(state);
|
||||
state.root.classList.toggle("is-single-frame", single);
|
||||
state.frameButton?.classList.toggle("active", single);
|
||||
if (state.cardMode !== single) {
|
||||
state.cardMode = single;
|
||||
mountCard(state);
|
||||
}
|
||||
state.card?.sync();
|
||||
}
|
||||
|
||||
// The playhead follows the player, so the card's marker and readout move
|
||||
// with the picture.
|
||||
function followPlayer(state) {
|
||||
const rate = clipInfo(state.metadata).fps;
|
||||
if (!(rate > 0)) return;
|
||||
const frame = Math.max(0, Math.floor(state.video.currentTime * rate + 1e-6));
|
||||
if (frame === state.head) return;
|
||||
state.head = frame;
|
||||
state.card?.sync();
|
||||
}
|
||||
|
||||
function installPlayback(state) {
|
||||
@@ -356,9 +203,9 @@ function installPlayback(state) {
|
||||
state.playRaf = 0;
|
||||
};
|
||||
const tick = () => {
|
||||
if (video.paused || video.ended) { stopTick(); return; }
|
||||
// Frame picking plays the whole source freely — the trim window is not
|
||||
// in force, so there is no boundary to loop or stop at.
|
||||
if (video.paused || video.ended) { stopTick(); followPlayer(state); return; }
|
||||
// Frame picking plays the whole source freely - the clip is not in
|
||||
// force, so there is no boundary to loop or stop at.
|
||||
if (!isSingleFrame(state)) {
|
||||
const bounds = currentBounds(state);
|
||||
const action = playbackBoundaryAction(video.currentTime, bounds, getLoopEnabled(state.node), TRIM_EPSILON);
|
||||
@@ -367,9 +214,11 @@ function installPlayback(state) {
|
||||
} else if (action === "stop") {
|
||||
video.pause();
|
||||
video.currentTime = Math.max(bounds.start, bounds.end - TRIM_EPSILON);
|
||||
followPlayer(state);
|
||||
return;
|
||||
}
|
||||
}
|
||||
followPlayer(state);
|
||||
state.playRaf = requestAnimationFrame(tick);
|
||||
};
|
||||
video.addEventListener("play", () => {
|
||||
@@ -390,7 +239,7 @@ function installPlayback(state) {
|
||||
state.correctingSeek = true;
|
||||
video.currentTime = target;
|
||||
}, { signal: state.abort.signal });
|
||||
video.addEventListener("seeked", () => { state.correctingSeek = false; }, { signal: state.abort.signal });
|
||||
video.addEventListener("seeked", () => { state.correctingSeek = false; followPlayer(state); }, { signal: state.abort.signal });
|
||||
state.stopTick = stopTick;
|
||||
}
|
||||
|
||||
@@ -401,8 +250,8 @@ function watchTrimWidget(state, widget) {
|
||||
widget.callback = function (...args) {
|
||||
const result = prior?.apply(this, args);
|
||||
if (!state.writing) {
|
||||
updateTrimFace(state);
|
||||
if (state.video.readyState >= 1) state.video.currentTime = homeTime(state);
|
||||
state.card?.sync();
|
||||
if (state.video.readyState >= 1 && state.video.paused) state.video.currentTime = homeTime(state);
|
||||
}
|
||||
return result;
|
||||
};
|
||||
@@ -416,7 +265,10 @@ function watchFrameWidget(state, widget) {
|
||||
const result = prior?.apply(this, args);
|
||||
if (!isSingleFrame(state)) {
|
||||
const reopened = trimAfterFramePick(state.startWidget.value, state.endWidget.value);
|
||||
if (reopened) writeTrim(state, reopened);
|
||||
if (reopened) {
|
||||
state.writing = true;
|
||||
try { setWidget(state.endWidget, reopened.end); } finally { state.writing = false; }
|
||||
}
|
||||
}
|
||||
applyMode(state);
|
||||
if (state.video.readyState >= 1) state.video.currentTime = homeTime(state);
|
||||
@@ -424,28 +276,6 @@ function watchFrameWidget(state, widget) {
|
||||
};
|
||||
}
|
||||
|
||||
function commitTimeInput(state, edge, input) {
|
||||
const duration = Number.isFinite(state.video.duration) ? state.video.duration : 0;
|
||||
if (duration <= 0) return;
|
||||
const seconds = parseTimecode(input.value);
|
||||
if (seconds === null) {
|
||||
// Malformed entry: restore the previous value and keep focus for a retry.
|
||||
const bounds = currentBounds(state);
|
||||
input.value = formatTimecode(edge === "start" ? bounds.start : bounds.end);
|
||||
return;
|
||||
}
|
||||
if (isSingleFrame(state)) {
|
||||
writeFrameTime(state, seconds);
|
||||
notifyAusbossChange();
|
||||
input.blur();
|
||||
return;
|
||||
}
|
||||
const fraction = seconds / duration;
|
||||
writeTrim(state, dragTrimHandle(duration, currentBounds(state), edge, fraction), edge);
|
||||
notifyAusbossChange();
|
||||
input.blur();
|
||||
}
|
||||
|
||||
function buildPreview(node) {
|
||||
if (node.__ausbossLoadVideo) return node.__ausbossLoadVideo;
|
||||
const videoWidget = findWidget(node, "video");
|
||||
@@ -461,6 +291,8 @@ function buildPreview(node) {
|
||||
hideCanvasWidget(startWidget);
|
||||
hideCanvasWidget(endWidget);
|
||||
if (frameWidget) hideCanvasWidget(frameWidget);
|
||||
// Every nth and the frame limit (the Duration) live in the timing card;
|
||||
// the widget card hides their widgets (js/widget_cards).
|
||||
|
||||
const root = document.createElement("div");
|
||||
root.className = "ausboss-video-root";
|
||||
@@ -475,7 +307,7 @@ function buildPreview(node) {
|
||||
status.textContent = "Choose or upload a video";
|
||||
const tools = document.createElement("div");
|
||||
tools.className = "ausboss-video-tools";
|
||||
const loopButton = makeToolButton("LOOP", "Loop playback inside the selected trim");
|
||||
const loopButton = makeToolButton("LOOP", "Play the clip again and again");
|
||||
loopButton.classList.add("ausboss-video-loop");
|
||||
const reloadButton = makeToolButton("↻", "Reload the selected source video");
|
||||
const frameButton = frameWidget
|
||||
@@ -485,45 +317,12 @@ function buildPreview(node) {
|
||||
tools.append(loopButton, reloadButton);
|
||||
stage.append(video, status, tools);
|
||||
|
||||
const trim = document.createElement("div");
|
||||
trim.className = "ausboss-video-trim";
|
||||
const range = document.createElement("div");
|
||||
range.className = "ausboss-video-range";
|
||||
range.title = "Drag either handle to set the trim; drag the selected range to move it";
|
||||
const selection = document.createElement("div");
|
||||
selection.className = "ausboss-video-selection";
|
||||
const startHandle = document.createElement("div");
|
||||
startHandle.className = "ausboss-video-handle";
|
||||
const endHandle = document.createElement("div");
|
||||
endHandle.className = "ausboss-video-handle is-end";
|
||||
range.append(selection, startHandle, endHandle);
|
||||
const values = document.createElement("div");
|
||||
values.className = "ausboss-video-values";
|
||||
const makeValue = (labelText) => {
|
||||
const wrap = document.createElement("div");
|
||||
wrap.className = "ausboss-video-value";
|
||||
const label = document.createElement("label");
|
||||
label.textContent = labelText;
|
||||
const input = document.createElement("input");
|
||||
input.type = "text";
|
||||
input.inputMode = "decimal";
|
||||
input.placeholder = "m:ss.s";
|
||||
input.title = `${labelText} point - type seconds (95.5) or a timecode (1:35.5, 1:02:03.5)`;
|
||||
input.addEventListener("keydown", (event) => {
|
||||
keepKeyInField(event);
|
||||
if (event.key === "Enter") input.dispatchEvent(new Event("change"));
|
||||
});
|
||||
wrap.append(label, input);
|
||||
return { wrap, input, label };
|
||||
};
|
||||
const startValue = makeValue("IN");
|
||||
const endValue = makeValue("OUT");
|
||||
endValue.wrap.classList.add("is-end");
|
||||
const durationLabel = document.createElement("div");
|
||||
durationLabel.className = "ausboss-video-duration";
|
||||
values.append(startValue.wrap, durationLabel, endValue.wrap);
|
||||
trim.append(range, values);
|
||||
root.append(stage, trim);
|
||||
// The timing card goes here (mountCard): the clip node's Start,
|
||||
// Duration and Ends at, or the frame picker while FRAME is on.
|
||||
const cardSlot = document.createElement("div");
|
||||
cardSlot.className = "ausboss-video-timing";
|
||||
cardSlot.style.flex = "none";
|
||||
root.append(stage, cardSlot);
|
||||
|
||||
const widget = node.addDOMWidget(PREVIEW_WIDGET, "ausboss_video", root, {
|
||||
serialize: false,
|
||||
@@ -544,15 +343,9 @@ function buildPreview(node) {
|
||||
const abort = new AbortController();
|
||||
const state = node.__ausbossLoadVideo = {
|
||||
node, root, stage, video, status, loopButton, reloadButton, frameButton,
|
||||
trim, range, selection, startHandle, endHandle, startInput: startValue.input,
|
||||
endInput: endValue.input, startLabel: startValue.label, durationLabel,
|
||||
cardSlot, card: null, cardMode: null, metadata: null, head: 0,
|
||||
videoWidget, startWidget, endWidget, frameWidget,
|
||||
// Absent on a backend without these widgets; the summary then reads
|
||||
// its defaults (keep every frame, no cap).
|
||||
everyNthWidget: findWidget(node, "every_nth"),
|
||||
maxFramesWidget: findWidget(node, "max_frames"),
|
||||
sourceFps: 0,
|
||||
widget, abort, writing: false, dragging: false, correctingSeek: false, playRaf: 0,
|
||||
widget, abort, writing: false, correctingSeek: false, playRaf: 0,
|
||||
};
|
||||
|
||||
const updateLoopButton = () => loopButton.classList.toggle("active", getLoopEnabled(node));
|
||||
@@ -583,9 +376,9 @@ function buildPreview(node) {
|
||||
if (!name || typeof name !== "string") {
|
||||
video.pause();
|
||||
video.removeAttribute("src");
|
||||
state.sourceFps = 0;
|
||||
state.metadata = null;
|
||||
setEmpty(state, "Choose or upload a video");
|
||||
updateTrimFace(state);
|
||||
state.card?.sync();
|
||||
return;
|
||||
}
|
||||
const parts = splitMediaName(name);
|
||||
@@ -595,7 +388,7 @@ function buildPreview(node) {
|
||||
const type = parseImageReference(name)?.type ?? "input";
|
||||
video.src = api.apiURL(`/view?${mediaViewQuery({ ...parts, type }, "input")}`);
|
||||
video.load();
|
||||
probeSourceFps(state, name);
|
||||
probeMetadata(state, name);
|
||||
};
|
||||
state.refresh = refresh;
|
||||
reloadButton.addEventListener("click", (event) => {
|
||||
@@ -605,31 +398,23 @@ function buildPreview(node) {
|
||||
});
|
||||
|
||||
video.addEventListener("loadedmetadata", () => {
|
||||
const single = isSingleFrame(state);
|
||||
const bounds = currentBounds(state);
|
||||
if (!single && bounds.end - bounds.start < MIN_TRIM_SECONDS && video.duration >= MIN_TRIM_SECONDS) {
|
||||
writeTrim(state, { start: Math.max(0, bounds.end - MIN_TRIM_SECONDS), end: bounds.end });
|
||||
}
|
||||
video.currentTime = homeTime(state);
|
||||
setReady(state, single
|
||||
? `frame at ${formatTimecode(homeTime(state))}`
|
||||
: `${formatTimecode(currentBounds(state).start)} → ${formatTimecode(currentBounds(state).end)}`);
|
||||
// Quiet source hint: resolution and duration from the loaded metadata.
|
||||
// The card says where the clip is; the stage only names the source on hover.
|
||||
setReady(state, "");
|
||||
const sourceInfo = mediaInfo(null, video);
|
||||
durationLabel.title = sourceInfo ? `Source: ${sourceInfo}` : "";
|
||||
updateTrimFace(state);
|
||||
stage.title = sourceInfo ? `Source: ${sourceInfo}` : "";
|
||||
state.card?.sync();
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
}, { signal: abort.signal });
|
||||
video.addEventListener("error", () => setEmpty(state, "Preview could not load this video"), { signal: abort.signal });
|
||||
|
||||
installTrimDrag(state);
|
||||
installPlayback(state);
|
||||
watchTrimWidget(state, startWidget);
|
||||
watchTrimWidget(state, endWidget);
|
||||
watchFrameWidget(state, frameWidget);
|
||||
watchTrimWidget(state, findWidget(node, "every_nth"));
|
||||
watchTrimWidget(state, findWidget(node, "max_frames"));
|
||||
applyMode(state);
|
||||
startValue.input.addEventListener("change", () => commitTimeInput(state, "start", startValue.input));
|
||||
endValue.input.addEventListener("change", () => commitTimeInput(state, "end", endValue.input));
|
||||
|
||||
const priorVideoCallback = videoWidget.callback;
|
||||
videoWidget.callback = function (...args) {
|
||||
@@ -664,6 +449,11 @@ app.registerExtension({
|
||||
}
|
||||
});
|
||||
});
|
||||
// A wire into a timing input changes what the card can show.
|
||||
chainCallback(nodeType.prototype, "onConnectionsChange", function () {
|
||||
const state = this.__ausbossLoadVideo;
|
||||
if (state) queueMicrotask(() => state.card?.sync());
|
||||
});
|
||||
chainCallback(nodeType.prototype, "onRemoved", function () {
|
||||
const state = this.__ausbossLoadVideo;
|
||||
state?.stopTick?.();
|
||||
|
||||
@@ -13,6 +13,7 @@ import { makeScrubInput } from "../shared/scrub_input.mjs";
|
||||
import { isAppShortcut, keepKeyInField } from "../shared/canvas_passthrough.mjs";
|
||||
import { resetUnknownValues } from "../shared/widget_card_math.mjs";
|
||||
import { boxRowInputs, boxRowsWidget, boxesOf, hostBox, rowsInputName, workflowNodes } from "../shared/lora_box.mjs";
|
||||
import { coreLoraList, syncFileWidgets } from "../shared/lora_file_widgets.mjs";
|
||||
import {
|
||||
SETTINGS_PROPERTY as SWITCHES_PROPERTY, releaseStackGroup, releaseStackRow, renameStackGroup, stackGroupOwners, stackRowOwners,
|
||||
} from "../shared/workflow_switches.mjs";
|
||||
@@ -724,9 +725,47 @@ function commitRows(state, rows, { structural = false, record = true } = {}) {
|
||||
state.node.graph?.setDirtyCanvas(true, true);
|
||||
if (structural) renderRows(state);
|
||||
else updateRowValues(state);
|
||||
syncLoraFiles(state.node);
|
||||
if (record) notifyAusbossChange();
|
||||
}
|
||||
|
||||
// ComfyUI's missing-models list sees a row's file only through a list
|
||||
// widget, so each row that needs a file gets a hidden one
|
||||
// (js/shared/lora_file_widgets.mjs). Rows that come in over a link (from a
|
||||
// box or another node) are not this node's to list.
|
||||
function syncLoraFiles(node) {
|
||||
if (node?.type !== NODE_CLASS) return;
|
||||
const widget = node.widgets?.find((item) => item.name === "loras");
|
||||
if (!widget) return;
|
||||
const fed = node.inputs?.find((item) => item?.name === "loras")?.link != null;
|
||||
syncFileWidgets(node, fed ? [] : parseRows(widget.value), {
|
||||
installed: () => coreLoraList(globalThis.LiteGraph?.registered_node_types?.LoraLoader?.nodeData)
|
||||
?? node.__ausbossLoraState?.available ?? null,
|
||||
hide: hideWidget,
|
||||
onPick: (rowId, name) => pickRowFile(node, rowId, name),
|
||||
});
|
||||
}
|
||||
|
||||
// ComfyUI put another file in a row's hidden list (a replacement picked in
|
||||
// its Issues panel): the row takes that file.
|
||||
function pickRowFile(node, rowId, name) {
|
||||
const widget = node.widgets?.find((item) => item.name === "loras");
|
||||
if (!widget) return;
|
||||
const rows = parseRows(widget.value);
|
||||
const row = rows.find((item) => item.id === rowId);
|
||||
if (!row || row.name === name) return;
|
||||
row.name = name;
|
||||
const state = node.__ausbossLoraState;
|
||||
if (state && !state.disposed && !state.readOnly && state.widget === widget) {
|
||||
commitRows(state, rows, { structural: true });
|
||||
return;
|
||||
}
|
||||
widget.value = serializeRows(rows);
|
||||
node.graph?.setDirtyCanvas(true, true);
|
||||
syncLoraFiles(node);
|
||||
notifyAusbossChange();
|
||||
}
|
||||
|
||||
// A box shows its loader's rows the way the loader does: one strength per
|
||||
// row, or model and CLIP apart.
|
||||
function linked(state) {
|
||||
@@ -3027,6 +3066,7 @@ function syncRows(state) {
|
||||
state.toggleSnapshot = snapshotEnabled(state.rows);
|
||||
renderRows(state);
|
||||
fitNode(state);
|
||||
syncLoraFiles(state.node);
|
||||
return true;
|
||||
}
|
||||
|
||||
@@ -3324,6 +3364,12 @@ app.registerExtension({
|
||||
});
|
||||
chainCallback(nodeType.prototype, "onConfigure", function () {
|
||||
resetUnknownValues(this.widgets, RESET_UNKNOWN);
|
||||
// Now, not later: ComfyUI reads its missing-models list right after
|
||||
// the workflow is restored.
|
||||
syncLoraFiles(this);
|
||||
});
|
||||
chainCallback(nodeType.prototype, "onConnectionsChange", function () {
|
||||
syncLoraFiles(this);
|
||||
});
|
||||
},
|
||||
// A box (subgraph node) shows the rows of a loader inside once they come
|
||||
|
||||
@@ -47,6 +47,24 @@ async function stopLook(node) {
|
||||
return Boolean(running || waiting.length);
|
||||
}
|
||||
|
||||
// Take this node's own runs that still wait out of the queue: the looks and
|
||||
// touches its picture asked for (they carry an `ask`), never a run of the
|
||||
// workflow. Says how many went. The one that is running is left alone.
|
||||
async function dropWaiting(node) {
|
||||
const id = executionIdOf(app.rootGraph ?? app.graph, node);
|
||||
if (!id) return 0;
|
||||
const queue = await (await api.fetchApi("/queue")).json();
|
||||
const ours = (entry) => {
|
||||
const step = entry?.[2]?.[id];
|
||||
return step?.class_type === NODE_CLASS && String(step.inputs?.ask ?? "") !== "";
|
||||
};
|
||||
const waiting = (queue?.queue_pending ?? []).filter(ours).map((entry) => entry[1]);
|
||||
if (waiting.length) {
|
||||
await api.fetchApi("/queue", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ delete: waiting }) });
|
||||
}
|
||||
return waiting.length;
|
||||
}
|
||||
|
||||
// The picture goes above the settings: the stage widget sits ahead of the
|
||||
// node's card, whichever of the two was built first.
|
||||
function stageFirst(node) {
|
||||
@@ -92,6 +110,7 @@ function build(node) {
|
||||
const face = mountTouchStage(node, {
|
||||
lookOnButton: () => lookOnButton,
|
||||
stopLook: () => stopLook(node).catch(() => false),
|
||||
dropWaiting: () => dropWaiting(node).catch(() => 0),
|
||||
// A workflow is being opened: nothing is run on the person's behalf.
|
||||
loading: () => Boolean(app.configuringGraph),
|
||||
// Bypassed or muted, or inside a box that is: nothing is run for it.
|
||||
|
||||
@@ -26,6 +26,7 @@ import { nodeByExecutionId } from "../shared/graph_ids.mjs";
|
||||
import {
|
||||
collapseRepeatedErrors,
|
||||
emptySourceMessage,
|
||||
explainRunError,
|
||||
findEmptySources,
|
||||
nodeLacksSource,
|
||||
} from "../shared/run_check.mjs";
|
||||
@@ -174,10 +175,20 @@ function collapseAnswers() {
|
||||
api.queuePrompt = collapsing;
|
||||
}
|
||||
|
||||
// A run that stops on an error the pack knows the cause of gets one toast
|
||||
// that says what to do. ComfyUI's own error report is left as it is.
|
||||
function explainKnownErrors() {
|
||||
api.addEventListener("execution_error", ({ detail }) => {
|
||||
const known = explainRunError(detail);
|
||||
if (known) showToast({ severity: "warn", summary: known.summary, detail: known.detail, life: 30000 });
|
||||
});
|
||||
}
|
||||
|
||||
app.registerExtension({
|
||||
name: "ausboss.run_check",
|
||||
setup() {
|
||||
guardRuns();
|
||||
collapseAnswers();
|
||||
explainKnownErrors();
|
||||
},
|
||||
});
|
||||
|
||||
+31
-3
@@ -33,7 +33,7 @@ const CARD_HEIGHT = 288;
|
||||
const WRAPPER_INSET = 16;
|
||||
const HIDDEN = [
|
||||
"filename_prefix", "format", "save_metadata", "exact_name", "on_existing",
|
||||
"output_dir", "caption", "name_counter", "name_date", "name_time", "name_size", "name_batch",
|
||||
"output_dir", "caption", "name_counter", "name_date", "name_time", "name_size", "name_batch", "autosave",
|
||||
];
|
||||
const FONT = `-apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif`;
|
||||
const MONO = `ui-monospace,SFMono-Regular,Menlo,Consolas,monospace`;
|
||||
@@ -60,7 +60,10 @@ function ensureCss() {
|
||||
.ausboss-save-chip:hover{color:#fff;border-color:${BRAND}}
|
||||
.ausboss-save-chip.on{background:${BRAND};border-color:${BRAND};color:#04201d}
|
||||
.ausboss-save-chips.off .ausboss-save-chip{opacity:.4;pointer-events:none}
|
||||
.ausboss-save-embed-label{flex:0 0 30%;color:#8ba3a1;font-size:11px;user-select:none}
|
||||
.ausboss-save-embed-label{flex:none;color:#8ba3a1;font-size:11px;white-space:nowrap;user-select:none}
|
||||
.ausboss-save-half{display:flex;align-items:center;gap:6px;flex:1 1 auto;min-width:0}
|
||||
.ausboss-save-half .ausboss-card-seg{flex:1 1 0;min-width:0}
|
||||
.ausboss-save-switches{gap:12px}
|
||||
.ausboss-save-browse{position:fixed;z-index:10000;width:260px;max-height:320px;overflow:auto;padding:4px;border:1px solid #3a4047;border-radius:7px;background:#1c1f23;box-shadow:0 8px 28px rgba(0,0,0,.5);font:12px/1.3 ${FONT};color:#d8ecea}
|
||||
.ausboss-save-browse-head{padding:5px 8px 6px;border-bottom:1px solid #2c3238;color:#78908e;font:10px/1 ${FONT};letter-spacing:.08em;text-transform:uppercase;white-space:nowrap;overflow:hidden;text-overflow:ellipsis}
|
||||
.ausboss-save-browse button{display:block;width:100%;padding:6px 8px;border:none;border-radius:5px;background:transparent;color:inherit;font:inherit;text-align:left;cursor:pointer}
|
||||
@@ -282,9 +285,29 @@ function buildCard(node) {
|
||||
button.addEventListener("click", () => { if (setValue(node, "save_metadata", on)) notifyAusbossChange(); state.refresh(); });
|
||||
embedSeg.append(button); embedButtons.set(on, button);
|
||||
}
|
||||
embedRow.append(el("span", "ausboss-save-embed-label", PANEL_LABELS.embed), embedSeg);
|
||||
const embedHalf = el("div", "ausboss-save-half");
|
||||
embedHalf.append(el("span", "ausboss-save-embed-label", PANEL_LABELS.embed), embedSeg);
|
||||
embedRow.classList.add("ausboss-save-switches");
|
||||
embedRow.append(embedHalf);
|
||||
root.append(embedRow);
|
||||
|
||||
// Auto save: off shows the pictures without saving them; the Save button
|
||||
// over the picture (js/input_preview) then keeps the ones that are wanted.
|
||||
// It shares the Embed workflow row, so the card is no taller than before.
|
||||
const autoHalf = el("div", "ausboss-save-half");
|
||||
const autoSeg = el("div", "ausboss-card-seg ausboss-card-bool");
|
||||
autoSeg.setAttribute("role", "radiogroup");
|
||||
const autoButtons = new Map();
|
||||
const autoTitle = "On: every run saves its pictures. Off: the pictures only show here, and the Save button over them keeps the ones you want.";
|
||||
for (const [on, text] of [[false, PANEL_LABELS.off], [true, PANEL_LABELS.on]]) {
|
||||
const button = el("button", "", text);
|
||||
button.type = "button"; button.title = autoTitle;
|
||||
button.addEventListener("click", () => { if (setValue(node, "autosave", on)) { state.savedPath = null; notifyAusbossChange(); } state.refresh(); });
|
||||
autoSeg.append(button); autoButtons.set(on, button);
|
||||
}
|
||||
autoHalf.append(el("span", "ausboss-save-embed-label", "Auto save"), autoSeg);
|
||||
embedRow.append(autoHalf);
|
||||
|
||||
root.addEventListener("pointerdown", (event) => {
|
||||
if (event.target.closest("button,input,select")) event.stopPropagation();
|
||||
});
|
||||
@@ -330,6 +353,11 @@ function buildCard(node) {
|
||||
button.classList.toggle("on", on === embed);
|
||||
button.setAttribute("aria-checked", String(on === embed));
|
||||
}
|
||||
const auto = values.autosave !== false;
|
||||
for (const [on, button] of autoButtons) {
|
||||
button.classList.toggle("on", on === auto);
|
||||
button.setAttribute("aria-checked", String(on === auto));
|
||||
}
|
||||
};
|
||||
|
||||
const domWidget = node.addDOMWidget(WIDGET_NAME, WIDGET_NAME, root, {
|
||||
|
||||
@@ -0,0 +1,59 @@
|
||||
// App mode and the pack's cards. Pure logic, no DOM: tested under node:test.
|
||||
//
|
||||
// App mode is ComfyUI's simple form view of a workflow: a saved list of
|
||||
// fields (extra.linearData.inputs) and a Run button. Each field is a
|
||||
// widget. A node with a card or panel hides its standard widgets and shows
|
||||
// its own controls instead, and App mode reads that hidden flag too, so the
|
||||
// form drew an empty row for every field on such a node.
|
||||
//
|
||||
// While a workflow is in App mode, the widgets its form lists are shown
|
||||
// again; App mode draws its own control for each. Back in the graph they are
|
||||
// hidden again, so the node's face is unchanged. Only listed widgets move.
|
||||
import { hideWidget, isWidgetHidden, showWidget } from "./widget_visibility.mjs";
|
||||
|
||||
export const APP_MODE = "app";
|
||||
|
||||
// "app", "graph" or "builder:..." for the open workflow.
|
||||
export function workflowMode(workflow) {
|
||||
return workflow?.activeMode ?? workflow?.initialMode ?? "graph";
|
||||
}
|
||||
|
||||
// The form's fields as { nodeId, widget } pairs. A field is saved as
|
||||
// [nodeId, widgetName] or [nodeId, widgetName, { description }].
|
||||
export function appFields(linearData) {
|
||||
const inputs = Array.isArray(linearData?.inputs) ? linearData.inputs : [];
|
||||
const fields = [];
|
||||
for (const entry of inputs) {
|
||||
if (!Array.isArray(entry) || entry.length < 2) continue;
|
||||
const [nodeId, widget] = entry;
|
||||
if ((typeof nodeId !== "string" && typeof nodeId !== "number") || typeof widget !== "string") continue;
|
||||
fields.push({ nodeId: String(nodeId), widget });
|
||||
}
|
||||
return fields;
|
||||
}
|
||||
|
||||
// find(nodeId, widgetName) -> { node, widget } or null
|
||||
// changed(node) is called after a widget on that node changed visibility.
|
||||
export function createAppModeSync({ find, changed = () => {} }) {
|
||||
const lent = new Map(); // widget -> node, shown for the form
|
||||
return function sync(mode, linearData) {
|
||||
if (mode === APP_MODE) {
|
||||
for (const { nodeId, widget: name } of appFields(linearData)) {
|
||||
const found = find(nodeId, name);
|
||||
if (!found?.widget || !isWidgetHidden(found.widget)) continue;
|
||||
showWidget(found.widget);
|
||||
// An explicit false, not a deleted flag: the frontend keeps its own
|
||||
// copy of the hidden state and only hears about a written value.
|
||||
if (found.widget.options) found.widget.options.hidden = false;
|
||||
lent.set(found.widget, found.node);
|
||||
changed(found.node);
|
||||
}
|
||||
return lent.size;
|
||||
}
|
||||
for (const [widget, node] of lent) {
|
||||
if (hideWidget(widget)) changed(node);
|
||||
}
|
||||
lent.clear();
|
||||
return 0;
|
||||
};
|
||||
}
|
||||
@@ -289,3 +289,86 @@ export function repointLinks(text, nodes, groups) {
|
||||
.replace(ARROW_RE, (all, _arrow, id) => (id === undefined ? all : `${all.slice(0, -id.length - 1)}${follow(`#${id}`)}`));
|
||||
return changed ? next : null;
|
||||
}
|
||||
|
||||
// Info card: a small card with a title, rows and small print, chosen from the
|
||||
// note's menu. Its text is the note's ordinary text with one addition: a line
|
||||
// with a | in it is a row, the words before it on the left and the rest on
|
||||
// the right. A row in **bold** is the one to notice and is drawn in teal. The
|
||||
// first line is the card's title; a clock, info, warning, bulb or tick emoji
|
||||
// at its start is drawn as a small icon. A plain callout reads none of this,
|
||||
// so every note written before looks as it did, and a release before this
|
||||
// one shows the card as a plain note with its rows written out.
|
||||
const CARD_ICONS = new Map([
|
||||
["⏱", "clock"], ["⏲", "clock"], ["⏰", "clock"], ["\u{1F552}", "clock"], ["⌛", "clock"], ["⏳", "clock"],
|
||||
["ℹ", "info"], ["⚠", "warning"], ["\u{1F4A1}", "tip"], ["✅", "check"], ["✔", "check"],
|
||||
]);
|
||||
const CARD_ICON_RE = new RegExp(`^(${[...CARD_ICONS.keys()].join("|")})${VS16}?\\s*`, "u");
|
||||
export const CARD_TEMPLATE = [
|
||||
"ℹ️ Title",
|
||||
`${SMALL_MARK}A line of small print under the title.`,
|
||||
"",
|
||||
"Label | value",
|
||||
"**The one to notice | value**",
|
||||
"",
|
||||
`${SMALL_MARK}A note at the bottom.`,
|
||||
].join("\n");
|
||||
|
||||
// Where a row splits: the first | that is not inside a link, or -1.
|
||||
function cellSplit(line) {
|
||||
const masked = line.replace(WORDS_LINK_RE, (all) => " ".repeat(all.length));
|
||||
return masked.indexOf("|");
|
||||
}
|
||||
|
||||
// A row's two sides, or null when the line is not a row. The outer pipes of
|
||||
// a pasted Markdown table row are dropped.
|
||||
export function cardCells(line) {
|
||||
const bare = String(line ?? "").trim().replace(/^\|/, "").replace(/\|$/, "");
|
||||
const at = cellSplit(bare);
|
||||
if (at < 0) return null;
|
||||
let label = bare.slice(0, at).trim();
|
||||
let value = bare.slice(at + 1).trim();
|
||||
if (!label || !value) return null;
|
||||
// **label | value** marks the whole row.
|
||||
if (label.startsWith("**") && !label.slice(2).includes("**") && value.endsWith("**") && !value.slice(0, -2).includes("**")) {
|
||||
label = `${label}**`;
|
||||
value = `**${value}`;
|
||||
}
|
||||
return [label, value];
|
||||
}
|
||||
|
||||
// Text -> blocks of lines. A line is { kind: "title", icon, pieces },
|
||||
// { kind: "row", cells: [pieces, pieces], highlight } or { kind: "text",
|
||||
// pieces }; small print carries `small: true`. A Markdown table's ---|---
|
||||
// line is dropped.
|
||||
export function parseCard(text) {
|
||||
const blocks = [];
|
||||
let lines = [];
|
||||
let first = true;
|
||||
const flush = () => {
|
||||
if (lines.length) blocks.push(lines);
|
||||
lines = [];
|
||||
};
|
||||
for (const raw of String(text ?? "").replace(/\r\n?/g, "\n").split("\n")) {
|
||||
const line = raw.replace(HEADING_RE, "").trim();
|
||||
if (!line) { flush(); continue; }
|
||||
if (/^\|?[\s:|-]*-[\s:|-]*\|?$/.test(line) && line.includes("|")) continue;
|
||||
const small = SMALL_RE.test(raw);
|
||||
const cells = cardCells(line);
|
||||
let item;
|
||||
if (first && !small && !cells) {
|
||||
const icon = CARD_ICON_RE.exec(line);
|
||||
item = { kind: "title", icon: icon ? CARD_ICONS.get(icon[1]) : null, pieces: pieces(icon ? line.slice(icon[0].length) : line) };
|
||||
} else if (cells) {
|
||||
const sides = cells.map((cell) => pieces(cell));
|
||||
const shown = sides.flat().filter((piece) => piece.arrow || piece.text?.trim());
|
||||
item = { kind: "row", cells: sides, highlight: shown.length > 0 && shown.every((piece) => piece.bold) };
|
||||
} else {
|
||||
item = { kind: "text", pieces: pieces(line) };
|
||||
}
|
||||
first = false;
|
||||
if (small) item.small = true;
|
||||
lines.push(item);
|
||||
}
|
||||
flush();
|
||||
return blocks;
|
||||
}
|
||||
|
||||
@@ -9,7 +9,9 @@
|
||||
// in js/canvas_passthrough/ wires them to the page.
|
||||
|
||||
// The wheel belongs to the panel while it can still scroll that way (a long
|
||||
// text box, a list of rows), or while a text field has the keyboard.
|
||||
// text box, a list of rows). A text box with the cursor in it is no
|
||||
// exception: when it has nothing left to scroll, the wheel zooms the graph,
|
||||
// as it does over ComfyUI's own text boxes.
|
||||
export function canScrollFurther(element, delta, axis = "y") {
|
||||
if (!element || !delta) return false;
|
||||
const horizontal = axis === "x";
|
||||
@@ -25,10 +27,11 @@ export function canScrollFurther(element, delta, axis = "y") {
|
||||
// and the panel's own root? `styleOf` is getComputedStyle, passed in so this
|
||||
// stays testable without a page. `host` is the frontend's wrapper around the
|
||||
// panel: a pointer on the wrapper itself is on nothing inside the panel.
|
||||
export function panelKeepsWheel(target, event, { root, host, styleOf, activeElement }) {
|
||||
// Ctrl + wheel, which is also how a trackpad pinch arrives, always zooms the
|
||||
// graph; left to a scrolling box it would zoom the whole page.
|
||||
export function panelKeepsWheel(target, event, { root, host, styleOf }) {
|
||||
if (event.ctrlKey || event.metaKey) return false;
|
||||
for (let element = target; element && element !== root && element !== host; element = element.parentElement) {
|
||||
const tag = String(element.tagName ?? "").toUpperCase();
|
||||
if ((tag === "TEXTAREA" || tag === "INPUT") && element === activeElement) return true;
|
||||
const style = styleOf(element);
|
||||
const overflow = style?.overflowY;
|
||||
if ((overflow === "auto" || overflow === "scroll") && canScrollFurther(element, event.deltaY)) return true;
|
||||
|
||||
+64
-16
@@ -1,30 +1,78 @@
|
||||
// Read only literal numeric sources. Calculated inputs must remain unknown
|
||||
// until execution; never evaluate an arbitrary upstream graph for a label.
|
||||
export function inputNumber(node, name, fallback) {
|
||||
const input = node.inputs?.find((slot) => slot.name === name || slot.widget?.name === name);
|
||||
if (input?.link == null) return fallback;
|
||||
const graph = node.graph;
|
||||
let linkId = input.link;
|
||||
// A subgraph's output that comes straight from a literal inside it (Person
|
||||
// Swap's "Pick the model files" hands out its video fps that way) counts as
|
||||
// that literal: it is still a typed number, not a calculation.
|
||||
function tableLink(graph, id) {
|
||||
if (id == null || !graph) return null;
|
||||
return graph.getLink?.(id) ?? graph.links?.get?.(id) ?? graph.links?.[id] ?? graph._links?.get?.(id) ?? null;
|
||||
}
|
||||
|
||||
// The link into one of a node's inputs: the frontend's own accessor first
|
||||
// (newer frontends warn on reading input.link), the link table after it.
|
||||
// undefined when nothing is connected.
|
||||
function linkInto(node, index, graph) {
|
||||
if (typeof node?.getInputLink === "function" && node.graph) {
|
||||
try {
|
||||
return node.getInputLink(index) ?? undefined;
|
||||
} catch {
|
||||
return undefined;
|
||||
}
|
||||
}
|
||||
const id = node?.inputs?.[index]?.link;
|
||||
return id == null ? undefined : tableLink(node.graph ?? graph, id);
|
||||
}
|
||||
|
||||
// Where a linked input really comes from: the node and output slot at the
|
||||
// far end, through Reroutes and into a subgraph whose output is wired
|
||||
// straight to a node inside it. null for a muted or bypassed source, a loop,
|
||||
// or a subgraph output with nothing behind it. undefined when the input is
|
||||
// not linked at all.
|
||||
export function linkOrigin(node, name) {
|
||||
const index = node?.inputs?.findIndex((slot) => slot?.name === name || slot?.widget?.name === name) ?? -1;
|
||||
if (index < 0) return undefined;
|
||||
let graph = node.graph;
|
||||
let link = linkInto(node, index, graph);
|
||||
if (link === undefined) return undefined;
|
||||
const seen = new Set();
|
||||
for (let depth = 0; depth < 16; depth++) {
|
||||
if (seen.has(linkId)) return null;
|
||||
seen.add(linkId);
|
||||
const link = graph?.links?.[linkId] ?? graph?._links?.get?.(linkId);
|
||||
const source = graph?.getNodeById?.(link?.origin_id);
|
||||
if (!link || seen.has(link)) return null;
|
||||
seen.add(link);
|
||||
const source = graph?.getNodeById?.(link.origin_id);
|
||||
if (!source || (source.mode ?? 0) !== 0) return null;
|
||||
if (source.type === "Reroute") {
|
||||
linkId = source.inputs?.[0]?.link;
|
||||
if (linkId == null) return null;
|
||||
if ((source.comfyClass ?? source.type) === "Reroute") {
|
||||
link = linkInto(source, 0, graph) ?? null;
|
||||
continue;
|
||||
}
|
||||
if (!["PrimitiveFloat", "PrimitiveInt", "AUSBOSS_NODES_Integer", "AUSBOSS_NODES_Float"].includes(source.type) || link.origin_slot !== 0) return null;
|
||||
if (source.inputs?.some((slot) => slot.name === "value" && slot.link != null)) return null;
|
||||
const value = source.widgets?.find((widget) => widget.name === "value")?.value;
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : null;
|
||||
const inner = source.subgraph;
|
||||
if (inner && typeof source.isSubgraphNode === "function" && source.isSubgraphNode()) {
|
||||
link = tableLink(inner, inner.outputs?.[link.origin_slot]?.linkIds?.[0]);
|
||||
graph = inner;
|
||||
continue;
|
||||
}
|
||||
return { node: source, slot: link.origin_slot };
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
const LITERALS = ["PrimitiveFloat", "PrimitiveInt", "AUSBOSS_NODES_Integer", "AUSBOSS_NODES_Float"];
|
||||
|
||||
export function inputNumber(node, name, fallback) {
|
||||
const origin = linkOrigin(node, name);
|
||||
if (origin === undefined) return fallback;
|
||||
if (!origin) return null;
|
||||
const { node: source, slot } = origin;
|
||||
// ComfyUI's older Primitive node holds the value of the widget it feeds.
|
||||
if ((source.comfyClass ?? source.type) === "PrimitiveNode") {
|
||||
const value = source.widgets?.[0]?.value;
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : null;
|
||||
}
|
||||
if (!LITERALS.includes(source.type) || slot !== 0) return null;
|
||||
if (source.inputs?.some((input) => input.name === "value" && input.link != null)) return null;
|
||||
const value = source.widgets?.find((widget) => widget.name === "value")?.value;
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : null;
|
||||
}
|
||||
|
||||
export function clipOutputRate(node, sourceFps, everyNth = 1) {
|
||||
const forced = inputNumber(node, "force_rate", 0);
|
||||
const nth = inputNumber(node, "every_nth", everyNth);
|
||||
|
||||
@@ -0,0 +1,24 @@
|
||||
// A click on the dark area around a modal <dialog> closes it, the way people
|
||||
// expect any popup to work. The browser reports a click on a dialog's
|
||||
// backdrop as a click on the dialog element itself, so a click counts as
|
||||
// outside when it lands on the dialog element beyond its own box. Both the
|
||||
// press and the release must be outside: a text selection or a drag that
|
||||
// starts on the card and ends on the backdrop keeps the popup open.
|
||||
|
||||
export function pointOutside(rect, x, y) {
|
||||
return x < rect.left || x > rect.right || y < rect.top || y > rect.bottom;
|
||||
}
|
||||
|
||||
export function closeOnBackdropClick(dialog, close = () => dialog.close()) {
|
||||
const outside = (event) => event.target === dialog
|
||||
&& pointOutside(dialog.getBoundingClientRect(), event.clientX, event.clientY);
|
||||
let pressedOutside = false;
|
||||
dialog.addEventListener("pointerdown", (event) => {
|
||||
pressedOutside = outside(event);
|
||||
});
|
||||
dialog.addEventListener("click", (event) => {
|
||||
const started = pressedOutside;
|
||||
pressedOutside = false;
|
||||
if (started && outside(event)) close();
|
||||
});
|
||||
}
|
||||
+114
-11
@@ -4,6 +4,7 @@
|
||||
// source of truth: a chip only changes their value.
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { BRAND, chainCallback } from "./index.mjs";
|
||||
import { THEME as T, FONT_UI } from "./theme.mjs";
|
||||
import { liftSocket } from "./widget_card.mjs";
|
||||
import { commitWidgetValue, comboValues } from "./widget_card_math.mjs";
|
||||
|
||||
@@ -21,9 +22,15 @@ function ensureCss() {
|
||||
.ausboss-chip img{flex:none;width:18px;height:18px;margin-left:-5px;border-radius:9px;object-fit:cover}
|
||||
.ausboss-chip b{flex:none;color:${BRAND};font:700 10px ui-monospace,SFMono-Regular,Menlo,monospace}
|
||||
.ausboss-chip em{flex:none;color:#6f8886;font-style:normal;font-weight:500}
|
||||
.ausboss-chip select{field-sizing:content;min-width:0;max-width:120px;height:22px;padding:0 11px 0 0;border:none;background:transparent url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='8' height='5' viewBox='0 0 8 5'%3E%3Cpath d='M1 1l3 3 3-3' fill='none' stroke='%236f8886' stroke-width='1.4'/%3E%3C/svg%3E") no-repeat right 1px center;color:#d8ecea;font:inherit;appearance:none;-webkit-appearance:none;cursor:pointer;text-overflow:ellipsis}
|
||||
.ausboss-chip select:focus-visible{outline:2px solid ${BRAND};outline-offset:2px}
|
||||
.ausboss-chip select option{background:#0f1516;color:#d8ecea}
|
||||
.ausboss-chip-pick{min-width:0;max-width:120px;height:22px;padding:0 12px 0 0;border:none;background:transparent url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='8' height='5' viewBox='0 0 8 5'%3E%3Cpath d='M1 1l3 3 3-3' fill='none' stroke='%236f8886' stroke-width='1.4'/%3E%3C/svg%3E") no-repeat right 1px center;color:#d8ecea;font:inherit;cursor:pointer;overflow:hidden;white-space:nowrap;text-overflow:ellipsis;text-align:left}
|
||||
.ausboss-chip-pick:focus-visible{outline:2px solid ${BRAND};outline-offset:2px;border-radius:3px}
|
||||
.ausboss-chip.open{border-color:${BRAND}}
|
||||
.ausboss-chip-menu{position:fixed;z-index:10000;box-sizing:border-box;padding:.33em;border:1px solid ${T.rule};border-radius:.66em;background:${T.card};box-shadow:0 10px 32px rgba(0,0,0,.55);font:500 12px/1.25 ${FONT_UI};color:${T.inkSoft}}
|
||||
.ausboss-chip-item{display:flex;align-items:center;gap:.5em;padding:.45em .75em .45em .5em;border-radius:.4em;white-space:nowrap;cursor:pointer;user-select:none}
|
||||
.ausboss-chip-item::before{content:"";flex:none;width:.9em;text-align:center;color:${T.brand}}
|
||||
.ausboss-chip-item.current::before{content:"\\2713"}
|
||||
.ausboss-chip-item.current{color:${T.brandInk}}
|
||||
.ausboss-chip-item.hot{background:${T.fill};color:${T.ink}}
|
||||
.ausboss-chip.off{border-style:dashed;background:transparent;color:#56706d}
|
||||
.ausboss-chip.off:hover{border-color:#2a3437}
|
||||
.ausboss-chip.wired{color:#8ba3a1}
|
||||
@@ -56,17 +63,109 @@ function add(parent, tag, text, className) {
|
||||
return child;
|
||||
}
|
||||
|
||||
// The open job list, one at a time across every chip on the page.
|
||||
let openMenu = null;
|
||||
|
||||
// The job list a chip opens. A native <select> lets the browser draw its own
|
||||
// list (small system font, light highlight, pinned to the select's left
|
||||
// edge); this one wears the node's colours, sits under the whole chip and
|
||||
// grows with the canvas zoom like the chip does.
|
||||
function openJobMenu(chipElement, button, values, current, choose) {
|
||||
openMenu?.close();
|
||||
const menu = add(document.body, "div", undefined, "ausboss-chip-menu");
|
||||
menu.setAttribute("role", "listbox");
|
||||
const items = values.map((value) => {
|
||||
const item = add(menu, "div", value, "ausboss-chip-item");
|
||||
item.setAttribute("role", "option");
|
||||
if (value === current) { item.classList.add("current"); item.setAttribute("aria-selected", "true"); }
|
||||
return item;
|
||||
});
|
||||
let hot = -1;
|
||||
const setHot = (index) => {
|
||||
items[hot]?.classList.remove("hot");
|
||||
hot = (index + items.length) % items.length;
|
||||
items[hot]?.classList.add("hot");
|
||||
items[hot]?.scrollIntoView({ block: "nearest" });
|
||||
};
|
||||
setHot(Math.max(0, values.indexOf(current)));
|
||||
|
||||
// The chip is drawn scaled with the canvas: the list matches that size,
|
||||
// starts at the chip's left edge and is at least as wide as the chip.
|
||||
function place() {
|
||||
const box = chipElement.getBoundingClientRect();
|
||||
const zoom = Math.min(2.5, Math.max(0.85, box.height / (chipElement.offsetHeight || 24)));
|
||||
menu.style.fontSize = `${12 * zoom}px`;
|
||||
menu.style.minWidth = `${box.width}px`;
|
||||
const width = menu.offsetWidth;
|
||||
const height = menu.offsetHeight;
|
||||
let top = box.bottom + 4;
|
||||
if (top + height > window.innerHeight - 8 && box.top - height - 4 >= 8) top = box.top - height - 4;
|
||||
menu.style.left = `${Math.max(8, Math.min(box.left, window.innerWidth - width - 8))}px`;
|
||||
menu.style.top = `${Math.max(8, top)}px`;
|
||||
}
|
||||
|
||||
const abort = new AbortController();
|
||||
const close = () => {
|
||||
abort.abort();
|
||||
menu.remove();
|
||||
chipElement.classList.remove("open");
|
||||
if (openMenu?.menu === menu) openMenu = null;
|
||||
};
|
||||
const pick = (index) => { close(); button.focus(); choose(values[index]); };
|
||||
items.forEach((item, index) => {
|
||||
item.addEventListener("pointermove", () => { if (hot !== index) setHot(index); });
|
||||
item.addEventListener("click", () => pick(index));
|
||||
});
|
||||
menu.addEventListener("pointerdown", (event) => event.stopPropagation());
|
||||
const outside = (event) => !menu.contains(event.target) && !button.contains(event.target);
|
||||
const on = { capture: true, signal: abort.signal };
|
||||
window.addEventListener("pointerdown", (event) => { if (outside(event)) close(); }, on);
|
||||
window.addEventListener("wheel", (event) => { if (outside(event)) close(); }, { ...on, passive: true });
|
||||
// The page fires resize on its own now and then: the list follows the chip
|
||||
// instead of vanishing, and goes only when the chip itself is gone.
|
||||
window.addEventListener("resize", () => (chipElement.isConnected ? place() : close()), on);
|
||||
// Only the window itself losing focus (another app, another tab). Pressing an item takes
|
||||
// focus off the chip's button, and a capturing blur listener would hear that and close
|
||||
// the list before the click lands. The press keeps focus on the button instead.
|
||||
window.addEventListener("blur", (event) => { if (event.target === window) close(); }, { signal: abort.signal });
|
||||
menu.addEventListener("mousedown", (event) => event.preventDefault());
|
||||
window.addEventListener("keydown", (event) => {
|
||||
const step = { ArrowDown: 1, ArrowUp: -1 }[event.key];
|
||||
if (step) setHot(hot + step);
|
||||
else if (event.key === "Home") setHot(0);
|
||||
else if (event.key === "End") setHot(items.length - 1);
|
||||
else if (event.key === "Enter" || event.key === " ") pick(hot);
|
||||
else if (event.key === "Escape" || event.key === "Tab") { close(); if (event.key === "Escape") button.focus(); }
|
||||
else return;
|
||||
if (event.key !== "Tab") { event.preventDefault(); event.stopPropagation(); }
|
||||
}, on);
|
||||
|
||||
place();
|
||||
chipElement.classList.add("open");
|
||||
openMenu = { owner: button, menu, close };
|
||||
}
|
||||
|
||||
function choice(node, name, into) {
|
||||
const widget = widgetOf(node, name);
|
||||
const select = add(into, "select");
|
||||
for (const value of comboValues(widget)) add(select, "option", value).value = value;
|
||||
select.value = widget?.value ?? "";
|
||||
select.title = widget?.options?.tooltip ?? "";
|
||||
select.addEventListener("change", () => {
|
||||
if (widget) commitWidgetValue(node, widget, select.value, app.canvas);
|
||||
const button = add(into, "button", undefined, "ausboss-chip-pick");
|
||||
button.type = "button";
|
||||
button.setAttribute("aria-haspopup", "listbox");
|
||||
button.title = widget?.options?.tooltip ?? "";
|
||||
const show = (value) => { button.textContent = String(value ?? ""); };
|
||||
show(widget?.value);
|
||||
const choose = (value) => {
|
||||
show(value);
|
||||
if (widget && value !== widget.value) commitWidgetValue(node, widget, value, app.canvas);
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
};
|
||||
button.addEventListener("pointerdown", (event) => event.stopPropagation());
|
||||
button.addEventListener("click", (event) => {
|
||||
event.preventDefault();
|
||||
if (openMenu?.owner === button) { openMenu.close(); return; }
|
||||
const values = comboValues(widget).map(String);
|
||||
openJobMenu(into, button, values, String(widget?.value ?? ""), choose);
|
||||
});
|
||||
return (values) => { if (document.activeElement !== select) select.value = values[name] ?? ""; };
|
||||
return (values) => { if (openMenu?.owner !== button) show(values[name]); };
|
||||
}
|
||||
|
||||
// The jobs have no socket of their own, but a page saved by an earlier
|
||||
@@ -90,6 +189,7 @@ export function castChips(node) {
|
||||
let syncs = [];
|
||||
let latest = {};
|
||||
const build = (values) => {
|
||||
if (openMenu && element.contains(openMenu.owner)) openMenu.close();
|
||||
element.textContent = "";
|
||||
syncs = [];
|
||||
for (const number of [1, 2, 3, 4]) {
|
||||
@@ -136,7 +236,10 @@ export function castChips(node) {
|
||||
for (const run of syncs) run(values);
|
||||
};
|
||||
const watch = setInterval(() => sync(latest), WATCH);
|
||||
chainCallback(node, "onRemoved", () => clearInterval(watch));
|
||||
chainCallback(node, "onRemoved", () => {
|
||||
clearInterval(watch);
|
||||
if (openMenu && element.contains(openMenu.owner)) openMenu.close();
|
||||
});
|
||||
chainCallback(node, "onConfigure", () => queueMicrotask(() => keepOldJobLinks(node)));
|
||||
return { element, sync };
|
||||
}
|
||||
|
||||
@@ -1,3 +1,5 @@
|
||||
import { linkOrigin } from "./clip_rate.mjs";
|
||||
|
||||
// The same live-input rule used by the writer timeline: muted/bypassed
|
||||
// sources do not reach this node. This inspects the wire, never media bytes.
|
||||
export function linkSource(node, name) {
|
||||
@@ -9,8 +11,45 @@ export function linkSource(node, name) {
|
||||
return (source?.mode ?? 0) === 0 ? source ?? null : null;
|
||||
}
|
||||
|
||||
// Keep in step with Python video_frames (node_h3_prompt_writer.py): H3 makes
|
||||
// 5 frames plus steps of 17, at most 15 s at 24 fps, rounding a clip down.
|
||||
export const H3_FPS = 24;
|
||||
export function h3VideoFrames(frames) {
|
||||
const count = Math.trunc(Number(frames));
|
||||
if (!(count > 0)) return null;
|
||||
return Math.max(1, Math.floor((Math.min(count, 15 * H3_FPS) - 5) / 17)) * 17 + 5;
|
||||
}
|
||||
|
||||
// What the clip node's face last worked out for its run ({frames,
|
||||
// startSeconds, endSeconds}), or null when the clip is not that node or its
|
||||
// count is only known at run time.
|
||||
export function clipTimingFor(node) {
|
||||
// Through Reroutes and subgraphs, as every linked value is read.
|
||||
const timing = linkOrigin(node, "video")?.node?.__ausbossClipTiming;
|
||||
return timing && timing.frames > 0 ? timing : null;
|
||||
}
|
||||
|
||||
// The words under "From the clip": what H3 will make of the clip's frames.
|
||||
export function clipLengthWords(timing) {
|
||||
if (!timing) return null;
|
||||
const made = h3VideoFrames(timing.frames);
|
||||
const seconds = (frames) => String(Number((frames / H3_FPS).toFixed(2)));
|
||||
if (made < timing.frames) {
|
||||
const up = made + 17;
|
||||
const room = timing.roomFrames ?? Infinity;
|
||||
const stops = timing.startSeconds + made / H3_FPS;
|
||||
return {
|
||||
title: `H3 makes ${made} of the clip's ${timing.frames} frames · ${seconds(made)} s`,
|
||||
note: `Stops at ${String(Number(stops.toFixed(2)))} s in the video. To use all of it, set the clip's Duration to ${seconds(made)} s${up <= Math.min(room, 15 * H3_FPS) ? ` or ${seconds(up)} s` : ""}.`,
|
||||
warn: true,
|
||||
};
|
||||
}
|
||||
if (made === timing.frames) return { title: `H3 uses all ${made} frames · ${seconds(made)} s`, note: "The clip is already a length H3 makes.", warn: false };
|
||||
return { title: `H3 makes ${made} frames · ${seconds(made)} s`, note: `That is its shortest length; the clip has ${timing.frames} frames.`, warn: true };
|
||||
}
|
||||
|
||||
// Keep in step with Python KEEPS_ITS_LENGTH. Other jobs use seconds normally.
|
||||
export function clipSetsDuration(node, values) {
|
||||
return ["swap the person", "new voice", "change something"].includes(values.video_job)
|
||||
return ["copy its moves", "swap the person", "new voice", "change something"].includes(values.video_job)
|
||||
&& linkSource(node, "video") !== null;
|
||||
}
|
||||
|
||||
+1
-1
@@ -9,7 +9,7 @@ import { app } from "../../../scripts/app.js";
|
||||
// Must match `version` in pyproject.toml — scripts/release_preflight.py
|
||||
// enforces it. The backend serves its copy at /ausboss/pack_version; a
|
||||
// mismatch means the browser cached JavaScript from an older install.
|
||||
export const AUSBOSS_JS_VERSION = "2.23.0";
|
||||
export const AUSBOSS_JS_VERSION = "2.26.3";
|
||||
|
||||
export const BRAND = "#00b4aa"; // AusBoss teal — keep node accents consistent.
|
||||
export const BRAND_DARK = "#007f78";
|
||||
|
||||
@@ -107,16 +107,59 @@ export function outputRecordQuery(record, index = null) {
|
||||
}).toString();
|
||||
}
|
||||
|
||||
// The /view query of the small copy Save Image writes for its face
|
||||
// ({ ausboss_preview: ["ausboss_preview/<name>.webp"] }, one for each entry
|
||||
// of images, in the temp folder). Null when the record has none: an older
|
||||
// run, another node, or the preview switched off.
|
||||
export function lightPreviewQuery(record, index = null) {
|
||||
const places = record?.ausboss_preview;
|
||||
const count = outputRecordCount(record);
|
||||
if (!count || !Array.isArray(places) || places.length !== count) return null;
|
||||
const at = Number.isInteger(index) && index >= 0 && index < count ? index : count - 1;
|
||||
const place = places[at];
|
||||
if (typeof place !== "string" || !place) return null;
|
||||
const slash = place.lastIndexOf("/");
|
||||
return new URLSearchParams({
|
||||
filename: place.slice(slash + 1),
|
||||
subfolder: slash >= 0 ? place.slice(0, slash) : "",
|
||||
type: "temp",
|
||||
}).toString();
|
||||
}
|
||||
|
||||
// How many pictures the node's last run left in its output record.
|
||||
export function outputRecordCount(record) {
|
||||
return Array.isArray(record?.images) ? record.images.length : 0;
|
||||
}
|
||||
|
||||
// Is node.imgs the picture of this output record?
|
||||
//
|
||||
// The frontend loads a result into node.imgs with a time limit (two tries of
|
||||
// about 8 seconds). A picture that takes longer, a big one over a slow link,
|
||||
// is dropped without a word and node.imgs keeps the run before. `loadedFor`
|
||||
// is the record that was current when node.imgs was last set; the file name
|
||||
// is compared as well, for a newer run that arrived while a picture loaded.
|
||||
export function ownResultIsCurrent(node, record, loadedFor) {
|
||||
const images = node?.imgs;
|
||||
const src = Array.isArray(images) && images.length ? images[images.length - 1]?.src : null;
|
||||
if (typeof src !== "string" || !src) return false;
|
||||
const wanted = outputRecordQuery(record);
|
||||
// No stored picture to compare with, or a frame streamed during the run.
|
||||
if (!wanted || !/[?&]filename=/.test(src)) return true;
|
||||
if (loadedFor !== record) return false;
|
||||
const have = new URLSearchParams(src.slice(src.indexOf("?") + 1));
|
||||
const want = new URLSearchParams(wanted);
|
||||
return ["filename", "subfolder", "type"].every((key) => (have.get(key) ?? "") === (want.get(key) ?? ""));
|
||||
}
|
||||
|
||||
// What the panel shows, best first: this node's result (node.imgs, then the
|
||||
// URL of its stored output), then whatever feeds it, then nothing.
|
||||
export function describeNodePreview(node, inputName, storedUrl = null) {
|
||||
return describeOwnResult(node)
|
||||
?? (storedUrl ? { kind: "url", url: storedUrl } : null)
|
||||
// `ownIsCurrent` false: node.imgs is an older run's (ownResultIsCurrent), so
|
||||
// the stored output goes first.
|
||||
export function describeNodePreview(node, inputName, storedUrl = null, ownIsCurrent = true) {
|
||||
const stored = storedUrl ? { kind: "url", url: storedUrl } : null;
|
||||
return (ownIsCurrent ? null : stored)
|
||||
?? describeOwnResult(node)
|
||||
?? stored
|
||||
?? describeSourcePreview(upstreamNode(node, inputName));
|
||||
}
|
||||
|
||||
|
||||
@@ -13,6 +13,7 @@
|
||||
// what it always was. The turn is made in canvas pixels, so a turned box is a
|
||||
// true rectangle on a canvas of any shape (nodes/_layout_boxes_helpers.py
|
||||
// turns it the same way).
|
||||
import { linkOrigin } from "./clip_rate.mjs";
|
||||
|
||||
export const DIM_MIN = 64;
|
||||
export const DIM_MAX = 16384;
|
||||
@@ -683,8 +684,10 @@ const isNumber = (value) => typeof value === "number" && Number.isFinite(value);
|
||||
// (Integer, Primitive). Reroutes are followed. Null when the input is not
|
||||
// wired or the value is only known once the workflow runs.
|
||||
export function wiredNumber(node, inputName) {
|
||||
const source = wiredSource(node, inputName);
|
||||
return source ? wiredFrom(source.node, source.slot) : null;
|
||||
// Through Reroutes and into subgraphs (js/shared/clip_rate.mjs), then
|
||||
// this node's own rule for which numbers count.
|
||||
const origin = linkOrigin(node, inputName);
|
||||
return origin ? wiredFrom(origin.node, origin.slot) : null;
|
||||
}
|
||||
|
||||
// The node (and its output slot) an input is wired to, through Reroutes.
|
||||
|
||||
@@ -0,0 +1,211 @@
|
||||
// The row under the file picker on Load Image: arrows that step through the
|
||||
// input folder, where the file is in the list ("14 / 212"), and the size the
|
||||
// picture goes in at and comes out at ("912 × 1632 → 1144 × 2048"). The
|
||||
// widgets underneath stay the source of truth: an arrow only changes the
|
||||
// image widget's value, the way picking from the list does.
|
||||
//
|
||||
// Under it, the preview: the picture as it goes in with the part that is
|
||||
// kept lit, and the picture as it comes out, both on one scale, with what
|
||||
// the size choice does in words. While the choice cuts something, a click
|
||||
// on the first picture picks the part to keep.
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { CROP_MODES, changeText, cropSideAt, keptPart, placeText, planLoadSize, shapeText, sizeText, stepFile, thumbSizes } from "./load_size.mjs";
|
||||
import { mediaViewQuery } from "./media_list.mjs";
|
||||
import { FONT_UI, RADIUS_CONTROL, THEME as T } from "./theme.mjs";
|
||||
import { commitWidgetValue, comboValues } from "./widget_card_math.mjs";
|
||||
|
||||
const CSS_ID = "ausboss-load-strip-css";
|
||||
// Sizes already read, by file: stepping back to a picture shows its size at once.
|
||||
const SIZES_KEPT = 200;
|
||||
const sizes = new Map();
|
||||
|
||||
function ensureCss() {
|
||||
if (document.getElementById(CSS_ID)) return;
|
||||
const style = document.createElement("style");
|
||||
style.id = CSS_ID;
|
||||
style.textContent = `
|
||||
.ausboss-load-strip{display:flex;align-items:center;gap:6px;height:100%;min-width:0;font:500 11px/1 ${FONT_UI};color:${T.label}}
|
||||
.ausboss-load-step{flex:none;width:26px;height:100%;padding:0;border:1px solid transparent;border-radius:${RADIUS_CONTROL}px;background:${T.fill};color:${T.inkSoft};font:600 11px/1 ${FONT_UI};cursor:pointer}
|
||||
.ausboss-load-step:hover{background:${T.fillHover};color:${T.inkHover}}
|
||||
.ausboss-load-step:focus-visible{outline:2px solid ${T.brand};outline-offset:1px}
|
||||
.ausboss-load-step:disabled{opacity:.4;cursor:default}
|
||||
.ausboss-load-place{flex:none;color:${T.faint};font-variant-numeric:tabular-nums;white-space:nowrap;user-select:none}
|
||||
.ausboss-load-size{flex:1 1 0;min-width:0;overflow:hidden;text-align:right;white-space:nowrap;text-overflow:ellipsis;color:${T.inkSoft};font-weight:600;font-variant-numeric:tabular-nums}
|
||||
.ausboss-load-size b{color:${T.brandInk};font-weight:700}
|
||||
.ausboss-load-size i{color:${T.faint};font-style:normal;margin:0 5px}
|
||||
.ausboss-load-view{display:flex;align-items:center;justify-content:center;gap:10px;height:100%;min-width:0;font:500 10px/1.25 ${FONT_UI};color:${T.faint}}
|
||||
.ausboss-load-thumb{position:relative;flex:none;overflow:hidden;border-radius:2px;background-color:${T.fill};background-repeat:no-repeat}
|
||||
.ausboss-load-thumb.in{background-size:100% 100%}
|
||||
.ausboss-load-thumb.in.pick{cursor:pointer}
|
||||
.ausboss-load-thumb.out{outline:1px solid ${T.brand};outline-offset:0}
|
||||
.ausboss-load-kept{position:absolute;box-sizing:border-box;border:1px solid ${T.brand};box-shadow:0 0 0 400px rgba(8,10,11,.68);pointer-events:none}
|
||||
.ausboss-load-note{flex:0 1 auto;min-width:0;display:flex;flex-direction:column;align-items:center;gap:2px;text-align:center;user-select:none}
|
||||
.ausboss-load-note b{color:${T.inkSoft};font-weight:600;font-variant-numeric:tabular-nums;white-space:nowrap}
|
||||
.ausboss-load-note span{white-space:nowrap}
|
||||
`;
|
||||
document.head.append(style);
|
||||
}
|
||||
|
||||
const viewUrl = (value) => api.apiURL(`/view?${mediaViewQuery(value)}`);
|
||||
|
||||
// The picture's own size, read from the file the way the browser shows it.
|
||||
function readSize(file, done) {
|
||||
const known = sizes.get(file);
|
||||
if (known) return known;
|
||||
const picture = new Image();
|
||||
picture.onload = () => {
|
||||
if (sizes.size >= SIZES_KEPT) sizes.delete(sizes.keys().next().value);
|
||||
sizes.set(file, { width: picture.naturalWidth, height: picture.naturalHeight });
|
||||
done();
|
||||
};
|
||||
picture.src = viewUrl(file);
|
||||
return null;
|
||||
}
|
||||
|
||||
// The widest and tallest a preview picture is drawn.
|
||||
const THUMB_WIDTH = 76;
|
||||
const THUMB_HEIGHT = 68;
|
||||
|
||||
export function loadImagePreview(node) {
|
||||
ensureCss();
|
||||
const root = document.createElement("div");
|
||||
root.className = "ausboss-load-view";
|
||||
const input = document.createElement("div");
|
||||
input.className = "ausboss-load-thumb in";
|
||||
const kept = document.createElement("div");
|
||||
kept.className = "ausboss-load-kept";
|
||||
input.append(kept);
|
||||
const note = document.createElement("div");
|
||||
note.className = "ausboss-load-note";
|
||||
const output = document.createElement("div");
|
||||
output.className = "ausboss-load-thumb out";
|
||||
root.append(input, note, output);
|
||||
let last = {};
|
||||
|
||||
const side = () => node.widgets?.find((item) => item.name === "crop_from") ?? null;
|
||||
const sideLinked = () => node.inputs?.find((slot) => slot?.widget?.name === "crop_from")?.link != null;
|
||||
input.addEventListener("click", (event) => {
|
||||
const widget = side();
|
||||
if (!widget || !input.classList.contains("pick")) return;
|
||||
const box = input.getBoundingClientRect();
|
||||
if (!(box.width > 0 && box.height > 0)) return;
|
||||
const next = cropSideAt((event.clientX - box.left) / box.width, (event.clientY - box.top) / box.height);
|
||||
if (next === widget.value) return;
|
||||
commitWidgetValue(node, widget, next, app.canvas);
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
});
|
||||
|
||||
const line = (tag, text) => {
|
||||
const item = document.createElement(tag);
|
||||
item.textContent = text;
|
||||
return item;
|
||||
};
|
||||
const sync = (values = last) => {
|
||||
last = values;
|
||||
const file = String(values.image ?? "");
|
||||
const source = file ? readSize(file, () => { if (String(last.image ?? "") === file) sync(); }) : null;
|
||||
const plan = source ? planLoadSize(source.width, source.height, values) : null;
|
||||
root.style.visibility = plan ? "" : "hidden";
|
||||
if (!plan) return;
|
||||
const picture = `url(${JSON.stringify(viewUrl(file))})`;
|
||||
const sizes = thumbSizes(source.width, source.height, plan, THUMB_WIDTH, THUMB_HEIGHT);
|
||||
const part = keptPart(plan);
|
||||
Object.assign(input.style, { width: `${sizes.input[0]}px`, height: `${sizes.input[1]}px`, backgroundImage: picture });
|
||||
Object.assign(kept.style, {
|
||||
left: `${part.x * 100}%`, top: `${part.y * 100}%`,
|
||||
width: `${part.width * 100}%`, height: `${part.height * 100}%`,
|
||||
});
|
||||
const spareWide = plan.resizeWidth - plan.width;
|
||||
const spareHigh = plan.resizeHeight - plan.height;
|
||||
Object.assign(output.style, {
|
||||
width: `${sizes.output[0]}px`, height: `${sizes.output[1]}px`, backgroundImage: picture,
|
||||
backgroundSize: `${(plan.resizeWidth / plan.width) * 100}% ${(plan.resizeHeight / plan.height) * 100}%`,
|
||||
backgroundPosition: `${spareWide > 0 ? (plan.cropX / spareWide) * 100 : 0}% ${spareHigh > 0 ? (plan.cropY / spareHigh) * 100 : 0}%`,
|
||||
});
|
||||
const picks = CROP_MODES.includes(values.size) && !sideLinked();
|
||||
input.classList.toggle("pick", picks);
|
||||
input.title = picks
|
||||
? "The picture going in. The lit part is kept. Click the part you want to keep."
|
||||
: "The picture going in. The lit part is kept.";
|
||||
output.title = "The picture coming out.";
|
||||
const from = shapeText(source.width, source.height);
|
||||
const to = shapeText(plan.width, plan.height);
|
||||
const change = changeText(source.width, source.height, plan);
|
||||
note.replaceChildren(
|
||||
line("b", from === to ? from : `${from} → ${to}`),
|
||||
...(change ? change.split(" · ").map((text) => line("span", text)) : [line("span", "no change")]),
|
||||
);
|
||||
};
|
||||
return { element: root, sync };
|
||||
}
|
||||
|
||||
export function loadImageStrip(node) {
|
||||
ensureCss();
|
||||
const root = document.createElement("div");
|
||||
root.className = "ausboss-load-strip";
|
||||
const widget = () => node.widgets?.find((item) => item.name === "image") ?? null;
|
||||
const linked = () => {
|
||||
const slot = node.inputs?.find((input) => input?.widget?.name === "image" || input?.name === "image");
|
||||
return slot?.link != null;
|
||||
};
|
||||
let last = {};
|
||||
|
||||
const step = (delta, label, title) => {
|
||||
const button = document.createElement("button");
|
||||
button.type = "button";
|
||||
button.className = "ausboss-load-step";
|
||||
button.textContent = label;
|
||||
button.title = title;
|
||||
button.addEventListener("click", () => {
|
||||
const image = widget();
|
||||
if (!image) return;
|
||||
const next = stepFile(comboValues(image), image.value, delta);
|
||||
if (next === null || next === image.value) return;
|
||||
commitWidgetValue(node, image, next, app.canvas);
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
});
|
||||
root.append(button);
|
||||
return button;
|
||||
};
|
||||
const back = step(-1, "◀", "The picture before this one in the input folder");
|
||||
const on = step(1, "▶", "The next picture in the input folder");
|
||||
const place = document.createElement("span");
|
||||
place.className = "ausboss-load-place";
|
||||
place.title = "Where this picture is in the input folder's list";
|
||||
const size = document.createElement("span");
|
||||
size.className = "ausboss-load-size";
|
||||
root.append(place, size);
|
||||
|
||||
const sync = (values = last) => {
|
||||
last = values;
|
||||
const image = widget();
|
||||
const files = comboValues(image);
|
||||
const file = String(values.image ?? "");
|
||||
place.textContent = placeText(files, file);
|
||||
const stuck = linked() || stepFile(files, file, 1) === null;
|
||||
back.disabled = stuck;
|
||||
on.disabled = stuck;
|
||||
const source = file ? readSize(file, () => { if (String(last.image ?? "") === file) sync(); }) : null;
|
||||
size.replaceChildren();
|
||||
if (!source) {
|
||||
size.title = "";
|
||||
return;
|
||||
}
|
||||
const plan = planLoadSize(source.width, source.height, values);
|
||||
const text = sizeText(source.width, source.height, plan);
|
||||
const [from, to] = text.split(" → ");
|
||||
size.append(from);
|
||||
if (to) {
|
||||
const arrow = document.createElement("i");
|
||||
arrow.textContent = "→";
|
||||
const out = document.createElement("b");
|
||||
out.textContent = to;
|
||||
size.append(arrow, out);
|
||||
}
|
||||
size.title = to
|
||||
? `The file is ${from}. The picture comes out at ${to}.`
|
||||
: `The picture comes out at its own size, ${from}.`;
|
||||
};
|
||||
return { element: root, sync };
|
||||
}
|
||||
@@ -0,0 +1,206 @@
|
||||
// Load Image: the size a picture comes out at, worked out in the browser so
|
||||
// the node can show it before a run. This repeats plan_load_size in
|
||||
// nodes/_load_image_helpers.py; tests/fixtures/load_size_cases.json holds
|
||||
// both to the same answers. No DOM in here.
|
||||
|
||||
export const SIZE_MODES = Object.freeze(["off", "megapixels", "longest side", "scale", "fit inside", "crop to fill", "crop", "crop to ratio"]);
|
||||
// Which part of the picture a crop keeps, as a 3 x 3 grid read row by row.
|
||||
export const CROP_GRID = Object.freeze(["top left", "top", "top right", "left", "center", "right", "bottom left", "bottom", "bottom right"]);
|
||||
// The choices that cut something off, so the kept part matters.
|
||||
export const CROP_MODES = Object.freeze(["crop to fill", "crop", "crop to ratio"]);
|
||||
|
||||
const roundHalfUp = (value) => Math.floor(value + 0.5);
|
||||
const whole = (value, fallback) => (Number.isFinite(Number(value)) ? Math.trunc(Number(value)) : fallback);
|
||||
// Nearest multiple, never below one step.
|
||||
const nearest = (value, step) => Math.max(step, roundHalfUp(value / step) * step);
|
||||
// The largest multiple that is not above the value, never below one step.
|
||||
const below = (value, step) => Math.max(step, Math.floor(value / step) * step);
|
||||
// Rounded down to a multiple, or left alone when smaller than one step.
|
||||
const inside = (value, step) => (value >= step ? Math.floor(value / step) * step : value);
|
||||
// Where a cut starts along one edge with `spare` pixels to lose.
|
||||
const cropStart = (spare, side, low, high) => (side.includes(low) ? 0 : side.includes(high) ? spare : Math.floor(spare / 2));
|
||||
|
||||
function ratioTerms(ratio, width, height) {
|
||||
if (ratio === "width : height") return [Math.max(1, width), Math.max(1, height)];
|
||||
const [across, down] = String(ratio ?? "1:1").split(":").map((term) => Math.trunc(Number(term)));
|
||||
return [Math.max(1, across || 1), Math.max(1, down || 1)];
|
||||
}
|
||||
|
||||
function target(sourceW, sourceH, size, { megapixels, longest_side: longestSide, scale, width, height }) {
|
||||
const scaled = (factor) => [Math.max(1, roundHalfUp(sourceW * factor)), Math.max(1, roundHalfUp(sourceH * factor))];
|
||||
if (size === "megapixels") {
|
||||
const amount = Number(megapixels);
|
||||
return amount > 0 ? scaled(Math.sqrt((amount * 1e6) / (sourceW * sourceH))) : [sourceW, sourceH];
|
||||
}
|
||||
if (size === "longest side") {
|
||||
const edge = whole(longestSide, 0);
|
||||
if (edge <= 0) return [sourceW, sourceH];
|
||||
const [w, h] = scaled(edge / Math.max(sourceW, sourceH));
|
||||
return sourceW >= sourceH ? [edge, h] : [w, edge];
|
||||
}
|
||||
if (size === "scale") {
|
||||
const factor = Number(scale);
|
||||
return factor > 0 ? scaled(factor) : [sourceW, sourceH];
|
||||
}
|
||||
const fit = Math.min(Math.max(1, whole(width, 1)) / sourceW, Math.max(1, whole(height, 1)) / sourceH);
|
||||
return scaled(fit);
|
||||
}
|
||||
|
||||
// { resizeWidth, resizeHeight, cropX, cropY, width, height } for a picture of
|
||||
// sourceWidth x sourceHeight and the node's values, or null when the picture
|
||||
// has no size yet or the choice is not one the node knows.
|
||||
export function planLoadSize(sourceWidth, sourceHeight, values = {}) {
|
||||
const sourceW = whole(sourceWidth, 0);
|
||||
const sourceH = whole(sourceHeight, 0);
|
||||
const size = values.size ?? "off";
|
||||
if (sourceW <= 0 || sourceH <= 0 || !SIZE_MODES.includes(size)) return null;
|
||||
const same = { resizeWidth: sourceW, resizeHeight: sourceH, cropX: 0, cropY: 0, width: sourceW, height: sourceH };
|
||||
if (size === "off") return same;
|
||||
const step = Math.max(1, whole(values.multiple, 1));
|
||||
const enlarge = values.enlarge !== false;
|
||||
const boxW = Math.max(1, whole(values.width, 1024));
|
||||
const boxH = Math.max(1, whole(values.height, 1024));
|
||||
const side = CROP_GRID.includes(values.crop_from) ? values.crop_from : "center";
|
||||
const cut = (resizeWidth, resizeHeight, width, height) => ({
|
||||
resizeWidth, resizeHeight,
|
||||
cropX: cropStart(resizeWidth - width, side, "left", "right"),
|
||||
cropY: cropStart(resizeHeight - height, side, "top", "bottom"),
|
||||
width, height,
|
||||
});
|
||||
|
||||
if (size === "crop") {
|
||||
return cut(sourceW, sourceH, inside(Math.min(boxW, sourceW), step), inside(Math.min(boxH, sourceH), step));
|
||||
}
|
||||
|
||||
if (size === "crop to ratio") {
|
||||
const [across, down] = ratioTerms(values.ratio, boxW, boxH);
|
||||
let width = sourceW;
|
||||
let height = sourceH;
|
||||
if (sourceW * down >= sourceH * across) width = Math.min(sourceW, Math.max(1, roundHalfUp((sourceH * across) / down)));
|
||||
else height = Math.min(sourceH, Math.max(1, roundHalfUp((sourceW * down) / across)));
|
||||
return cut(sourceW, sourceH, inside(width, step), inside(height, step));
|
||||
}
|
||||
|
||||
if (size === "crop to fill") {
|
||||
let width = boxW;
|
||||
let height = boxH;
|
||||
if (enlarge) {
|
||||
width = nearest(width, step);
|
||||
height = nearest(height, step);
|
||||
} else {
|
||||
const grow = Math.max(width / sourceW, height / sourceH);
|
||||
if (grow > 1) {
|
||||
width = Math.max(1, Math.floor(width / grow + 1e-6));
|
||||
height = Math.max(1, Math.floor(height / grow + 1e-6));
|
||||
}
|
||||
width = Math.min(below(width, step), Math.max(step, sourceW));
|
||||
height = Math.min(below(height, step), Math.max(step, sourceH));
|
||||
}
|
||||
const cover = Math.max(width / sourceW, height / sourceH);
|
||||
const resizeWidth = Math.max(width, Math.ceil(sourceW * cover - 1e-6));
|
||||
const resizeHeight = Math.max(height, Math.ceil(sourceH * cover - 1e-6));
|
||||
return cut(resizeWidth, resizeHeight, width, height);
|
||||
}
|
||||
|
||||
let [targetW, targetH] = target(sourceW, sourceH, size, { ...values, width: boxW, height: boxH });
|
||||
if (!enlarge && (targetW > sourceW || targetH > sourceH)) [targetW, targetH] = [sourceW, sourceH];
|
||||
const rounded = (wanted, source, box) => {
|
||||
let value = nearest(wanted, step);
|
||||
if (size === "fit inside" && value > box) value = below(box, step);
|
||||
if (!enlarge && value > source) value = below(source, step);
|
||||
return value;
|
||||
};
|
||||
const width = rounded(targetW, sourceW, boxW);
|
||||
const height = rounded(targetH, sourceH, boxH);
|
||||
return { resizeWidth: width, resizeHeight: height, cropX: 0, cropY: 0, width, height };
|
||||
}
|
||||
|
||||
// "912 × 1632" for a picture that comes out as it went in, "912 × 1632 →
|
||||
// 1144 × 2048" when its size changes, "" while the picture's size is unknown.
|
||||
export function sizeText(sourceWidth, sourceHeight, plan) {
|
||||
if (!(sourceWidth > 0 && sourceHeight > 0)) return "";
|
||||
const from = `${sourceWidth} × ${sourceHeight}`;
|
||||
if (!plan || (plan.width === sourceWidth && plan.height === sourceHeight)) return from;
|
||||
return `${from} → ${plan.width} × ${plan.height}`;
|
||||
}
|
||||
|
||||
// What a plan does to the picture, in a few words for the node: how much it
|
||||
// is resized and how much is cut off, in the file's own pixels. "" when the
|
||||
// picture comes out as it went in.
|
||||
export function changeText(sourceWidth, sourceHeight, plan) {
|
||||
if (!plan || !(sourceWidth > 0 && sourceHeight > 0)) return "";
|
||||
const parts = [];
|
||||
const percent = Math.round((plan.resizeWidth / sourceWidth) * 100);
|
||||
if (plan.resizeWidth !== sourceWidth || plan.resizeHeight !== sourceHeight) parts.push(`${percent}% size`);
|
||||
const cutWide = Math.round(((plan.resizeWidth - plan.width) * sourceWidth) / plan.resizeWidth);
|
||||
const cutHigh = Math.round(((plan.resizeHeight - plan.height) * sourceHeight) / plan.resizeHeight);
|
||||
if (cutWide > 0 && cutHigh > 0) parts.push(`cuts ${cutWide} × ${cutHigh} px`);
|
||||
else if (cutWide > 0) parts.push(`cuts ${cutWide} px of width`);
|
||||
else if (cutHigh > 0) parts.push(`cuts ${cutHigh} px of height`);
|
||||
return parts.join(" · ");
|
||||
}
|
||||
|
||||
const SHAPES = [[1, 1], [5, 4], [4, 3], [3, 2], [16, 10], [16, 9], [2, 1], [21, 9]];
|
||||
|
||||
// A picture's shape in words people use: "3:4", "16:9". A size within one
|
||||
// percent of a common shape gets that name (rounding to a multiple moves a
|
||||
// 16:9 picture a few pixels off it); any other gets "~1.34:1" or "~1:1.34".
|
||||
export function shapeText(width, height) {
|
||||
if (!(width > 0 && height > 0)) return "";
|
||||
const wide = width >= height;
|
||||
const value = wide ? width / height : height / width;
|
||||
const near = SHAPES.find(([across, down]) => Math.abs(value - across / down) <= (across / down) * 0.01);
|
||||
if (near) return wide ? `${near[0]}:${near[1]}` : `${near[1]}:${near[0]}`;
|
||||
const rounded = Math.round(value * 100) / 100;
|
||||
return wide ? `~${rounded}:1` : `~1:${rounded}`;
|
||||
}
|
||||
|
||||
// The part of the file a plan keeps, as fractions of the file's width and
|
||||
// height: { x, y, width, height }, all 0 to 1.
|
||||
export function keptPart(plan) {
|
||||
if (!plan) return { x: 0, y: 0, width: 1, height: 1 };
|
||||
return {
|
||||
x: plan.cropX / plan.resizeWidth,
|
||||
y: plan.cropY / plan.resizeHeight,
|
||||
width: Math.min(1, plan.width / plan.resizeWidth),
|
||||
height: Math.min(1, plan.height / plan.resizeHeight),
|
||||
};
|
||||
}
|
||||
|
||||
// The sizes two small pictures are drawn at so their difference shows: both
|
||||
// on one scale, the larger filling a box of maxWidth x maxHeight, and neither
|
||||
// edge under `least` pixels.
|
||||
export function thumbSizes(sourceWidth, sourceHeight, plan, maxWidth, maxHeight, least = 10) {
|
||||
const outW = plan?.width ?? sourceWidth;
|
||||
const outH = plan?.height ?? sourceHeight;
|
||||
const scale = Math.min(maxWidth / Math.max(sourceWidth, outW), maxHeight / Math.max(sourceHeight, outH));
|
||||
const at = (value) => Math.max(least, Math.round(value * scale));
|
||||
return { input: [at(sourceWidth), at(sourceHeight)], output: [at(outW), at(outH)] };
|
||||
}
|
||||
|
||||
// The crop side under a click at (x, y), each 0 to 1 across the picture.
|
||||
export function cropSideAt(x, y) {
|
||||
const third = (value) => (value < 1 / 3 ? 0 : value < 2 / 3 ? 1 : 2);
|
||||
return CROP_GRID[third(y) * 3 + third(x)];
|
||||
}
|
||||
|
||||
// The file one step before or after the current one in the list, going
|
||||
// round at the ends. A file that is not in the list steps to the first or
|
||||
// the last. null when the list has nothing to step to.
|
||||
export function stepFile(files, current, delta) {
|
||||
const list = (Array.isArray(files) ? files : []).filter((name) => name !== "" && name != null);
|
||||
if (!list.length) return null;
|
||||
const at = list.indexOf(current);
|
||||
if (at < 0) return delta < 0 ? list[list.length - 1] : list[0];
|
||||
if (list.length === 1) return null;
|
||||
return list[(at + (delta < 0 ? -1 : 1) + list.length) % list.length];
|
||||
}
|
||||
|
||||
// "14 / 212": where the current file is in the list. "" for an empty list,
|
||||
// "– / 212" for a file that is not in it.
|
||||
export function placeText(files, current) {
|
||||
const list = (Array.isArray(files) ? files : []).filter((name) => name !== "" && name != null);
|
||||
if (!list.length) return "";
|
||||
const at = list.indexOf(current);
|
||||
return `${at < 0 ? "–" : at + 1} / ${list.length}`;
|
||||
}
|
||||
@@ -0,0 +1,87 @@
|
||||
// ComfyUI's missing-models list (the Issues panel, "Download all") only
|
||||
// looks at list widgets: a model is listed when a list on some node holds a
|
||||
// file name that is not one of its choices. LoRA Loader keeps its files in
|
||||
// the rows' text, so a LoRA a workflow needs never made that list, and
|
||||
// "Download all" then Run stopped on "missing LoRA file".
|
||||
//
|
||||
// The loader therefore carries one hidden list widget per row that needs a
|
||||
// file. Nothing else reads them: they are kept out of the saved workflow and
|
||||
// the prompt, have no socket, and are made again from the rows every time.
|
||||
// The rows stay the one source of truth.
|
||||
|
||||
import { resolveLoraName } from "./lora_stack.mjs";
|
||||
|
||||
export const FILE_WIDGET_PREFIX = "ausboss_lora_file_";
|
||||
export const FILE_WIDGET_FLAG = "ausbossLoraFile";
|
||||
|
||||
// Rows the server needs a file for when it runs: named, on, in a group that
|
||||
// is on, and not at zero strength (nodes/_lora_helpers.py, missing_lora_rows).
|
||||
export function rowsNeedingFiles(rows) {
|
||||
return (Array.isArray(rows) ? rows : []).filter((row) => {
|
||||
const name = typeof row?.name === "string" ? row.name.trim() : "";
|
||||
if (!name || row.enabled === false || row.group_off === true) return false;
|
||||
const strength = Number(row.strength ?? 1);
|
||||
const clip = Number(row.strength_clip ?? strength);
|
||||
return !(strength === 0 && clip === 0);
|
||||
});
|
||||
}
|
||||
|
||||
// The list ComfyUI checks a row's file against. A file the loader would
|
||||
// find anyway (saved in another folder, or with other capitals) counts as
|
||||
// installed; so does every file while the install's list is unknown, so a
|
||||
// slow list never shows a false alarm.
|
||||
export function fileChoices(name, installed) {
|
||||
if (!Array.isArray(installed)) return [name];
|
||||
return resolveLoraName(name, installed).status === "missing" ? installed : [...installed, name];
|
||||
}
|
||||
|
||||
// The install's LoRA files as ComfyUI's own LoRA loader lists them. The node
|
||||
// list arrives before any workflow opens, so this answers at once; null when
|
||||
// there is no such node to ask.
|
||||
export function coreLoraList(definition) {
|
||||
const spec = definition?.input?.required?.lora_name;
|
||||
if (!Array.isArray(spec)) return null;
|
||||
if (Array.isArray(spec[0])) return spec[0];
|
||||
if (spec[0] === "COMBO" && Array.isArray(spec[1]?.options)) return spec[1].options;
|
||||
return null;
|
||||
}
|
||||
|
||||
export function isFileWidget(widget) {
|
||||
return Boolean(widget?.[FILE_WIDGET_FLAG]);
|
||||
}
|
||||
|
||||
// Bring the node's file widgets in line with its rows. `installed()` returns
|
||||
// the install's files (or null), `hide(widget)` folds a widget away, and
|
||||
// `onPick(rowId, name)` runs when ComfyUI puts another file in a widget
|
||||
// (choosing a replacement in the Issues panel). Returns true when a widget
|
||||
// was added or removed.
|
||||
export function syncFileWidgets(node, rows, { installed, hide, onPick }) {
|
||||
if (!node || !Array.isArray(node.widgets)) return false;
|
||||
const needed = rowsNeedingFiles(rows);
|
||||
const current = node.widgets.filter(isFileWidget);
|
||||
let changed = false;
|
||||
needed.forEach((row, index) => {
|
||||
let widget = current[index];
|
||||
if (!widget) {
|
||||
widget = node.addWidget("combo", `${FILE_WIDGET_PREFIX}${index + 1}`, row.name, (value) => {
|
||||
if (typeof value === "string" && value !== widget.ausbossRowName) onPick?.(widget.ausbossRowId, value);
|
||||
}, { values: () => fileChoices(widget.value, installed?.()), serialize: false });
|
||||
widget[FILE_WIDGET_FLAG] = true;
|
||||
widget.serialize = false;
|
||||
hide?.(widget);
|
||||
changed = true;
|
||||
}
|
||||
widget.value = row.name;
|
||||
widget.ausbossRowName = row.name;
|
||||
widget.ausbossRowId = row.id;
|
||||
});
|
||||
for (const widget of current.slice(needed.length)) {
|
||||
if (typeof node.removeWidget === "function") node.removeWidget(widget);
|
||||
else {
|
||||
const at = node.widgets.indexOf(widget);
|
||||
if (at >= 0) node.widgets.splice(at, 1);
|
||||
}
|
||||
changed = true;
|
||||
}
|
||||
return changed;
|
||||
}
|
||||
@@ -2,7 +2,7 @@
|
||||
// bar as long as the clip, with the cuts as marks. The server reads the script with the same code a run uses,
|
||||
// so the bar shows what will happen. It also colours the script box, so the
|
||||
// words the node reads as instructions (@1, 3s:, a cut) stand out.
|
||||
import { clipSetsDuration, linkSource } from "./h3_duration.mjs";
|
||||
import { clipLengthWords, clipSetsDuration, clipTimingFor, linkSource } from "./h3_duration.mjs";
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { BRAND, chainCallback } from "./index.mjs";
|
||||
@@ -38,7 +38,8 @@ function ensureCss() {
|
||||
.ausboss-strip-handle::before{content:"";position:absolute;left:5px;top:5px;bottom:5px;width:2px;border-radius:1px;background:rgba(207,232,229,.35)}
|
||||
.ausboss-strip-handle:hover::before,.ausboss-strip-handle.dragging::before{background:#cfe8e5}
|
||||
.ausboss-strip-handle span{position:absolute;bottom:100%;left:50%;transform:translateX(-50%);padding:1px 4px;border-radius:3px;background:#cfe8e5;color:#0b0f10;font-size:9px;white-space:nowrap;display:none}
|
||||
.ausboss-strip-handle.dragging span{display:block}
|
||||
.ausboss-strip-handle:hover span,.ausboss-strip-handle.dragging span{display:block}
|
||||
.ausboss-strip-handle.cuts::before{background:${BRAND}}
|
||||
.ausboss-strip-block.movable{cursor:grab;touch-action:none}
|
||||
.ausboss-strip-block.moving{z-index:3;cursor:grabbing;opacity:.85;box-shadow:0 4px 14px rgba(0,0,0,.5)}
|
||||
.ausboss-strip-drop{position:absolute;top:-3px;bottom:-3px;z-index:4;width:3px;margin-left:-2px;border-radius:2px;background:#f5b53d;pointer-events:none}
|
||||
@@ -108,7 +109,12 @@ export function momentStrip(node) {
|
||||
const label = document.createElement("span");
|
||||
handle.append(label);
|
||||
handle.style.left = `${rows[index].from}%`;
|
||||
handle.title = range ? "Drag to set when this moment starts. Click to add or remove a cut." : "Click to add or remove a cut.";
|
||||
const isCut = Boolean(rows[index].cut);
|
||||
handle.classList.toggle("cuts", isCut);
|
||||
// Shown on hover: what a click does here. Dragging swaps it for the time.
|
||||
const hint = isCut ? "\u2702 join" : "\u2702 cut";
|
||||
label.textContent = hint;
|
||||
handle.title = `${isCut ? "Click to join these two shots." : "Click to cut to a new shot here."}${range ? " Drag to set when this moment starts." : ""}`;
|
||||
track.append(handle);
|
||||
handle.addEventListener("pointerdown", (event) => {
|
||||
if (event.button !== 0) return;
|
||||
@@ -134,6 +140,7 @@ export function momentStrip(node) {
|
||||
handle.removeEventListener("pointerup", stop);
|
||||
handle.removeEventListener("pointercancel", stop);
|
||||
handle.classList.remove("dragging");
|
||||
label.textContent = hint;
|
||||
if (!moved) commitScript((text) => toggleCut(text, rows[index].line));
|
||||
else if (second !== null && second !== rows[index].start) commitScript((text) => retime(text, rows[index].line, second));
|
||||
};
|
||||
@@ -290,12 +297,15 @@ export function momentStrip(node) {
|
||||
asked += 1;
|
||||
ruler.textContent = "";
|
||||
track.textContent = "";
|
||||
const message = el("ausboss-strip-words", "Length comes from the clip");
|
||||
// Say what H3 will make of the clip's frames, when the clip node knows
|
||||
// its count; it rounds down to 5 + 17n, so a 48-frame clip gives 39.
|
||||
const words = clipLengthWords(clipTimingFor(node));
|
||||
const message = el("ausboss-strip-words", words?.title ?? "Length comes from the clip");
|
||||
message.style.cssText = "display:flex;align-items:center;height:100%";
|
||||
track.append(message);
|
||||
note.textContent = "Set the clip's length upstream. Timing controls are unavailable here.";
|
||||
note.textContent = words?.note ?? "Set the clip's length on the clip node. H3 makes 5 + 17n frames (22, 39, 56 ...), rounding the clip down.";
|
||||
note.title = note.textContent;
|
||||
note.classList.remove("warn");
|
||||
note.classList.toggle("warn", Boolean(words?.warn));
|
||||
return;
|
||||
}
|
||||
if (wired(node, "script")) {
|
||||
|
||||
@@ -59,14 +59,15 @@ export function dragRange(rows, index, seconds) {
|
||||
export function stripCaption(report) {
|
||||
if (report?.warning) return { text: report.warning, warn: true };
|
||||
const moments = report?.moments ?? [];
|
||||
if (!moments.length) return { text: "Each line becomes a moment.", warn: false };
|
||||
if (!moments.length) return { text: "Each line becomes a moment. A blank line between lines is a cut.", warn: false };
|
||||
const count = moments.length === 1 ? "1 moment" : `${moments.length} moments`;
|
||||
const cuts = moments.filter((moment) => moment.cut).length;
|
||||
if (cuts) return { text: `${count}, ${cuts === 1 ? "1 cut" : `${cuts} cuts`}.`, warn: false };
|
||||
// Say how to cut, since nothing on the bar shows it until you hover an edge.
|
||||
if (cuts) return { text: `${count}, ${cuts === 1 ? "1 cut" : `${cuts} cuts`}. Click a cut on the bar to join the shots.`, warn: false };
|
||||
const each = moments.map((moment) => trimmed(moment.end - moment.start));
|
||||
if (moments.length === 1) return { text: `${count}, ${each[0]} s, one take.`, warn: false };
|
||||
if (moments.length === 1) return { text: `${count}, ${each[0]} s, one take. Add a line for the next moment, a blank line to cut.`, warn: false };
|
||||
const even = each.every((length) => length === each[0]);
|
||||
return { text: even ? `${count}, ${each[0]} s each, one take.` : `${count}, one take.`, warn: false };
|
||||
return { text: `${even ? `${count}, ${each[0]} s each` : count}, one take. Click between two moments to cut there.`, warn: false };
|
||||
}
|
||||
|
||||
// The seconds the bar under the script marks: every second up to 10 s,
|
||||
|
||||
@@ -5,6 +5,11 @@
|
||||
// 2.0 draws each node as an element on the page, which can run past those
|
||||
// bounds (its frame, a footer), so there the element itself is measured.
|
||||
// Groups are drawn on the canvas in both looks.
|
||||
//
|
||||
// A frameless tile (a link tile, a title, a note tile) draws only a card
|
||||
// inside a transparent node, so what people see is the card, not the node's
|
||||
// bounds. Such a node names that card as `ausbossOutlineElement`, and its
|
||||
// outline is the card's, in both looks.
|
||||
|
||||
function graphRectOf(element, canvas) {
|
||||
const box = element.getBoundingClientRect();
|
||||
@@ -34,6 +39,11 @@ export function nodeElements() {
|
||||
}
|
||||
|
||||
export function outlineOf(item, canvas, { node = true, elements = null } = {}) {
|
||||
const face = node && !item?.flags?.collapsed ? item?.ausbossOutlineElement : null;
|
||||
if (face?.isConnected && canvas?.canvas && canvas.ds) {
|
||||
const rect = graphRectOf(face, canvas);
|
||||
if (rect) return rect;
|
||||
}
|
||||
if (node && globalThis.LiteGraph?.vueNodesMode === true && item?.id != null && canvas?.canvas && canvas.ds) {
|
||||
const element = elements
|
||||
? elements.get(String(item.id))
|
||||
|
||||
@@ -25,7 +25,11 @@ export function ensurePanelControls() {
|
||||
.ausboss-panel-switch i::after{content:"";position:absolute;top:2px;left:2px;width:9px;height:9px;border-radius:50%;background:${T.inkSoft}}
|
||||
.ausboss-panel-switch[aria-checked=true] i{background:${T.brand}}
|
||||
.ausboss-panel-switch[aria-checked=true] i::after{left:13px;background:${T.inkHover}}
|
||||
.ausboss-compact-action:focus-visible,.ausboss-panel-switch:focus-visible{outline:2px solid ${T.brand};outline-offset:1px}
|
||||
.ausboss-primary-action{box-sizing:border-box;flex:none;height:22px;min-width:24px;padding:0 14px;border:0;border-radius:${RADIUS_XS}px;background:${T.brand};color:${T.onBrand};font:700 11px/1 ${controlFont};white-space:nowrap;cursor:pointer;pointer-events:auto}
|
||||
.ausboss-primary-action:hover:enabled{filter:brightness(1.12)}
|
||||
.ausboss-primary-action.quiet{background:${T.fill};color:${T.inkSoft}}
|
||||
.ausboss-primary-action:disabled{opacity:.6;cursor:default}
|
||||
.ausboss-compact-action:focus-visible,.ausboss-primary-action:focus-visible,.ausboss-panel-switch:focus-visible{outline:2px solid ${T.brand};outline-offset:1px}
|
||||
`;
|
||||
document.head.append(style);
|
||||
}
|
||||
@@ -40,6 +44,19 @@ export function compactAction(label, title = "") {
|
||||
return button;
|
||||
}
|
||||
|
||||
// The one thing a panel wants done next, as a solid button: Save Image's
|
||||
// "Save picture" over a picture that is not saved yet. `quiet` (a class the
|
||||
// panel sets) is the same button once it has nothing left to do.
|
||||
export function primaryAction(label, title = "") {
|
||||
ensurePanelControls();
|
||||
const button = document.createElement("button");
|
||||
button.type = "button";
|
||||
button.className = "ausboss-primary-action";
|
||||
button.textContent = label;
|
||||
button.title = title;
|
||||
return button;
|
||||
}
|
||||
|
||||
export function panelSwitch(label, title) {
|
||||
ensurePanelControls();
|
||||
const button = document.createElement("button");
|
||||
|
||||
@@ -52,9 +52,10 @@ export function makePictureStepper({ noun = "picture", height = 24, onStep, sign
|
||||
root.append(arrow("‹", -1, `Previous ${noun}`), label, arrow("›", 1, `Next ${noun}`));
|
||||
return {
|
||||
root,
|
||||
set(index, count) {
|
||||
// `text` replaces the "2 / 5" for a stop that is not one picture.
|
||||
set(index, count, text) {
|
||||
root.hidden = !(count > 1);
|
||||
label.textContent = stepperLabel(index, count);
|
||||
label.textContent = text ?? stepperLabel(index, count);
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
@@ -19,3 +19,75 @@ export function stepIndex(index, count, step) {
|
||||
export function stepperLabel(index, count) {
|
||||
return count > 1 ? `${keepIndex(index, count) + 1} / ${count}` : "";
|
||||
}
|
||||
|
||||
// The preview panel has one more stop, before picture 1: every picture at
|
||||
// once. ALL stands for it where an index would be.
|
||||
export const ALL = -1;
|
||||
|
||||
// One step back or on through "all, 1, 2 ... count", round at both ends.
|
||||
export function stepWithAll(index, count, step) {
|
||||
if (!(count > 0)) return ALL;
|
||||
const stops = count + 1;
|
||||
const at = index === ALL ? 0 : keepIndex(index, count) + 1;
|
||||
const next = (((at + Math.sign(Number(step) || 0)) % stops) + stops) % stops;
|
||||
return next === 0 ? ALL : next - 1;
|
||||
}
|
||||
|
||||
export function allLabel(count) {
|
||||
return count > 1 ? `all ${count}` : "";
|
||||
}
|
||||
|
||||
// How to lay `count` pictures out in a box so each one draws as big as it
|
||||
// can: the columns and rows that leave a picture of this shape (`aspect`,
|
||||
// width over height) the most room.
|
||||
export function gridShape(count, width, height, aspect = 1) {
|
||||
const total = Math.max(1, Math.trunc(Number(count) || 0));
|
||||
const shape = Number.isFinite(aspect) && aspect > 0 ? aspect : 1;
|
||||
if (!(width > 0) || !(height > 0)) {
|
||||
const columns = Math.ceil(Math.sqrt(total));
|
||||
return { columns, rows: Math.ceil(total / columns) };
|
||||
}
|
||||
let best = { columns: 1, rows: total, size: 0 };
|
||||
for (let columns = 1; columns <= total; columns += 1) {
|
||||
const rows = Math.ceil(total / columns);
|
||||
const size = Math.min(width / columns, (height / rows) * shape);
|
||||
if (size > best.size) best = { columns, rows, size };
|
||||
}
|
||||
return { columns: best.columns, rows: best.rows };
|
||||
}
|
||||
|
||||
// The held pictures in a Save Image result, in the order they show: each
|
||||
// one's run token and its place in that run. A list run saves once for each
|
||||
// item, so one result can hold several tokens. Empty when the result holds
|
||||
// none, or when a picture in it is not one of the held ones.
|
||||
export function heldPlaces(record) {
|
||||
const tokens = new Set(Array.isArray(record?.ausboss_held) ? record.ausboss_held : []);
|
||||
if (!tokens.size) return [];
|
||||
const held = (Array.isArray(record.images) ? record.images : []).map((image) => {
|
||||
const match = /^([0-9a-f]{16})_(\d{3})\./.exec(String(image?.filename ?? ""));
|
||||
return match && tokens.has(match[1]) ? { token: match[1], index: Number(match[2]) } : null;
|
||||
});
|
||||
return held.length && held.every(Boolean) ? held : [];
|
||||
}
|
||||
|
||||
// What the row over held pictures says, and what its button does.
|
||||
// held how many pictures the node holds
|
||||
// kept how many of them are saved already
|
||||
// all a click keeps every picture (one picture, the grid, preview off)
|
||||
// done nothing is left to keep where the click would act
|
||||
// saving a save is on its way
|
||||
// name the file a single picture was saved as, when known
|
||||
// `quiet` is the button once it has nothing left to do.
|
||||
export function keepWords({ held = 0, kept = 0, all = true, done = false, saving = false, name = "" } = {}) {
|
||||
const note = !kept
|
||||
? "Not saved yet"
|
||||
: held < 2
|
||||
? (name ? `Saved as ${name}` : "Saved")
|
||||
: kept >= held ? `All ${held} saved` : `${kept} of ${held} saved`;
|
||||
if (saving) return { note, button: "Saving...", quiet: false };
|
||||
if (done) return { note, button: "Saved", quiet: true };
|
||||
const button = held < 2
|
||||
? "Save picture"
|
||||
: !all ? "Save this one" : kept ? "Save the rest" : `Save all ${held}`;
|
||||
return { note, button, quiet: false };
|
||||
}
|
||||
|
||||
@@ -37,6 +37,7 @@ export const SOURCE_LOADERS = Object.freeze({
|
||||
LoadImageOutput: picture("image"),
|
||||
LoadVideo: video("file"),
|
||||
AUSBOSS_NODES_ImageCropRotatePad: picture("image"),
|
||||
AUSBOSS_NODES_LoadImage: picture("image"),
|
||||
// A wired source_image replaces the file, so the file may stay blank.
|
||||
AUSBOSS_NODES_LoadImagePad: (inputs) => (linked(inputs?.source_image) ? null : picture("image")(inputs)),
|
||||
AUSBOSS_NODES_LoadAudio: sound("audio"),
|
||||
@@ -160,3 +161,51 @@ export function collapseRepeatedErrors(nodeErrors) {
|
||||
}
|
||||
return nodeErrors;
|
||||
}
|
||||
|
||||
// ------------------------------------------------------------ known run errors
|
||||
//
|
||||
// Some errors that stop a run come from outside the workflow and say nothing
|
||||
// a person can act on ("KeyError: 'latent'"). Each entry here names one such
|
||||
// error by what is in ComfyUI's own report, and says in plain words what to
|
||||
// do. An entry is added only after the cause was found for real.
|
||||
const KNOWN_RUN_ERRORS = Object.freeze([
|
||||
{
|
||||
// Muse MiniMax Director replaces part of H3 for every workflow, and that
|
||||
// part reads a picture from every guide. ComfyUI's own H3 nodes also make
|
||||
// guides that hold sound alone, which is how a clip keeps its sound.
|
||||
matches: ({ type, message, traceback }) =>
|
||||
type === "KeyError" && message.includes("'latent'") && traceback.includes("muse_minimax_hybrid_patch"),
|
||||
summary: "Another node pack stopped this run",
|
||||
detail:
|
||||
"This error comes from the Muse MiniMax Director pack, not from this workflow. Turn that pack off in "
|
||||
+ "Manager, restart ComfyUI and run again.",
|
||||
},
|
||||
{
|
||||
// The w6a8 H3 files load from ComfyUI 0.38.0 on; older versions stop in
|
||||
// the model loader on the name of the format.
|
||||
matches: ({ message }) => message.includes("w6a8_int8"),
|
||||
summary: "This model needs a newer ComfyUI",
|
||||
detail:
|
||||
"Your ComfyUI is too old to read this model file. Update ComfyUI to 0.38 or newer and run again.",
|
||||
},
|
||||
]);
|
||||
|
||||
/**
|
||||
* Plain advice for a run error the pack knows the cause of, from the
|
||||
* `execution_error` event's detail. `null` for every other error.
|
||||
*/
|
||||
export function explainRunError(detail) {
|
||||
const report = {
|
||||
type: String(detail?.exception_type ?? "").split(".").pop(),
|
||||
message: String(detail?.exception_message ?? ""),
|
||||
traceback: [].concat(detail?.traceback ?? []).join("\n"),
|
||||
};
|
||||
for (const known of KNOWN_RUN_ERRORS) {
|
||||
try {
|
||||
if (known.matches(report)) return { summary: known.summary, detail: known.detail };
|
||||
} catch (_error) {
|
||||
// A report in a shape this entry did not expect is simply not a match.
|
||||
}
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
@@ -28,11 +28,15 @@ export function setTileFace(node, root, enabled, onEdit) {
|
||||
});
|
||||
}
|
||||
const listeners = new AbortController();
|
||||
// Snap into line and Line up use the card people see, not the
|
||||
// transparent node around it (js/shared/outline.mjs).
|
||||
node.ausbossOutlineElement = root.querySelector(".ausboss-callout-card,.ausboss-note-card") || root;
|
||||
installTileDrag(node, root, listeners.signal, onEdit);
|
||||
installTileResize(node, root, listeners.signal);
|
||||
originalDescriptors.set(node, { descriptors, listeners });
|
||||
} else {
|
||||
previous.listeners.abort();
|
||||
delete node.ausbossOutlineElement;
|
||||
for (const [key, descriptor] of previous.descriptors) {
|
||||
if (descriptor) Object.defineProperty(node, key, descriptor);
|
||||
else delete node[key];
|
||||
@@ -44,6 +48,9 @@ export function setTileFace(node, root, enabled, onEdit) {
|
||||
|
||||
function skipBadges() {}
|
||||
|
||||
// Sent from a tile's own drag after each move: js/snap_lines/ lines it up.
|
||||
export const TILE_MOVE = "ausboss-tile-move";
|
||||
|
||||
// A normal node moves by its header. Give the tile that gesture while
|
||||
// keeping a short click on a link or the info button as its usual action.
|
||||
function installTileDrag(node, root, signal, onEdit) {
|
||||
@@ -87,6 +94,9 @@ function installTileDrag(node, root, signal, onEdit) {
|
||||
for (const start of press.starts) start.item.pos = [start.x + dx / zoom, start.y + dy / zoom];
|
||||
node.setDirtyCanvas?.(true, true);
|
||||
event.stopPropagation();
|
||||
// The move stops here (Nodes 2.0 would pan the canvas with it), so tell
|
||||
// Snap into line separately, after the tile has been placed.
|
||||
root.dispatchEvent(new CustomEvent(TILE_MOVE, { bubbles: true, detail: event }));
|
||||
}, capture);
|
||||
const finish = (event) => {
|
||||
if (event.pointerId !== press?.id) return;
|
||||
|
||||
@@ -217,6 +217,16 @@ export function latestFirstFor(info, frames, everyNth = 1) {
|
||||
return Math.max(0, info.count - 1 - (lastFrameFor(0, frames, everyNth)));
|
||||
}
|
||||
|
||||
// How far a drag can move IN while a Length is on. A Length that fits stops
|
||||
// IN where OUT meets the source's end, so the count stays whole. A Length
|
||||
// as long as the clip or longer cannot keep its count anywhere: it only caps
|
||||
// the clip, the source's end is OUT, and IN moves over the whole clip.
|
||||
export function latestInFor(info, frames, everyNth = 1) {
|
||||
if (!info?.count) return 0;
|
||||
const latest = latestFirstFor(info, frames, everyNth);
|
||||
return latest > 0 ? latest : info.count - 1;
|
||||
}
|
||||
|
||||
// Everything the rail draws for the clip node, from the stored values:
|
||||
// mode which way sets the length (lengthMode)
|
||||
// first IN, the first source frame the run keeps
|
||||
@@ -263,7 +273,14 @@ export function clipLengthPlan(info, values = {}) {
|
||||
latestStart = clamp(Math.floor((end - lastTick - TRIM_EPSILON) * info.fps), 0, endFrame);
|
||||
}
|
||||
}
|
||||
if (values.resampled) return { ...base, last: cap > 0 ? null : window.last, requested, latestStart };
|
||||
if (values.resampled) {
|
||||
// A known output rate places every output frame: the backend samples
|
||||
// and holds on ticks at start + k / rate (decode_video_range), so the
|
||||
// count, the last source frame shown and the end are all known here.
|
||||
if (!(values.outputFps > 0) || !(info.fps > 0)) return { ...base, last: cap > 0 ? null : window.last, requested, latestStart };
|
||||
const resampled = resampledFrames(info, values, nth, requested, snap);
|
||||
return { ...base, ...resampled, requested, latestStart, truncated: requested != null && resampled.frames < requested };
|
||||
}
|
||||
const kept = keptFrames(window, nth, cap, snap);
|
||||
return { ...base, last: kept.lastKept, frames: kept.frames, requested, latestStart, truncated: requested != null && kept.frames < requested };
|
||||
};
|
||||
@@ -278,6 +295,27 @@ export function clipLengthPlan(info, values = {}) {
|
||||
return capped(mode === "length" ? stored : 0);
|
||||
}
|
||||
|
||||
// Output frames of a run that resamples to a known rate: ticks at
|
||||
// start + k / tickRate while before the end, thinned by every nth, capped,
|
||||
// then snapped. `last` is the source frame the final tick holds.
|
||||
export function resampledFrames(info, values, everyNth = 1, cap = null, snap = "free") {
|
||||
const nth = Math.max(1, Math.floor(finite(everyNth, 1)) || 1);
|
||||
const tickRate = finite(values.outputFps) * nth;
|
||||
const duration = info.duration || info.count / info.fps;
|
||||
const start = Math.max(0, finite(values.start));
|
||||
const end = finite(values.end) > 0 ? Math.min(finite(values.end), duration) : duration;
|
||||
const ticks = Math.max(0, Math.ceil((end - start) * tickRate - TRIM_EPSILON));
|
||||
let frames = Math.ceil(ticks / nth);
|
||||
if (cap != null && cap > 0) frames = Math.min(frames, cap);
|
||||
frames = snapFrameCount(frames, snap);
|
||||
const first = frameWindow(info, start, values.end).first;
|
||||
const lastTime = start + Math.max(0, frames - 1) * nth / tickRate;
|
||||
const last = clamp(Math.floor(lastTime * info.fps + TRIM_EPSILON), first, info.count - 1);
|
||||
const windowLast = frameWindow(info, start, values.end).last;
|
||||
// The last frame is held for one output frame, but never past the video.
|
||||
return { first, last, windowLast, frames, endSeconds: Math.min(end, start + frames / finite(values.outputFps)) };
|
||||
}
|
||||
|
||||
// Playback rate for a label, trimmed of float noise: 24 -> "24",
|
||||
// 23.976023... -> "23.976".
|
||||
export function formatFps(value) {
|
||||
|
||||
@@ -0,0 +1,40 @@
|
||||
// Small pieces every timing card shares (the clip node, Load Video, Load
|
||||
// Audio): the seconds | frames choice and the "linked" button that goes to
|
||||
// the node feeding a value.
|
||||
import { app } from "../../../scripts/app.js";
|
||||
|
||||
// Seconds or frames on a timing card: a per-browser choice, like a view
|
||||
// setting. It never changes a stored value.
|
||||
const UNITS_KEY = "AusBoss.ClipTiming.Units";
|
||||
let unitsCache = null;
|
||||
export function timingUnits(next) {
|
||||
if (next === "s" || next === "f") {
|
||||
unitsCache = next;
|
||||
try { localStorage.setItem(UNITS_KEY, next); } catch { /* private window */ }
|
||||
return next;
|
||||
}
|
||||
if (unitsCache) return unitsCache;
|
||||
try { unitsCache = localStorage.getItem(UNITS_KEY) === "f" ? "f" : "s"; } catch { unitsCache = "s"; }
|
||||
return unitsCache;
|
||||
}
|
||||
|
||||
// The node feeding a linked input, by its title, with a way to go to it.
|
||||
export function linkedNodeInfo(node, name) {
|
||||
const slot = node.inputs?.find((input) => input?.name === name || input?.widget?.name === name);
|
||||
if (slot?.link == null) return null;
|
||||
const graph = node.graph;
|
||||
const link = graph?.links?.get?.(slot.link) ?? graph?.links?.[slot.link] ?? graph?._links?.get?.(slot.link);
|
||||
const source = graph?.getNodeById?.(link?.origin_id);
|
||||
if (!source) return { title: "the subgraph's input", go: null };
|
||||
const title = source.title || source.type || "another node";
|
||||
return {
|
||||
title,
|
||||
go: () => {
|
||||
try {
|
||||
app.canvas?.selectNode?.(source);
|
||||
app.canvas?.centerOnNode?.(source);
|
||||
app.canvas?.setDirty?.(true, true);
|
||||
} catch { /* an older canvas */ }
|
||||
},
|
||||
};
|
||||
}
|
||||
@@ -57,9 +57,9 @@ export function goesBehind(flying, ask, epoch) {
|
||||
|
||||
// The pack's own picture loaders. One of them wired straight into Mask by
|
||||
// Touch is run by the node itself, the way a press on Show picture runs it:
|
||||
// it may crop, turn or pad its file, so the file alone is not the picture,
|
||||
// it may crop, turn, pad or resize its file, so the file alone is not the picture,
|
||||
// but running it costs next to nothing.
|
||||
export const OWN_LOADERS = new Set(["AUSBOSS_NODES_ImageCropRotatePad", "AUSBOSS_NODES_LoadImagePad"]);
|
||||
export const OWN_LOADERS = new Set(["AUSBOSS_NODES_ImageCropRotatePad", "AUSBOSS_NODES_LoadImage", "AUSBOSS_NODES_LoadImagePad"]);
|
||||
|
||||
// What a loader will hand on, as far as the page can tell: its file and its
|
||||
// settings, as one text. Another text means another picture. Widgets that
|
||||
@@ -74,6 +74,25 @@ export function loaderSignature(widgets) {
|
||||
return JSON.stringify(settings);
|
||||
}
|
||||
|
||||
// What the picture says when a run stopped at the node. ComfyUI's own text
|
||||
// for a full graphics card is ten lines of numbers; here it is one plain
|
||||
// line. Anything else is said as it came.
|
||||
export const OUT_OF_MEMORY = "The graphics card ran out of memory, so the node could not look. Free some memory on it, then click again.";
|
||||
export function stopMessage(message) {
|
||||
const text = String(message ?? "");
|
||||
return /out of memory|OutOfMemory/i.test(text) ? OUT_OF_MEMORY : text;
|
||||
}
|
||||
|
||||
// A name as it is typed under the picture: "-face" (or the minus sign a
|
||||
// phone types) takes the name away, "+face" adds it; without a sign it does
|
||||
// what `takes` says. Returns { word, add }, the word cleaned.
|
||||
export function typedName(text, takes = false) {
|
||||
const typed = String(text ?? "").trim();
|
||||
const sign = /^[-\u2212\u2013+]/.exec(typed)?.[0] ?? "";
|
||||
const word = cleanWord(sign ? typed.slice(1) : typed);
|
||||
return { word, add: sign ? sign === "+" : !takes };
|
||||
}
|
||||
|
||||
// A word as the node sends it to SAM 3: commas, colons and brackets are the
|
||||
// text reader's own syntax and go; spaces are collapsed.
|
||||
export function cleanWord(word) {
|
||||
@@ -1044,19 +1063,46 @@ export function strokePieces(planes, list, sizes, pixels, stroke, ranks, wholes
|
||||
// picked area, and as untouched while almost none of it is.
|
||||
const NEARLY = 0.03;
|
||||
|
||||
// Which pieces hold a whole pick: nearly all of one picked area (`areas`,
|
||||
// on/off masks, one per chip) lies inside the piece. A person holds the
|
||||
// sweater picked on them; hair that a picked jacket only laps over does not.
|
||||
export const HOLDS = 0.9;
|
||||
export function piecesHolding(planes, count, areas) {
|
||||
const holds = new Uint8Array(count);
|
||||
for (const area of areas ?? []) {
|
||||
if (!area) continue;
|
||||
let size = 0;
|
||||
for (let index = 0; index < area.length; index += 1) if (area[index]) size += 1;
|
||||
if (!size) continue;
|
||||
const inside = piecesInside(planes, count, area);
|
||||
for (let piece = 0; piece < count; piece += 1) if (inside[piece] >= HOLDS * size) holds[piece] = 1;
|
||||
}
|
||||
return holds;
|
||||
}
|
||||
|
||||
// Which piece a pointer at pixel `index` lights up: the first piece there,
|
||||
// in the order they are offered (`ranks`; without it, smallest first), that
|
||||
// would still change something. Adding, that is a piece not yet (nearly)
|
||||
// all picked, so clicking the same spot again goes on from the glasses to
|
||||
// the face to the person. Taking away, it is a piece that has some of the
|
||||
// picked area in it. -1 when there is none.
|
||||
export function pieceAt(planes, index, { sizes, inside, adding = true, ranks = null } = {}) {
|
||||
//
|
||||
// Taking away from a spot that is not picked itself (`here` false) never
|
||||
// means a thing that holds a whole pick somewhere else (`holds`, from
|
||||
// piecesHolding): with a sweater picked, pointing at the face does not
|
||||
// offer the person, which would take the sweater along.
|
||||
export function pieceAt(planes, index, { sizes, inside, adding = true, ranks = null, here = true, holds = null } = {}) {
|
||||
for (const piece of candidatesAt(planes, index, ranks)) {
|
||||
const size = sizes?.[piece] ?? 0;
|
||||
if (!size) continue;
|
||||
const picked = inside?.[piece] ?? 0;
|
||||
const slack = Math.max(4, size * NEARLY);
|
||||
if (adding ? size - picked > slack : picked > slack) return piece;
|
||||
if (adding) {
|
||||
if (size - picked > slack) return piece;
|
||||
continue;
|
||||
}
|
||||
if (picked <= slack || (!here && holds?.[piece])) continue;
|
||||
return piece;
|
||||
}
|
||||
return -1;
|
||||
}
|
||||
|
||||
+106
-19
@@ -77,6 +77,7 @@ import {
|
||||
pieceRanks,
|
||||
pieceWholes,
|
||||
pieceSizes,
|
||||
piecesHolding,
|
||||
piecesInside,
|
||||
prefixStamps,
|
||||
previewPixels,
|
||||
@@ -85,9 +86,11 @@ import {
|
||||
sharedPixels,
|
||||
shrinkMask,
|
||||
stampOf,
|
||||
stopMessage,
|
||||
strokePieces,
|
||||
thinStroke,
|
||||
touchFits,
|
||||
typedName,
|
||||
} from "./touch_math.mjs";
|
||||
|
||||
const CSS_ID = "ausboss-touch-css";
|
||||
@@ -190,9 +193,17 @@ function ensureCss() {
|
||||
.ausboss-touch-chip>button:hover{opacity:1;background:rgba(255,255,255,.14);}
|
||||
.ausboss-touch-chip>button.more{padding:0 6px;background:rgba(0,0,0,.35);opacity:.95;}
|
||||
.ausboss-touch-chip.open>button.more{background:rgba(0,0,0,.2);}
|
||||
.ausboss-touch-name{box-sizing:border-box;flex:none;width:104px;height:20px;padding:0 9px;border:1px dashed rgba(0,180,170,.5);border-radius:10px;background:transparent;color:#dffcf9;font:600 10px/18px -apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif;outline:none;}
|
||||
.ausboss-touch-namebox{box-sizing:border-box;flex:none;display:inline-flex;align-items:stretch;height:20px;border:1px dashed rgba(0,180,170,.5);border-radius:10px;overflow:hidden;}
|
||||
.ausboss-touch-namebox:focus-within{border-style:solid;border-color:${BRAND};background:rgba(0,0,0,.4);}
|
||||
.ausboss-touch-namebox.away{border-color:rgba(${LIT_AWAY.join(",")},.6);}
|
||||
.ausboss-touch-namebox.away:focus-within{border-color:rgb(${LIT_AWAY.join(",")});}
|
||||
.ausboss-touch-sign{box-sizing:border-box;flex:none;width:22px;padding:0;border:0;border-right:1px solid rgba(0,180,170,.35);background:rgba(0,180,170,.16);color:#dffcf9;font:700 13px/18px -apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif;cursor:pointer;}
|
||||
.ausboss-touch-sign:hover{background:rgba(0,180,170,.34);color:#fff;}
|
||||
.ausboss-touch-sign:disabled{opacity:.4;cursor:default;}
|
||||
.ausboss-touch-namebox.away .ausboss-touch-sign{border-right-color:rgba(${LIT_AWAY.join(",")},.4);background:rgba(${LIT_AWAY.join(",")},.24);color:#ffe9e2;}
|
||||
.ausboss-touch-namebox.away .ausboss-touch-sign:hover{background:rgba(${LIT_AWAY.join(",")},.42);}
|
||||
.ausboss-touch-name{box-sizing:border-box;flex:none;width:118px;height:18px;padding:0 8px 0 6px;border:0;background:transparent;color:#dffcf9;font:600 10px/18px -apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif;outline:none;}
|
||||
.ausboss-touch-name::placeholder{color:#78908e;font-weight:500;}
|
||||
.ausboss-touch-name:focus{border-style:solid;border-color:${BRAND};background:rgba(0,0,0,.4);}
|
||||
.ausboss-touch-gap{flex:1 1 auto;min-width:0;}
|
||||
.ausboss-touch-stage{position:relative;flex:1 1 auto;min-height:0;width:100%;overflow:hidden;border:1px solid rgba(0,180,170,.27);border-radius:6px;background:rgba(0,0,0,.28);}
|
||||
[data-node-id] .ausboss-touch-stage{box-sizing:border-box;min-height:${STAGE_MIN}px;}
|
||||
@@ -224,9 +235,12 @@ function ensureCss() {
|
||||
.ausboss-touch.fining .ausboss-touch-fining{display:flex;}
|
||||
.ausboss-touch-fining::before{content:"";flex:none;width:7px;height:7px;border-radius:50%;background:${BRAND};animation:ausboss-touch-pulse 1s ease-in-out infinite;}
|
||||
@keyframes ausboss-touch-pulse{0%,100%{opacity:.25;}50%{opacity:1;}}
|
||||
.ausboss-touch-btn.lookpill{display:inline-flex;align-items:center;gap:5px;height:${BAR_HEIGHT}px;padding:0 9px 0 8px;border:1px solid rgba(0,180,170,.6);border-radius:10px;letter-spacing:0;text-transform:none;font:600 10px/18px -apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif;color:#dffcf9;background:linear-gradient(90deg,rgba(0,180,170,.38) var(--done,0%),rgba(0,180,170,.08) var(--done,0%));}
|
||||
.ausboss-touch-btn.lookpill{display:inline-flex;align-items:center;gap:5px;height:${BAR_HEIGHT}px;padding:0 9px 0 0;overflow:hidden;border:1px solid rgba(0,180,170,.8);border-radius:10px;letter-spacing:0;text-transform:none;font:600 10px/18px -apple-system,BlinkMacSystemFont,"Segoe UI",sans-serif;color:#dffcf9;background:linear-gradient(90deg,rgba(0,180,170,.38) var(--done,0%),rgba(0,180,170,.08) var(--done,0%));}
|
||||
.ausboss-touch-btn.lookpill:hover{border-color:${BRAND};color:#fff;background:linear-gradient(90deg,rgba(0,180,170,.5) var(--done,0%),rgba(0,180,170,.16) var(--done,0%));}
|
||||
.ausboss-touch-btn.lookpill .icon{flex:none;font-size:8px;letter-spacing:-1px;opacity:.9;}
|
||||
.ausboss-touch-btn.lookpill .key{flex:none;display:inline-flex;align-items:center;gap:4px;align-self:stretch;padding:0 7px 0 8px;background:${BRAND};color:#04201e;font-weight:700;}
|
||||
.ausboss-touch-btn.lookpill:hover .key{background:#35d6cb;}
|
||||
.ausboss-touch-btn.lookpill.paused .key{background:#9fb4b2;}
|
||||
.ausboss-touch-btn.lookpill .icon{flex:none;font-size:8px;letter-spacing:-1px;}
|
||||
.ausboss-touch-btn.lookpill.paused{border-color:rgba(120,144,142,.6);color:#c8dddd;background:linear-gradient(90deg,rgba(120,144,142,.3) var(--done,0%),rgba(120,144,142,.06) var(--done,0%));}
|
||||
.ausboss-touch-btn.lookpill.paused .icon{letter-spacing:0;}
|
||||
.ausboss-touch-btn.lookpill .count{opacity:.8;}
|
||||
@@ -527,6 +541,7 @@ function resolve(face) {
|
||||
face.raw = raw;
|
||||
face.exact = exact;
|
||||
face.chipAreas = null;
|
||||
face.holds = null;
|
||||
face.inside = pic.pieces ? piecesInside(pic.pieces.planes, pic.pieces.list.length, raw) : null;
|
||||
face.hoverMask = null;
|
||||
}
|
||||
@@ -697,6 +712,11 @@ function requestRun(face, ask) {
|
||||
return false;
|
||||
}
|
||||
const run = { ask, stamp: face.stamp, epoch: face.sourceEpoch, queued: false };
|
||||
// Runs of this node that are at a picture it no longer has. The one that
|
||||
// is running ends by itself once the next job waits; the ones that still
|
||||
// wait in the queue would each run for nothing, one after the other, so
|
||||
// they are taken out before this one is sent.
|
||||
const stale = face.flying.filter((other) => other.queued && other.epoch !== run.epoch);
|
||||
face.flying.push(run);
|
||||
face.inFlight = true;
|
||||
face.askedAt = performance.now();
|
||||
@@ -707,7 +727,15 @@ function requestRun(face, ask) {
|
||||
face.flying = face.flying.filter((other) => other !== run);
|
||||
if (!face.flying.length) settle(face);
|
||||
};
|
||||
Promise.resolve(face.host.run(ask === "scope" ? scopeAsk(face) : ask)).then((queued) => {
|
||||
const cleared = stale.length && face.host.dropWaiting
|
||||
? Promise.resolve(face.host.dropWaiting()).then((gone) => {
|
||||
// What was taken out never answers: the newest of the stale ones,
|
||||
// because a run of this node goes to the front of the queue.
|
||||
const out = new Set(stale.slice(Math.max(0, stale.length - (Number(gone) || 0))));
|
||||
if (out.size) face.flying = face.flying.filter((other) => !out.has(other));
|
||||
}, () => {})
|
||||
: Promise.resolve();
|
||||
cleared.then(() => face.host.run(ask === "scope" ? scopeAsk(face) : ask)).then((queued) => {
|
||||
if (queued === false) {
|
||||
failed();
|
||||
return;
|
||||
@@ -1258,11 +1286,23 @@ function takeAll(face, index) {
|
||||
refreshHover(face);
|
||||
}
|
||||
|
||||
// Whether a typed name is taken away: what the sign in front of the field
|
||||
// was set to, or else what the tool does right now (Erase, or Ctrl held).
|
||||
function nameTakes(face) {
|
||||
return face.nameAway ?? face.away;
|
||||
}
|
||||
|
||||
// A typed name, as every thing of that name. "-face" takes it away and
|
||||
// "+face" adds it, whatever the sign says.
|
||||
function addName(face, text) {
|
||||
const word = cleanWord(text);
|
||||
if (!word || !face.pic || face.held) return false;
|
||||
const typed = typedName(text, nameTakes(face));
|
||||
if (!typed.word || !face.pic || face.held) return false;
|
||||
flushPending(face);
|
||||
addItem(face, { t: "all", op: face.away ? 0 : 1, w: word });
|
||||
if (!typed.add && !face.items.some((item) => item.op === 1)) {
|
||||
setNote(face, "Nothing is picked yet. Pick something first, then take a name away from it.", { life: 3200 });
|
||||
return false;
|
||||
}
|
||||
addItem(face, { t: "all", op: typed.add ? 1 : 0, w: typed.word });
|
||||
return true;
|
||||
}
|
||||
|
||||
@@ -1413,9 +1453,19 @@ function syncTools(face) {
|
||||
face.lookButton.style.display = offer ? "" : "none";
|
||||
}
|
||||
if (face.nameField) {
|
||||
face.nameField.disabled = !live || face.held > 0;
|
||||
// A typed name is added, or taken away while the tool takes away.
|
||||
face.nameField.placeholder = face.away ? "\u2212 type a name" : "+ type a name";
|
||||
const off = !live || face.held > 0;
|
||||
const takes = nameTakes(face);
|
||||
face.nameField.disabled = off;
|
||||
face.nameSign.disabled = off;
|
||||
face.nameBox.classList.toggle("away", takes);
|
||||
face.nameSign.textContent = takes ? "\u2212" : "+";
|
||||
face.nameSign.title = takes
|
||||
? "A name you type is taken away from the mask. Click to add by name instead."
|
||||
: "A name you type is added to the mask. Click to take away by name instead.";
|
||||
face.nameField.placeholder = takes ? "take away by name" : "add by name";
|
||||
face.nameField.title = takes
|
||||
? "Type what to take out of the mask, like \"face\" or \"hair\", and press Enter. Every thing of that name is taken away."
|
||||
: "Type what to pick, like \"screen\" or \"earring\", and press Enter. Every thing of that name is taken. Start with a minus, like \"-face\", to take it away instead.";
|
||||
}
|
||||
if (face.newButton) face.newButton.disabled = !live || face.held > 0 || Boolean(face.fresh);
|
||||
// The switch shows what a press does right now: with Ctrl held, Paint
|
||||
@@ -1949,12 +1999,31 @@ function pointOf(face, event) {
|
||||
return [clamp(x, 0, face.pic.width - 1), clamp(y, 0, face.pic.height - 1)];
|
||||
}
|
||||
|
||||
// The pieces that hold a whole pick, worked out once per change to the
|
||||
// picks and only when something is being taken away.
|
||||
function holdsOf(face) {
|
||||
const pic = face.pic;
|
||||
if (!pic?.pieces || !face.raw) return null;
|
||||
if (!face.holds || face.holds.length !== pic.pieces.list.length) {
|
||||
const areas = chipGroups(face.items).filter((chip) => !chip.away).map((chip) => chipArea(face, chip));
|
||||
face.holds = piecesHolding(pic.pieces.planes, pic.pieces.list.length, areas);
|
||||
}
|
||||
return face.holds;
|
||||
}
|
||||
|
||||
function pieceUnder(face, point, adding) {
|
||||
const pic = face.pic;
|
||||
if (!pic?.pieces || !point) return -1;
|
||||
const x = clamp(Math.floor(point[0] * pic.scale), 0, pic.pw - 1);
|
||||
const y = clamp(Math.floor(point[1] * pic.scale), 0, pic.ph - 1);
|
||||
return pieceAt(pic.pieces.planes, y * pic.pw + x, { sizes: pic.pieces.sizes, inside: face.inside, adding, ranks: pic.pieces.ranks });
|
||||
const at = y * pic.pw + x;
|
||||
if (adding) return pieceAt(pic.pieces.planes, at, { sizes: pic.pieces.sizes, inside: face.inside, ranks: pic.pieces.ranks });
|
||||
// Taking away from a spot that is not picked: never a thing that holds a
|
||||
// whole pick somewhere else (the person, for the sweater picked on them).
|
||||
return pieceAt(pic.pieces.planes, at, {
|
||||
sizes: pic.pieces.sizes, inside: face.inside, adding, ranks: pic.pieces.ranks,
|
||||
here: face.raw?.[at] >= 0.5, holds: holdsOf(face),
|
||||
});
|
||||
}
|
||||
|
||||
// The pixel of the small copy under a point of the picture.
|
||||
@@ -2324,7 +2393,8 @@ function syncLooking(face) {
|
||||
pill.classList.toggle("paused", paused);
|
||||
pill.style.setProperty("--done", `${total ? Math.round((done / total) * 100) : 0}%`);
|
||||
face.lookIcon.textContent = paused ? "\u25b6" : "\u275a\u275a";
|
||||
face.lookWord.textContent = paused ? "Paused" : "Looking";
|
||||
face.lookVerb.textContent = paused ? "Look on" : "Pause";
|
||||
face.lookWord.textContent = paused ? "paused" : "looking";
|
||||
face.lookCount.textContent = total ? `${done}/${total}` : "\u2026";
|
||||
const far = total ? ` ${done} of ${total}` : "";
|
||||
pill.title = paused
|
||||
@@ -2717,6 +2787,8 @@ function setSoft(face, soft) {
|
||||
|
||||
// Pick takes whole things; Paint and Erase are the brush.
|
||||
function setTool(face, tool) {
|
||||
// The sign in front of the name field follows the tool again.
|
||||
face.nameAway = null;
|
||||
if (!TOOLS.includes(tool)) return;
|
||||
face.tool = tool;
|
||||
face.snap = tool === "pick";
|
||||
@@ -2773,16 +2845,21 @@ export function mountTouchStage(node, host) {
|
||||
const bar = el("div", "ausboss-touch-bar");
|
||||
const chips = el("div", "ausboss-touch-chips");
|
||||
const chipList = el("span", "ausboss-touch-chiplist");
|
||||
// A name typed here picks every thing of that name, or takes it away:
|
||||
// the sign in front says which, and a press on it changes it.
|
||||
const nameBox = el("span", "ausboss-touch-namebox");
|
||||
const nameSign = el("button", "ausboss-touch-sign", "+");
|
||||
nameSign.type = "button";
|
||||
const nameField = el("input", "ausboss-touch-name");
|
||||
nameField.type = "text";
|
||||
nameField.placeholder = "+ type a name";
|
||||
nameField.placeholder = "add by name";
|
||||
nameField.maxLength = 60;
|
||||
nameField.spellcheck = false;
|
||||
nameField.title = "Type what to pick, like \"screen\" or \"earring\", and press Enter. Every thing of that name is taken.";
|
||||
nameBox.append(nameSign, nameField);
|
||||
const newButton = el("button", "ausboss-touch-new", "+ new");
|
||||
newButton.type = "button";
|
||||
newButton.title = "Start a new area: what you paint next gets a chip of its own, also where it touches another area.";
|
||||
chips.append(chipList, newButton, nameField);
|
||||
chips.append(chipList, newButton, nameBox);
|
||||
root.append(tools, stage, bar, chips);
|
||||
|
||||
const frame = el("div", "ausboss-touch-frame");
|
||||
@@ -2827,7 +2904,7 @@ export function mountTouchStage(node, host) {
|
||||
|
||||
const face = node.__ausbossTouch = {
|
||||
node, host, root, stage, frame, img, tint, lit, ink, path, dots, ringBack, ring, coreRing, dab, brushLabel, tag, hint, noteEl, abort,
|
||||
chips, chipList, newButton, nameField, chipsShown: null,
|
||||
chips, chipList, newButton, nameField, nameBox, nameSign, nameAway: null, chipsShown: null,
|
||||
flying: [],
|
||||
pendingTimer: 0,
|
||||
alive: true,
|
||||
@@ -3031,10 +3108,15 @@ export function mountTouchStage(node, host) {
|
||||
if (face.pic?.paused && !face.inFlight) resumeLook(face);
|
||||
else pauseLook(face);
|
||||
}, signal);
|
||||
// The press it takes comes first, as a button of its own look: "Pause"
|
||||
// while the node looks, "Look on" once it is paused.
|
||||
face.lookIcon = el("span", "icon");
|
||||
face.lookVerb = el("span", "verb");
|
||||
face.lookWord = el("span", "word");
|
||||
face.lookCount = el("span", "count");
|
||||
face.lookPill.append(face.lookIcon, face.lookWord, face.lookCount);
|
||||
const lookKey = el("span", "key");
|
||||
lookKey.append(face.lookIcon, face.lookVerb);
|
||||
face.lookPill.append(lookKey, face.lookWord, face.lookCount);
|
||||
face.lookPill.style.display = "none";
|
||||
// First in the line under the picture: what the node is doing with the
|
||||
// picture, then what is picked; the buttons that pick come after.
|
||||
@@ -3044,6 +3126,11 @@ export function mountTouchStage(node, host) {
|
||||
for (const type of ["keydown", "keyup", "keypress", "pointerdown", "mousedown", "dblclick"]) {
|
||||
chips.addEventListener(type, (event) => event.stopPropagation(), { signal });
|
||||
}
|
||||
nameSign.addEventListener("click", () => {
|
||||
face.nameAway = !nameTakes(face);
|
||||
syncTools(face);
|
||||
nameField.focus();
|
||||
}, { signal });
|
||||
nameField.addEventListener("keydown", (event) => {
|
||||
if (event.key === "Enter") {
|
||||
event.preventDefault();
|
||||
@@ -3096,7 +3183,7 @@ export function mountTouchStage(node, host) {
|
||||
// The run stopped at this node with a message for the person.
|
||||
face.stopped = (message) => {
|
||||
settle(face);
|
||||
setNote(face, String(message ?? ""));
|
||||
setNote(face, stopMessage(message));
|
||||
// It stays up until something else is said or the next result comes in.
|
||||
face.stopNote = true;
|
||||
};
|
||||
|
||||
@@ -1,5 +1,6 @@
|
||||
import { api } from "../../../scripts/api.js";
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { linkedNodeInfo, timingUnits as clipUnits } from "./timing_links.mjs";
|
||||
import { PANEL_LABELS } from "./panel_controls.mjs";
|
||||
import { hideInputsInDef, hideWidget } from "./widget_visibility.mjs";
|
||||
import { mountTransformTrim } from "./transform_trim.mjs";
|
||||
@@ -81,6 +82,7 @@ import {
|
||||
} from "./transform_geometry.mjs";
|
||||
import { clampFrame, clipInfo, frameTime } from "./timeline_math.mjs";
|
||||
import { liftSocket } from "./widget_card.mjs";
|
||||
import { COPY_EDITS_SETTING, copyPicture, transformPicture } from "./transform_picture.mjs";
|
||||
|
||||
const HIDDEN_WIDGETS = [
|
||||
"image", "upload",
|
||||
@@ -103,11 +105,10 @@ const trimInputDriven = (node, name) => node.inputs?.some(
|
||||
|
||||
const RESIZE_METHODS = ["lanczos", "area", "bicubic", "bilinear", "nearest-exact"];
|
||||
// The widest the control rows get on a big node; the picture takes the rest.
|
||||
const CONTROL_MAX_WIDTH = 460;
|
||||
const TRANSFORM_DEFAULTS = resetTransformValues(false);
|
||||
const CORE_IMAGE_PREVIEW_WIDGET = "$$canvas-image-preview";
|
||||
|
||||
// The gear on the image node's source card. Display choices only, kept for
|
||||
// The gear on the image node's source card. Browser preferences, kept for
|
||||
// every image node on this browser (settings_menu.mjs stores them).
|
||||
const SETTINGS_SCOPE = "image_crop_rotate_pad";
|
||||
const SETTINGS_SCHEMA = [
|
||||
@@ -115,6 +116,7 @@ const SETTINGS_SCHEMA = [
|
||||
key: "show_mask", label: "Show the mask on the picture", type: "toggle", default: true,
|
||||
hint: "Teal over the parts of your picture the model paints: a mask drawn in the MaskEditor, or see-through parts of a PNG.",
|
||||
},
|
||||
COPY_EDITS_SETTING,
|
||||
];
|
||||
let transformSettings = null;
|
||||
function settings() {
|
||||
@@ -150,7 +152,7 @@ function installStyles() {
|
||||
style.id = "ausboss-transform-styles";
|
||||
style.textContent = `
|
||||
.ausboss-transform-panel{display:flex;flex-direction:column;gap:8px;padding:8px;color:#ddd;font:12px system-ui;box-sizing:border-box;width:100%;height:100%;overflow:hidden}
|
||||
.ausboss-transform-panel>:not(.ausboss-transform-preview){width:100%;max-width:${CONTROL_MAX_WIDTH}px;align-self:center;box-sizing:border-box}
|
||||
.ausboss-transform-panel>:not(.ausboss-transform-preview){width:100%;align-self:stretch;box-sizing:border-box}
|
||||
.ausboss-transform-preview{width:100%;flex:1 1 180px;min-height:0;border:1px solid #50555b;border-radius:8px;background:#111;display:block;touch-action:none}
|
||||
.ausboss-transform-source{display:flex;flex-direction:column;gap:7px;flex:0 0 auto;padding:8px;border:1px solid rgba(0,184,174,.28);border-radius:8px;background:rgba(0,0,0,.24)}
|
||||
.ausboss-transform-source-image{display:grid;grid-template-columns:minmax(0,1fr) auto;align-items:center}
|
||||
@@ -189,13 +191,14 @@ function installStyles() {
|
||||
.ausboss-transform-group-meta{min-width:0;overflow:hidden;text-overflow:ellipsis}
|
||||
.ausboss-transform-group-body{display:flex;flex-direction:column;gap:7px;padding:8px;box-sizing:border-box}
|
||||
.ausboss-transform-group-body[hidden]{display:none}
|
||||
.ausboss-transform-readout{flex:0 0 auto;margin-top:-3px;color:#c9d0d6;font:11px/14px system-ui;font-variant-numeric:tabular-nums;text-align:center;overflow:hidden;cursor:default;user-select:none}
|
||||
.ausboss-transform-readout{flex:0 0 auto;margin:-5px 0 -3px;color:#c9d0d6;font:11px/14px system-ui;font-variant-numeric:tabular-nums;text-align:center;overflow:hidden;cursor:default;user-select:none}
|
||||
.ausboss-transform-readout:empty{display:none}
|
||||
.ausboss-transform-readout span{color:#8ca8a5}.ausboss-transform-readout b{color:#fff;font-weight:600}
|
||||
.ausboss-transform-readout>div,.ausboss-transform-readout i{display:flex;align-items:center;justify-content:center;gap:3px;height:14px;white-space:nowrap}
|
||||
.ausboss-transform-readout .ausboss-transform-ellipsis{min-width:0;overflow:hidden;text-overflow:ellipsis;color:inherit}
|
||||
.ausboss-transform-readout b,.ausboss-transform-readout button{flex:none}
|
||||
.ausboss-transform-readout i{color:#ffc46b;font-style:normal}
|
||||
.ausboss-transform-readout i:empty{display:none}
|
||||
.ausboss-transform-readout em{color:#8ca8a5;font-style:normal}
|
||||
.ausboss-transform-readout em.note.ausboss-transform-ellipsis{color:#ffc46b}
|
||||
.ausboss-transform-readout button{margin:0 0 0 6px;height:14px;padding:0 6px;border:1px solid #6b5a33;border-radius:3px;background:#2a2417;color:#ffd79a;font:600 9.5px/12px system-ui;cursor:pointer;vertical-align:top}
|
||||
@@ -2320,6 +2323,11 @@ function buildTrim(state) {
|
||||
onTogglePlayback: () => timelineCommand(state, "play"),
|
||||
onSeek: (frame, settled) => scrubTo(state, frame, settled),
|
||||
onCommit: notifyAusbossChange,
|
||||
linkInfo: (name) => linkedNodeInfo(state.node, name),
|
||||
// H3 Prompt Writer reads this to say what H3 makes of the clip.
|
||||
publish: (timing) => { state.node.__ausbossClipTiming = timing; },
|
||||
units: () => clipUnits(),
|
||||
setUnits: (next) => { clipUnits(next); for (const view of state.trimViews) view.sync(); },
|
||||
});
|
||||
state.trimViews.add(control);
|
||||
return control.root;
|
||||
@@ -2340,6 +2348,7 @@ function setPlayhead(state, index) {
|
||||
// also snaps the widgets to the decoded frame.
|
||||
function scrubTo(state, index, settled) {
|
||||
if (state.playing) stopPlayback(state, false);
|
||||
state.pausedAt = null;
|
||||
const frame = setPlayhead(state, index);
|
||||
syncTimelineRange(state);
|
||||
if (settled) { void seekFrame(state); return; }
|
||||
@@ -2400,7 +2409,11 @@ function startPlayback(state) {
|
||||
const trim = [...state.trimViews][0];
|
||||
const { first, last } = trim.window();
|
||||
const head = clampFrame(value(state.node, "frame_index", 0), info);
|
||||
let next = head < first || head >= last ? first : head + 1;
|
||||
// Play shows the clip from its start, so what plays is what the run
|
||||
// keeps; only a resume right after a pause inside the clip carries on.
|
||||
const resume = state.pausedAt != null && state.pausedAt === head && head >= first && head < last;
|
||||
let next = resume ? head + 1 : first;
|
||||
state.pausedAt = null;
|
||||
state.playing = true;
|
||||
syncTimelineRange(state);
|
||||
const session = ++state.playbackSession;
|
||||
@@ -2421,6 +2434,7 @@ function startPlayback(state) {
|
||||
function stopPlayback(state, settle = true) {
|
||||
const wasPlaying = Boolean(state.playing);
|
||||
state.playing = false;
|
||||
if (wasPlaying) state.pausedAt = clampFrame(value(state.node, "frame_index", 0), clipInfo(state.metadata));
|
||||
state.playbackSession = (state.playbackSession || 0) + 1; // orphan in-flight ticks
|
||||
if (state.playbackTimer) clearTimeout(state.playbackTimer);
|
||||
state.playbackTimer = null;
|
||||
@@ -3094,6 +3108,68 @@ function drawRotateGlyph(context, x, y, radius, color) {
|
||||
// node itself can still drag, and pointercancel plus mouseleave both end a
|
||||
// gesture in case the capture was refused.
|
||||
function attachStageHandlers(state, canvas, signal) {
|
||||
if (state.kind === "image") {
|
||||
canvas.addEventListener("contextmenu", (event) => {
|
||||
event.preventDefault(); event.stopPropagation();
|
||||
if (!state.image) return;
|
||||
const picture = async () => {
|
||||
let source = state.image;
|
||||
const sourceWidth = state.sourceWidth, sourceHeight = state.sourceHeight;
|
||||
const current = values(state.node), resize = resizeRequest(state.node);
|
||||
const includeEdits = settings().copy_edits;
|
||||
if (includeEdits && maskLayer(state)) {
|
||||
if (value(state.node, "painted_area", "fill") === "keep picture") {
|
||||
// The browser loses colours under fully transparent pixels. Core's
|
||||
// RGB view restores the stored picture under MaskEditor strokes.
|
||||
source = new Image();
|
||||
source.src = api.apiURL(`/view?${mediaViewQuery(value(state.node, "image", ""))}&channel=rgb`);
|
||||
await source.decode();
|
||||
} else {
|
||||
// The stage's mask layer is capped at 2048px. Copy from the source
|
||||
// at full size, using the same rule for which pixels stay picture.
|
||||
const full = document.createElement("canvas");
|
||||
full.width = sourceWidth; full.height = sourceHeight;
|
||||
const context = full.getContext("2d");
|
||||
context.drawImage(source, 0, 0);
|
||||
const pixels = context.getImageData(0, 0, full.width, full.height);
|
||||
const map = seeThroughMap(pixels.data);
|
||||
if (map) for (let i = 0; i < map.length; i++) pixels.data[i * 4 + 3] = map[i] ? 0 : 255;
|
||||
context.putImageData(pixels, 0, 0);
|
||||
source = full;
|
||||
}
|
||||
}
|
||||
return transformPicture({
|
||||
sourceWidth, sourceHeight, values: current, resize, includeEdits,
|
||||
fill: normalizeColor(current.fill_color),
|
||||
drawSource: (context) => context.drawImage(source, -sourceWidth / 2, -sourceHeight / 2, sourceWidth, sourceHeight),
|
||||
});
|
||||
};
|
||||
const report = (error) => showToast({ severity: "error", summary: "Copy image", detail: error?.message ?? String(error) });
|
||||
new globalThis.LiteGraph.ContextMenu([
|
||||
{ content: "Copy image", callback: () => {
|
||||
try {
|
||||
Promise.resolve(copyPicture(picture)).then(
|
||||
() => showToast({ summary: "Copy image", severity: "success", detail: "The picture is on the clipboard." }), report,
|
||||
);
|
||||
} catch (error) { report(error); }
|
||||
} },
|
||||
{ content: "Open image", callback: async () => {
|
||||
try {
|
||||
// A real image retains the browser's native Copy image on plain
|
||||
// HTTP too. A data URL stays usable for as long as the tab is open.
|
||||
const tab = window.open("about:blank", "_blank");
|
||||
if (!tab) throw new Error("Allow pop-ups to open the picture.");
|
||||
tab.opener = null;
|
||||
const image = tab.document.createElement("img");
|
||||
image.src = (await picture()).toDataURL("image/png");
|
||||
image.style.maxWidth = "100%";
|
||||
tab.document.title = "Crop image — right-click to copy";
|
||||
tab.document.body.append(image);
|
||||
} catch (error) { report(error); }
|
||||
} },
|
||||
], { event });
|
||||
}, { signal });
|
||||
}
|
||||
canvas.addEventListener("pointerdown", (event) => pointerDown(state, canvas, event), { signal });
|
||||
canvas.addEventListener("pointermove", (event) => pointerMove(state, canvas, event), { signal });
|
||||
canvas.addEventListener("pointerup", (event) => pointerUp(state, canvas, event), { signal });
|
||||
|
||||
@@ -0,0 +1,54 @@
|
||||
import { keptCrop, resolveCrop, resolvePadding, rotatedSize, scaleToMegapixels } from "./transform_geometry.mjs";
|
||||
|
||||
export const COPY_EDITS_SETTING = Object.freeze({
|
||||
key: "copy_edits", label: "Copy with crop, rotation and padding", type: "toggle", default: false,
|
||||
hint: "Off: copy the original picture. On: include crop, rotation, padding and resize. Handles and guides are never copied.",
|
||||
});
|
||||
|
||||
// Render picture pixels into a separate canvas. The editor's canvas contains
|
||||
// handles, hatching and stage margins and must never be a clipboard source.
|
||||
export function transformPicture({ sourceWidth, sourceHeight, values, resize, fill, drawSource, includeEdits = false, createCanvas = () => document.createElement("canvas") }) {
|
||||
const canvas = createCanvas();
|
||||
if (!includeEdits) {
|
||||
canvas.width = sourceWidth; canvas.height = sourceHeight;
|
||||
const context = canvas.getContext("2d");
|
||||
if (!context) throw new Error("The browser could not make the picture.");
|
||||
context.translate(sourceWidth / 2, sourceHeight / 2);
|
||||
drawSource(context);
|
||||
return canvas;
|
||||
}
|
||||
const source = rotatedSize(sourceWidth, sourceHeight, values.rotation_degrees);
|
||||
const whole = resolveCrop(values, source);
|
||||
const padding = resolvePadding(values, whole);
|
||||
const crop = keptCrop(whole, padding);
|
||||
const size = resize
|
||||
? scaleToMegapixels(padding.outputWidth, padding.outputHeight, resize.megapixels, resize.steps)
|
||||
: { width: padding.outputWidth, height: padding.outputHeight };
|
||||
canvas.width = size.width; canvas.height = size.height;
|
||||
const context = canvas.getContext("2d");
|
||||
if (!context) throw new Error("The browser could not make the picture.");
|
||||
context.fillStyle = fill;
|
||||
context.fillRect(0, 0, canvas.width, canvas.height);
|
||||
context.save();
|
||||
context.scale(size.width / padding.outputWidth, size.height / padding.outputHeight);
|
||||
context.beginPath();
|
||||
context.rect(padding.left, padding.top, crop.width, crop.height);
|
||||
context.clip();
|
||||
context.translate(padding.left - crop.x + source.width / 2, padding.top - crop.y + source.height / 2);
|
||||
context.rotate((Number(values.rotation_degrees) || 0) * Math.PI / 180);
|
||||
drawSource(context);
|
||||
context.restore();
|
||||
return canvas;
|
||||
}
|
||||
|
||||
// Start the clipboard write in the click's activation, including in browsers
|
||||
// that require a promised PNG rather than a write after toBlob has finished.
|
||||
export function copyPicture(picture, { clipboard = globalThis.navigator?.clipboard, ClipboardItem = globalThis.ClipboardItem } = {}) {
|
||||
if (!clipboard?.write || !ClipboardItem) throw new Error("Open ComfyUI on localhost or HTTPS to copy directly. Choose Open image, then right-click the picture to copy it instead.");
|
||||
const png = Promise.resolve().then(() => typeof picture === "function" ? picture() : picture).then((canvas) => new Promise((resolve, reject) => {
|
||||
canvas.toBlob((blob) => blob ? resolve(blob) : reject(new Error("The browser could not make the picture.")), "image/png");
|
||||
}));
|
||||
// Also observe encoding failure when write() is refused before it reads png.
|
||||
png.catch(() => {});
|
||||
return clipboard.write([new ClipboardItem({ "image/png": png })]);
|
||||
}
|
||||
+390
-169
@@ -32,7 +32,7 @@ import {
|
||||
framesThrough,
|
||||
keyboardStep,
|
||||
lastFrameFor,
|
||||
latestFirstFor,
|
||||
latestInFor,
|
||||
railZone,
|
||||
setTrimFrame,
|
||||
slideRange,
|
||||
@@ -41,7 +41,7 @@ import {
|
||||
windowSeconds,
|
||||
} from "./timeline_math.mjs";
|
||||
|
||||
const CSS_ID = "ausboss-transform-trim-css-v4";
|
||||
const CSS_ID = "ausboss-transform-trim-css-v5";
|
||||
const HANDLE_HIT_PX = 9;
|
||||
// Connected inputs that set the length themselves, in the backend's order of
|
||||
// precedence: fixed_frames replaces the window, a cap replaces the Length.
|
||||
@@ -101,6 +101,25 @@ function installCss() {
|
||||
.ausboss-transform-trim-reset{flex:none;height:18px;box-sizing:border-box;border:1px solid #2c4d4b;border-radius:4px;background:#1b2627;color:#cadddb;font:600 10px/16px system-ui;cursor:pointer;padding:0 8px;white-space:nowrap}
|
||||
.ausboss-transform-trim-reset:hover{border-color:#8fa3a1;color:#fff}
|
||||
.ausboss-transform-trim-reset:disabled{opacity:.45;cursor:not-allowed}
|
||||
.ausboss-transform-trim-span.is-room{background:repeating-linear-gradient(90deg,rgba(0,180,170,.38) 0 6px,rgba(0,180,170,.1) 6px 12px)}
|
||||
.ausboss-transform-trim-line .ausboss-transform-trim-readout{text-align:left}
|
||||
.ausboss-transform-trim-inside{color:#45d6cb}
|
||||
.ausboss-transform-trim-outside{color:#8ca8a5}
|
||||
.ausboss-transform-trim-clip{flex:1 1 0;min-width:0;color:#45d6cb;font-weight:600;white-space:nowrap;overflow:hidden;text-overflow:ellipsis;font-variant-numeric:tabular-nums}
|
||||
.ausboss-transform-trim-fields{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:8px;min-width:0}
|
||||
.ausboss-transform-trim-field{display:flex;flex-direction:column;gap:3px;min-width:0}
|
||||
.ausboss-transform-trim-field .ausboss-scrub{width:100%;box-sizing:border-box}
|
||||
.ausboss-transform-trim-field-label{color:#cadddb;display:flex;align-items:center;gap:6px;min-width:0}
|
||||
.ausboss-transform-trim-go[hidden]{display:none}
|
||||
.ausboss-transform-trim-sub{color:#8ca8a5;font-size:10px;white-space:nowrap;overflow:hidden;text-overflow:ellipsis;font-variant-numeric:tabular-nums;min-height:12px}
|
||||
.ausboss-transform-trim-link{color:#9fe3dc;font-size:10px;white-space:nowrap;overflow:hidden;text-overflow:ellipsis}
|
||||
.ausboss-transform-trim-link[hidden]{display:none}
|
||||
.ausboss-transform-trim-go{border:0;background:transparent;padding:0;color:#45d6cb;font:600 10px system-ui;text-decoration:underline;cursor:pointer}
|
||||
.ausboss-transform-trim-note{display:flex;align-items:center;gap:6px;flex-wrap:wrap;color:#9fe3dc;font-size:11px;line-height:1.35}
|
||||
.ausboss-transform-trim-note[hidden]{display:none}
|
||||
.ausboss-transform-trim-note>span{flex:1 1 180px;min-width:0}
|
||||
.ausboss-transform-trim-note.warn{background:#3a2a10;box-shadow:inset 3px 0 0 #e0a040;border-radius:6px;padding:6px 8px;color:#ffdca8}
|
||||
.ausboss-transform-trim-note button[hidden]{display:none}
|
||||
`;
|
||||
document.head.append(style);
|
||||
}
|
||||
@@ -113,6 +132,8 @@ function el(tag, className = "", text = "") {
|
||||
}
|
||||
|
||||
const pct = (fraction) => `${(fraction * 100).toFixed(3)}%`;
|
||||
// Seconds for people: at most two decimals (1, 1.5, 2.17).
|
||||
const secondsText = (value) => String(Number((Number(value) || 0).toFixed(2)));
|
||||
|
||||
// options:
|
||||
// get(name, fallback) / set(name, value) widget access
|
||||
@@ -126,7 +147,12 @@ const pct = (fraction) => `${(fraction * 100).toFixed(3)}%`;
|
||||
// (the clip node) or playhead only
|
||||
// onSeek(frame, settled) the playhead moved; settled on release
|
||||
// onCommit() a stored value settled (undo point)
|
||||
export function mountTransformTrim({ get, set, has = () => true, driven = () => false, number = (name, fallback) => get(name, fallback), outputRate, metadata, trim = true, isPlaying, onTogglePlayback, onSeek, onCommit }) {
|
||||
// linkInfo(name) {title, go()} of the node feeding
|
||||
// a linked input, or null
|
||||
// publish(timing) the clip's {frames, startSeconds,
|
||||
// endSeconds, roomFrames}, or null
|
||||
// units() / setUnits(units) "s" (seconds) or "f" (frames)
|
||||
export function mountTransformTrim({ get, set, has = () => true, driven = () => false, number = (name, fallback) => get(name, fallback), outputRate, metadata, trim = true, isPlaying, onTogglePlayback, onSeek, onCommit, linkInfo = () => null, publish, units: getUnits = () => "s", setUnits = () => {}, play: showPlay = true }) {
|
||||
installCss();
|
||||
const root = el("div", "ausboss-transform-trim");
|
||||
const rail = el("div", "ausboss-transform-trim-rail");
|
||||
@@ -144,7 +170,9 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
const play = el("button", "ausboss-transform-trim-play");
|
||||
play.type = "button";
|
||||
play.addEventListener("click", onTogglePlayback);
|
||||
transport.append(play, rail);
|
||||
// A host with its own player (Load Video) leaves out the Play button.
|
||||
if (showPlay) transport.append(play);
|
||||
transport.append(rail);
|
||||
const info = () => clipInfo(metadata());
|
||||
const everyNth = () => Math.max(1, Math.trunc(Number(number("every_nth", get("every_nth", 1))) || 1));
|
||||
const snapRule = () => String(get("frame_snap", "free"));
|
||||
@@ -160,11 +188,10 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
return number(name, get(name, 0)) ?? get(name, 0);
|
||||
};
|
||||
const countAtRunTime = () => COUNT_INPUTS.some((name) => driven(name) && number(name, null) == null);
|
||||
const plan = () => {
|
||||
if (!trim) return null;
|
||||
const clip = info();
|
||||
const outFps = () => (outputRate ? outputRate() : info().fps / everyNth());
|
||||
const planValues = () => {
|
||||
const wired = wiredBy();
|
||||
return clipLengthPlan(clip, {
|
||||
return {
|
||||
start: boundSeconds("start"),
|
||||
end: boundSeconds("end"),
|
||||
everyNth: everyNth(),
|
||||
@@ -173,9 +200,48 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
fixedFrames: driven("fixed_frames") ? 0 : get("fixed_frames", 0),
|
||||
wired,
|
||||
wiredFrames: wired ? number(wired, null) : null,
|
||||
outputFps: outputRate ? outputRate() : clip.fps,
|
||||
outputFps: outputRate ? outputRate() : info().fps,
|
||||
resampled: driven("force_rate"),
|
||||
});
|
||||
};
|
||||
};
|
||||
const plan = () => (trim ? clipLengthPlan(info(), planValues()) : null);
|
||||
// Output frames that fit between IN and the end of the video: the most a
|
||||
// Duration can be from here.
|
||||
const roomFrames = () => {
|
||||
if (!trim) return null;
|
||||
const room = clipLengthPlan(info(), { ...planValues(), maxFrames: 0, fixedFrames: 0, wired: null, wiredFrames: null });
|
||||
return room.frames ?? null;
|
||||
};
|
||||
// Where the clip ends, in source seconds (exclusive), when it is known.
|
||||
const endSecondsOf = (current) => {
|
||||
const clip = info();
|
||||
if (current?.endSeconds != null) return current.endSeconds;
|
||||
if (current?.last == null || !clip.fps) return null;
|
||||
return (current.last + 1) / clip.fps;
|
||||
};
|
||||
const endDriven = () => driven("end_frame") || driven("end_seconds");
|
||||
// The latest IN that keeps the whole Duration. With a connected rate the
|
||||
// count is in output frames, so the source grid would stop IN far too soon
|
||||
// (the 2.25 face stopped a 240-frame Length at IN 7 on a 247-frame clip).
|
||||
const latestIn = (current) => {
|
||||
const clip = info();
|
||||
if (driven("force_rate")) {
|
||||
if (current.latestStart == null) return clip.count - 1;
|
||||
return current.latestStart > 0 ? current.latestStart : clip.count - 1;
|
||||
}
|
||||
return latestInFor(clip, current.requested ?? get("max_frames", 1), everyNth());
|
||||
};
|
||||
// A Duration as long as the whole video only caps it: IN moves freely and
|
||||
// the clip runs to the end.
|
||||
const durationFills = (current) => {
|
||||
if (current.mode !== "length" || current.requested == null) return false;
|
||||
const whole = roomFromStart();
|
||||
return whole != null && current.requested >= whole;
|
||||
};
|
||||
const roomFromStart = () => {
|
||||
if (!trim) return null;
|
||||
const room = clipLengthPlan(info(), { ...planValues(), start: 0, maxFrames: 0, fixedFrames: 0, wired: null, wiredFrames: null });
|
||||
return room.frames ?? null;
|
||||
};
|
||||
const currentWindow = () => {
|
||||
const clip = info();
|
||||
@@ -185,7 +251,8 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
};
|
||||
const playheadFrame = () => clampFrame(get("frame_index", 0), info());
|
||||
const controls = {};
|
||||
let syncLength = () => {};
|
||||
// A short line under the fields after a drag stops at the video's end.
|
||||
let stopNote = "";
|
||||
// IN is locked by a connected bound; OUT by a connected bound in free
|
||||
// trim, and always when a connected input sets the length.
|
||||
const locked = (edge) => {
|
||||
@@ -196,10 +263,10 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
return false;
|
||||
};
|
||||
const lockTip = (edge) => edge === "start"
|
||||
? "The starting frame is supplied by a connected input. Disconnect it to edit IN here."
|
||||
? "The starting frame is supplied by a connected input. Disconnect it to set Start here."
|
||||
: wiredBy()
|
||||
? `The connected ${wiredBy()} input sets the length, so the clip ends where it says. Disconnect it to use OUT or Length.`
|
||||
: "The ending frame is supplied by a connected input. Disconnect it to edit OUT here.";
|
||||
? `The connected ${wiredBy()} input sets the length, so the clip ends where it says. Disconnect it to set Duration or Ends at here.`
|
||||
: "The ending frame is supplied by a connected input. Disconnect it to set Ends at here.";
|
||||
const disableControl = (control, disabled, tip) => {
|
||||
if (!control) return;
|
||||
control.root.style.opacity = disabled ? ".45" : "";
|
||||
@@ -217,7 +284,16 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
const writeEnd = (first, last) => set("end_seconds", windowSeconds(info(), first, last).end_seconds);
|
||||
// An older face's Fixed frames reads as a Length; the first edit makes it
|
||||
// one, keeping the window where the fixed count had placed it.
|
||||
// The face has no modes: a clip is a Start and a Duration. A window saved
|
||||
// as IN -> OUT (Length off) becomes a Duration of the same frames on its
|
||||
// first edit, so the run outputs exactly what it did. A connected end
|
||||
// keeps the window as it is.
|
||||
const adoptLength = (current) => {
|
||||
if (current.mode === "free" && !endDriven() && current.frames != null) {
|
||||
set("max_frames", Math.max(1, current.frames));
|
||||
set("end_seconds", 0);
|
||||
return plan();
|
||||
}
|
||||
if (current.mode !== "fixed") return current;
|
||||
set("fixed_frames", 0);
|
||||
set("max_frames", Math.max(1, current.requested ?? current.frames ?? 1));
|
||||
@@ -235,8 +311,14 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
let parked;
|
||||
if (edge === "start") {
|
||||
if (current.mode === "length") {
|
||||
// A Length drags the whole window, stopping where OUT meets the end.
|
||||
parked = Math.max(0, Math.min(clampFrame(frame, clip), latestFirstFor(clip, current.requested ?? get("max_frames", 1), nth)));
|
||||
// A Duration drags the whole clip, stopping where its end meets the
|
||||
// video's end; a Duration past the video's end only caps it, so IN
|
||||
// moves freely.
|
||||
const latest = latestIn(current);
|
||||
parked = Math.max(0, Math.min(clampFrame(frame, clip), latest));
|
||||
stopNote = clampFrame(frame, clip) > latest
|
||||
? `Start stops at ${secondsText(parked / clip.fps)} s so all ${current.requested} frames fit.`
|
||||
: "";
|
||||
set("end_seconds", 0);
|
||||
} else if (current.mode === "wired" && current.latestStart != null && !countAtRunTime()) {
|
||||
parked = Math.min(clampFrame(frame, clip), current.latestStart);
|
||||
@@ -246,11 +328,18 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
writeStart(parked);
|
||||
} else {
|
||||
const target = Math.max(current.first, clampFrame(frame, clip));
|
||||
const frames = framesThrough(current.first, target, nth, snapRule());
|
||||
stopNote = "";
|
||||
let frames = framesThrough(current.first, target, nth, snapRule());
|
||||
if (driven("force_rate") && outFps() > 0) {
|
||||
// Output frames at the linked rate from IN to the handle's boundary.
|
||||
frames = snapToValid(Math.max(1, Math.round(((target + 1) / clip.fps - boundSeconds("start")) * outFps())), snapRule());
|
||||
const room = roomFrames();
|
||||
if (room != null) frames = Math.min(frames, Math.max(1, room));
|
||||
}
|
||||
if (current.mode === "length") {
|
||||
set("max_frames", frames);
|
||||
set("end_seconds", 0);
|
||||
parked = Math.min(clip.count - 1, lastFrameFor(current.first, frames, nth));
|
||||
parked = plan()?.last ?? Math.min(clip.count - 1, lastFrameFor(current.first, frames, nth));
|
||||
} else {
|
||||
parked = driven("force_rate") ? target : lastFrameFor(current.first, frames, nth);
|
||||
writeEnd(current.first, parked);
|
||||
@@ -262,7 +351,8 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
// The grip slides the kept part: a Length moves with IN (OUT follows), a
|
||||
// free window keeps its span and both edges move.
|
||||
const canSlide = () => {
|
||||
if (!trim || locked("start")) return false;
|
||||
// With a connected rate, IN itself slides the clip at its Duration.
|
||||
if (!trim || locked("start") || driven("force_rate")) return false;
|
||||
const current = plan();
|
||||
if (!current || !["length", "fixed", "free"].includes(current.mode)) return false;
|
||||
if (current.mode === "free" && locked("end")) return false;
|
||||
@@ -284,26 +374,6 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
}
|
||||
moveEdge("start", first, settled);
|
||||
};
|
||||
// Length on keeps what the rail shows now as the count; off turns the
|
||||
// Length into an OUT at the same frame. Either way nothing moves.
|
||||
const setLength = (on) => {
|
||||
const current = plan();
|
||||
if (!current || current.mode === "wired" || (on === (current.mode !== "free"))) return;
|
||||
if (on) {
|
||||
const frames = current.frames ?? framesThrough(current.first, current.windowLast, everyNth(), snapRule());
|
||||
set("max_frames", Math.max(1, frames));
|
||||
set("end_seconds", 0);
|
||||
} else {
|
||||
const last = current.last ?? current.windowLast;
|
||||
writeStart(current.first);
|
||||
writeEnd(current.first, last);
|
||||
set("max_frames", 0);
|
||||
if (!driven("fixed_frames")) set("fixed_frames", 0);
|
||||
}
|
||||
sync();
|
||||
onCommit?.();
|
||||
};
|
||||
|
||||
const readout = el("span", "ausboss-transform-trim-readout");
|
||||
const summary = el("div", "ausboss-transform-trim-summary");
|
||||
|
||||
@@ -311,7 +381,7 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
const clip = info();
|
||||
const playing = isPlaying();
|
||||
play.setAttribute("aria-label", playing ? "Pause" : "Play");
|
||||
play.title = playing ? "Pause on this frame" : trim ? "Play the selected clip" : "Play the video to pick a frame";
|
||||
play.title = playing ? "Pause on this frame" : trim ? "Play the clip from its start" : "Play the video to pick a frame";
|
||||
play.classList.toggle("is-playing", playing);
|
||||
play.disabled = !clip.count;
|
||||
const head = playheadFrame();
|
||||
@@ -319,33 +389,41 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
playhead.style.left = pct(fractionOfFrame(head, clip, "center"));
|
||||
playhead.style.visibility = clip.count ? "" : "hidden";
|
||||
readout.replaceChildren();
|
||||
const current = trim ? plan() : null;
|
||||
if (clip.count) {
|
||||
readout.append(el("b", "", `fr ${head}`), document.createTextNode(` · ${formatTimecode(frameTime(head, clip))}`));
|
||||
readout.title = `Playhead: frame ${head} of ${clip.count} at ${frameTime(head, clip).toFixed(3)} s. Press or drag the rail to scrub; ${trim ? "the trim handles carry it with them." : "this is the frame the node outputs."}`;
|
||||
readout.append(document.createTextNode("Preview "), el("b", "", `${secondsText(frameTime(head, clip))} s`), document.createTextNode(` · src fr ${head}`));
|
||||
if (trim) {
|
||||
const end = countAtRunTime() ? null : current.last;
|
||||
const inside = head >= current.first && (end == null || head <= end);
|
||||
readout.append(el("span", inside ? "ausboss-transform-trim-inside" : "ausboss-transform-trim-outside", inside ? " · in the clip" : " · outside the clip"));
|
||||
}
|
||||
readout.title = `The picture shows frame ${head} of ${clip.count}, ${frameTime(head, clip).toFixed(2)} seconds into the video. Press or drag the bar to look at another frame${trim ? "; Use as start starts the clip here." : "; this is the frame the node outputs."}`;
|
||||
}
|
||||
controls.frame?.set(head);
|
||||
if (!trim) return;
|
||||
const current = plan();
|
||||
const wired = wiredBy();
|
||||
const unknown = countAtRunTime();
|
||||
const hasOut = current.last != null && current.mode !== "wired";
|
||||
const hasOut = current.last != null && current.mode !== "wired" && !unknown;
|
||||
const outFrame = current.last ?? current.windowLast;
|
||||
const startFraction = fractionOfFrame(current.first, clip, "start");
|
||||
const endFraction = fractionOfFrame(outFrame, clip, "end");
|
||||
const endSec = endSecondsOf(current);
|
||||
const endFraction = endSec != null && !unknown
|
||||
? Math.max(startFraction, Math.min(1, endSec * clip.fps / clip.count))
|
||||
: fractionOfFrame(outFrame, clip, "end");
|
||||
// Dim: frames the free window holds but the run drops (every nth, snap).
|
||||
const tailFraction = current.mode === "free" ? fractionOfFrame(current.windowLast, clip, "end") : endFraction;
|
||||
span.style.left = pct(startFraction);
|
||||
span.style.width = pct(current.last == null && current.mode !== "free" ? 0 : Math.max(0, tailFraction - startFraction));
|
||||
span.classList.toggle("is-room", unknown || current.last == null);
|
||||
// A length only the run knows: the dim part is the room the clip has.
|
||||
span.style.width = pct(unknown || current.last == null ? 1 - startFraction : Math.max(0, tailFraction - startFraction));
|
||||
kept.style.left = pct(startFraction);
|
||||
kept.style.width = current.last == null || unknown ? "0%" : pct(Math.max(0, endFraction - startFraction));
|
||||
const fills = lengthFillsClip(clip, current, everyNth());
|
||||
span.title = unknown || current.last == null ? "Room for the clip: its length is set when the graph runs." : "";
|
||||
const railWidth = rail.getBoundingClientRect().width || 0;
|
||||
const keptWidth = Math.max(0, endFraction - startFraction) * railWidth;
|
||||
grip.hidden = !canSlide() || unknown || (railWidth > 0 && keptWidth < 34);
|
||||
grip.style.left = pct((startFraction + endFraction) / 2);
|
||||
grip.title = current.mode === "free"
|
||||
? "Drag to slide the kept part: IN and OUT move together."
|
||||
: "Drag to slide the clip: OUT follows at the Length.";
|
||||
grip.title = "Drag to slide the clip; its Duration stays.";
|
||||
for (const edge of ["start", "end"]) {
|
||||
const handle = handles[edge];
|
||||
const frame = edge === "start" ? current.first : outFrame;
|
||||
@@ -358,10 +436,9 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
handle.style.cursor = isLocked ? "not-allowed" : "";
|
||||
handle.title = isLocked ? lockTip(edge)
|
||||
: edge === "start"
|
||||
? current.mode === "free" ? "Drag IN to trim; Shift + arrow moves one frame."
|
||||
: fills ? "The Length is the whole clip, so IN has nowhere to go. Shorten the Length to move IN."
|
||||
: "Drag IN to move the clip; OUT follows at the Length. Shift + arrow moves one frame."
|
||||
: current.mode === "free" ? "Drag OUT to trim; it lands on the last frame the run keeps." : "Drag OUT to change the Length; IN stays.";
|
||||
? durationFills(current) ? "Drag to start later. The Duration is longer than the video, so the clip runs to its end."
|
||||
: "Drag to move the clip; its Duration stays. Shift + arrow moves one frame."
|
||||
: "Drag to change the Duration; the start stays.";
|
||||
handle.setAttribute("aria-valuemin", "0");
|
||||
const maxFrame = edge === "start" && current.mode === "wired" && !unknown
|
||||
? current.latestStart ?? current.windowLast : Math.max(0, clip.count - 1);
|
||||
@@ -369,16 +446,16 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
handle.setAttribute("aria-valuenow", String(frame));
|
||||
handle.setAttribute("aria-valuetext", `frame ${frame}, ${formatTimecode(frameTime(frame, clip))}`);
|
||||
}
|
||||
controls.start_seconds?.set(current.first);
|
||||
controls.end_seconds?.set(outFrame);
|
||||
disableControl(controls.start_seconds, locked("start"), lockTip("start"));
|
||||
disableControl(controls.end_seconds, locked("end") || !hasOut, lockTip("end"));
|
||||
syncLength(current, wired);
|
||||
if (useStart) {
|
||||
useStart.disabled = !clip.count || locked("start");
|
||||
useStart.title = locked("start") ? lockTip("start") : "Start the clip at the frame on the stage. The Duration stays.";
|
||||
}
|
||||
if (clip.count) syncFields(clip, current, wired, unknown);
|
||||
if (controls.reset) {
|
||||
const free = current.mode === "free";
|
||||
controls.reset.textContent = free ? "Full clip" : "To start";
|
||||
controls.reset.title = free ? "Reset IN/OUT to the full source. Keeps Every nth and Frames for." : "Move IN to the start of the source; the length stays.";
|
||||
controls.reset.disabled = !free && locked("start");
|
||||
const free = current.mode !== "wired";
|
||||
controls.reset.textContent = free ? "Whole video" : "To start";
|
||||
controls.reset.title = free ? `Use the whole video: start at 0 with no Duration limit. Keeps Every nth${has("frame_snap") ? " and Frames for" : ""}.` : "Move the start to the beginning of the video; the linked length stays.";
|
||||
controls.reset.disabled = locked("start") && !free;
|
||||
}
|
||||
controls.every_nth?.set(get("every_nth", 1));
|
||||
disableControl(controls.every_nth, driven("every_nth"), "Every nth is supplied by a connected input.");
|
||||
@@ -386,50 +463,21 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
controls.frame_snap.value = snapRule();
|
||||
const exact = current.mode === "fixed" || (current.mode === "wired" && wired === "fixed_frames");
|
||||
controls.frame_snap.disabled = exact || driven("frame_snap");
|
||||
controls.frame_snap.title = exact ? "An exact fixed_frames count is used as it is; this does not apply." : "Keep a frame count the video model takes: LTX wants 8n+1 (49, 97, 121), Wan 4n+1. OUT and Length step to those counts. Any keeps every frame.";
|
||||
controls.frame_snap.title = exact ? "An exact fixed_frames count is used as it is; this does not apply." : "Keep a frame count the video model takes: LTX wants 8n+1 (49, 97, 121), Wan 4n+1. The Duration steps to those counts. Any keeps every frame.";
|
||||
}
|
||||
syncSummary(clip, current, wired, unknown);
|
||||
syncSummary(clip);
|
||||
};
|
||||
|
||||
const syncSummary = (clip, current, wired, unknown) => {
|
||||
const syncSummary = (clip) => {
|
||||
if (!clip.count) {
|
||||
summary.textContent = "Choose a source to set the clip window";
|
||||
summary.textContent = "Choose a source to set the clip";
|
||||
summary.title = "";
|
||||
return;
|
||||
}
|
||||
const rate = outputRate?.();
|
||||
const fps = `Source ${formatFps(clip.fps)} fps → Output ${rate == null ? "rate at run time" : `${formatFps(rate)} fps`}`;
|
||||
if (current.mode === "wired") {
|
||||
summary.textContent = current.frames == null ? `${fps} · length from ${wired} at run time` : `${fps} · ${current.frames} frames from ${wired}`;
|
||||
summary.title = `The connected ${wired} input sets the length, so only IN is set here. Disconnect it to use OUT or a Length.`;
|
||||
return;
|
||||
}
|
||||
if (unknown || current.frames == null) {
|
||||
summary.textContent = `${fps} · frame count at run time`;
|
||||
summary.title = "IN, OUT and the playhead refer to source frames. A connected input decides the exact frame count when the workflow runs. Connect the fps output to the next video node's fps input.";
|
||||
return;
|
||||
}
|
||||
const from = formatTimecode(frameTime(current.first, clip));
|
||||
const to = formatTimecode((current.last + 1) / clip.fps);
|
||||
if (current.mode === "fixed") {
|
||||
summary.textContent = current.tooLong
|
||||
? `Fixed length needs more than the source's ${clip.duration.toFixed(3)}s`
|
||||
: `${fps} · exactly ${current.frames} frames · ${from} → ${to}`;
|
||||
summary.title = "An older Fixed frames value: an exact count from IN. Any change here turns it into a Length.";
|
||||
return;
|
||||
}
|
||||
const total = current.windowLast - current.first + 1;
|
||||
const count = current.mode === "length"
|
||||
? current.truncated ? `${current.frames} of ${current.requested} frames, source ends` : `${current.frames} frames`
|
||||
: current.frames < total ? `${current.frames} of ${total} frames` : `${current.frames} frames`;
|
||||
summary.textContent = `${fps} · ${count} · ${from} → ${to}`;
|
||||
summary.title = current.mode === "length"
|
||||
? current.truncated
|
||||
? `The source ends ${current.frames} frames after IN, before the Length of ${current.requested}. Move IN earlier or shorten the Length.`
|
||||
: "Length keeps OUT this many output frames after IN; drag IN to move the clip, OUT to change the Length."
|
||||
: current.frames < total
|
||||
? `The window holds ${total} frames; every nth and the snap rule keep ${current.frames}, ending at frame ${current.last}, where OUT sits. The dim part is dropped.`
|
||||
: "Frames the run outputs, at the fps the node reports.";
|
||||
summary.textContent = `Source ${formatFps(clip.fps)} fps · ${clip.count} frames · ends at ${secondsText(clip.duration || clip.count / clip.fps)} s · clip out at ${rate == null ? "the rate set at run time" : `${formatFps(rate)} fps`}`;
|
||||
summary.title = "Start and Ends at are places in the source video. Duration counts the frames the clip sends on, at the output rate.";
|
||||
if (trim) rail.title = [clipLine?.textContent, summary.textContent, readout.textContent].filter(Boolean).join("\n");
|
||||
};
|
||||
|
||||
// --- rail gestures --------------------------------------------------------
|
||||
@@ -531,21 +579,19 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
}
|
||||
|
||||
// --- rows -----------------------------------------------------------------
|
||||
const timeRow = el("div", "ausboss-transform-trim-line");
|
||||
// The clip node: Preview (the playhead, and Use as start), the clip in
|
||||
// words with a Seconds | Frames switch, then Start, Duration and Ends at.
|
||||
// Start and Ends at are positions in the source video; Duration counts the
|
||||
// frames the clip sends on, at the output rate. No modes: dragging IN
|
||||
// slides the clip, dragging OUT or typing a Duration or an end changes the
|
||||
// Duration.
|
||||
const previewRow = el("div", "ausboss-transform-trim-line");
|
||||
let useStart = null;
|
||||
if (trim) {
|
||||
for (const [name, caption, edge] of [["start_seconds", "IN", "start"], ["end_seconds", "OUT", "end"]]) {
|
||||
const label = el("label");
|
||||
label.append(el("span", "ausboss-transform-trim-caption", caption));
|
||||
controls[name] = makeScrubInput({
|
||||
value: 0, min: 0, max: 10000000, step: 1, decimals: 0,
|
||||
title: edge === "start" ? "IN frame: the first frame the run keeps (stored as start_seconds)." : "OUT frame: the last frame the run keeps. With Length on it follows IN; typing it changes the Length.",
|
||||
onChange: (frame) => moveEdge(edge, frame, false),
|
||||
onSettle: () => { const window = currentWindow(); seek(edge === "start" ? window.first : window.last, true); onCommit?.(); },
|
||||
});
|
||||
label.append(controls[name].root);
|
||||
timeRow.append(label);
|
||||
if (edge === "start") timeRow.append(readout);
|
||||
}
|
||||
useStart = el("button", "ausboss-transform-trim-reset", "Use as start");
|
||||
useStart.type = "button";
|
||||
useStart.title = "Start the clip at the frame on the stage. The Duration stays.";
|
||||
useStart.addEventListener("click", () => moveEdge("start", playheadFrame(), true));
|
||||
} else {
|
||||
const label = el("label");
|
||||
label.append(el("span", "ausboss-transform-trim-caption", "FRAME"));
|
||||
@@ -556,67 +602,237 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
onSettle: () => seek(playheadFrame(), true),
|
||||
});
|
||||
label.append(controls.frame.root);
|
||||
timeRow.append(label, readout);
|
||||
previewRow.append(label, readout);
|
||||
}
|
||||
root.append(transport, timeRow);
|
||||
root.append(transport);
|
||||
if (!trim) root.append(previewRow);
|
||||
|
||||
// Length: off, OUT ends the clip; on, a frame count from IN (max_frames,
|
||||
// end left open), so OUT follows IN and the count survives a new source.
|
||||
if (trim && has("max_frames")) {
|
||||
const lengthRow = el("div", "ausboss-transform-trim-line wrap");
|
||||
const label = el("label", "", "Length");
|
||||
const pill = el("div", "ausboss-transform-trim-pill");
|
||||
const off = el("button", "", "off");
|
||||
const on = el("button", "", "on");
|
||||
off.type = on.type = "button";
|
||||
off.title = "OUT ends the clip: drag both handles freely.";
|
||||
on.title = "Set the clip by a frame count from IN: OUT follows IN, and the count stays when you swap the video.";
|
||||
off.addEventListener("click", () => setLength(false));
|
||||
on.addEventListener("click", () => setLength(true));
|
||||
pill.append(off, on);
|
||||
const hint = el("span", "ausboss-transform-trim-hint");
|
||||
const count = makeScrubInput({ value: 1, min: 1, max: 100000, step: 1, decimals: 0, unit: "fr", width: 72,
|
||||
title: "Output frames from IN. Under a snap rule it steps to the counts the model keeps (97, 105, ... for 8n+1).",
|
||||
onChange: (next) => {
|
||||
const current = adoptLength(plan());
|
||||
if (current.mode !== "length") { sync(); return; }
|
||||
const was = current.requested ?? get("max_frames", 1);
|
||||
set("max_frames", snapToValid(next, snapRule(), Math.sign(next - was)));
|
||||
set("end_seconds", 0);
|
||||
sync();
|
||||
},
|
||||
onSettle: () => { const window = currentWindow(); seek(window.last, true); onCommit?.(); },
|
||||
});
|
||||
label.append(pill, count.root);
|
||||
lengthRow.append(label, hint);
|
||||
root.append(lengthRow);
|
||||
syncLength = (current, wired) => {
|
||||
const active = current.mode !== "free";
|
||||
const blocked = current.mode === "wired";
|
||||
off.classList.toggle("on", !active);
|
||||
on.classList.toggle("on", active);
|
||||
off.setAttribute("aria-pressed", String(!active));
|
||||
on.setAttribute("aria-pressed", String(active));
|
||||
pill.classList.toggle("is-disabled", blocked);
|
||||
off.disabled = on.disabled = blocked;
|
||||
pill.title = blocked ? lockTip("end") : "";
|
||||
count.root.style.display = active ? "" : "none";
|
||||
const step = snapStep(snapRule());
|
||||
count.setStep(step, step);
|
||||
count.set(blocked ? (current.requested ?? current.frames ?? 0) : (current.requested ?? get("max_frames", 1)));
|
||||
disableControl(count, blocked, lockTip("end"));
|
||||
const fills = lengthFillsClip(info(), current, everyNth());
|
||||
hint.textContent = blocked ? `from ${wired}` : current.mode === "fixed" ? "exact (older setting)" : fills ? "= the whole clip. Shorten it to move IN" : active ? "" : "OUT ends the clip";
|
||||
hint.classList.toggle("note", fills && !blocked);
|
||||
hint.title = fills && !blocked ? `The Length (${current.requested ?? current.frames} frames) covers the whole clip, so IN and the clip cannot move. The bar still scrubs.` : "";
|
||||
const fields = {};
|
||||
let clipLine = null;
|
||||
let unitPill = null;
|
||||
let noteRow = null;
|
||||
let noteText = null;
|
||||
const noteButtons = [];
|
||||
if (trim) {
|
||||
const clipRow = el("div", "ausboss-transform-trim-line");
|
||||
clipLine = el("span", "ausboss-transform-trim-clip");
|
||||
unitPill = el("div", "ausboss-transform-trim-pill");
|
||||
for (const [key, words, tip] of [["s", "s", "Show times in seconds."], ["f", "fr", "Show times in frames: source frames for Start and Ends at, clip frames for Duration."]]) {
|
||||
const button = el("button", "", words);
|
||||
button.type = "button";
|
||||
button.dataset.units = key;
|
||||
button.title = tip;
|
||||
button.addEventListener("click", () => { setUnits(key); sync(); });
|
||||
unitPill.append(button);
|
||||
}
|
||||
|
||||
const grid = el("div", "ausboss-transform-trim-fields");
|
||||
const toFrames = {
|
||||
start: (value) => (getUnits() === "f" ? Math.round(value) : Math.round(value * info().fps)),
|
||||
};
|
||||
// Output frames for a Duration typed in the current units.
|
||||
const durationFrames = (value) => {
|
||||
if (getUnits() === "f") return Math.round(value);
|
||||
const rate = outFps();
|
||||
return rate > 0 ? Math.round(value * rate) : null;
|
||||
};
|
||||
// Output frames from IN to an end typed in the current units.
|
||||
const framesToEnd = (value) => {
|
||||
const clip = info();
|
||||
const current = plan();
|
||||
const end = getUnits() === "f" ? (Math.round(value) + 1) / clip.fps : value;
|
||||
if (driven("force_rate")) {
|
||||
const rate = outFps();
|
||||
return rate > 0 ? Math.round((end - boundSeconds("start")) * rate) : null;
|
||||
}
|
||||
return framesThrough(current.first, Math.ceil(end * clip.fps - 1e-7) - 1, everyNth(), snapRule());
|
||||
};
|
||||
const setDuration = (frames) => {
|
||||
if (frames == null || !Number.isFinite(frames)) return;
|
||||
const current = adoptLength(plan());
|
||||
if (current.mode !== "length") return;
|
||||
const was = current.requested ?? get("max_frames", 1);
|
||||
set("max_frames", snapToValid(Math.max(1, frames), snapRule(), Math.sign(frames - was)));
|
||||
set("end_seconds", 0);
|
||||
stopNote = "";
|
||||
};
|
||||
const defs = [
|
||||
["start", "Start", "Where the clip starts in the source video."],
|
||||
["duration", "Duration", "How long the clip is. Counted in clip frames at the output rate."],
|
||||
["end", "Ends at", "Where the clip ends in the source video. Typing it changes the Duration."],
|
||||
];
|
||||
for (const [key, caption, tip] of defs) {
|
||||
const cell = el("div", "ausboss-transform-trim-field");
|
||||
const label = el("span", "ausboss-transform-trim-field-label", caption);
|
||||
const go = el("button", "ausboss-transform-trim-go", "linked");
|
||||
go.type = "button";
|
||||
label.append(go);
|
||||
const control = makeScrubInput({
|
||||
value: 0, min: 0, max: 10000000, step: 1, decimals: 2, minDecimals: 0, unit: "s", unitWidth: 34,
|
||||
title: tip,
|
||||
onChange: (value) => {
|
||||
if (key === "start") {
|
||||
if (locked("start")) return;
|
||||
// A typed start is kept even when the Duration no longer fits;
|
||||
// the note below says what fits and offers both fixes.
|
||||
adoptLength(plan());
|
||||
const frame = clampFrame(toFrames.start(value), info());
|
||||
writeStart(frame);
|
||||
stopNote = "";
|
||||
seek(frame, false);
|
||||
return;
|
||||
}
|
||||
if (key === "duration") setDuration(durationFrames(value));
|
||||
else setDuration(framesToEnd(value));
|
||||
sync();
|
||||
},
|
||||
onSettle: () => { const window = currentWindow(); seek(key === "start" ? window.first : window.last, true); onCommit?.(); },
|
||||
});
|
||||
cell.append(label, control.root);
|
||||
grid.append(cell);
|
||||
fields[key] = { control, go, cell, label };
|
||||
}
|
||||
|
||||
noteRow = el("div", "ausboss-transform-trim-note");
|
||||
noteText = el("span");
|
||||
noteRow.append(noteText);
|
||||
for (let index = 0; index < 2; index++) {
|
||||
const button = el("button", "ausboss-transform-trim-reset");
|
||||
button.type = "button";
|
||||
noteButtons.push(button);
|
||||
noteRow.append(button);
|
||||
}
|
||||
root.append(grid, noteRow);
|
||||
}
|
||||
|
||||
// One field's look: its value in the current units, the line under it,
|
||||
// and, when another node sets it, that node's name and a Go to button.
|
||||
const showField = (key, { value, unit, decimals, sub, lockedBy, readOnly, placeholder = "" }) => {
|
||||
const field = fields[key];
|
||||
if (!field) return;
|
||||
field.control.setRange({ decimals, minDecimals: 0, step: decimals ? 0.1 : 1, fineStep: decimals ? 0.01 : 1 });
|
||||
field.control.set(value);
|
||||
if (value == null) field.control.input.value = "";
|
||||
field.control.input.placeholder = placeholder;
|
||||
const unitSlot = field.control.root.querySelector(".ausboss-scrub-unit");
|
||||
if (unitSlot) unitSlot.textContent = unit;
|
||||
const tip = lockedBy ? `Set by ${lockedBy.title}: ${sub}. Click "linked" to go to that node.` : sub;
|
||||
disableControl(field.control, readOnly, tip);
|
||||
field.cell.title = tip;
|
||||
field.go.hidden = !lockedBy;
|
||||
if (lockedBy) {
|
||||
field.go.title = `Set by ${lockedBy.title}. Click to go to that node.`;
|
||||
field.go.onclick = () => lockedBy.go?.();
|
||||
}
|
||||
};
|
||||
const showNote = (text, kind = "", actions = []) => {
|
||||
if (!noteRow) return;
|
||||
noteRow.hidden = !text;
|
||||
noteRow.classList.toggle("warn", kind === "warn");
|
||||
noteText.textContent = text;
|
||||
noteButtons.forEach((button, index) => {
|
||||
const action = actions[index];
|
||||
button.hidden = !action;
|
||||
if (action) { button.textContent = action.label; button.onclick = () => { action.run(); sync(); onCommit?.(); }; }
|
||||
});
|
||||
};
|
||||
|
||||
const syncFields = (clip, current, wired, unknown) => {
|
||||
const seconds = getUnits() !== "f";
|
||||
for (const button of unitPill.children) button.classList.toggle("on", button.dataset.units === (seconds ? "s" : "f"));
|
||||
const rate = outFps();
|
||||
const rateText = rate > 0 ? `${formatFps(rate)} fps` : "the output rate";
|
||||
const first = current.first;
|
||||
const endSec = unknown ? null : endSecondsOf(current);
|
||||
const frames = unknown ? null : current.frames;
|
||||
const startLink = locked("start") ? linkInfo(driven("start_frame") ? "start_frame" : "start_seconds") : null;
|
||||
const lengthLink = current.mode === "wired" && wired ? linkInfo(wired) : null;
|
||||
const endLink = endDriven() && current.mode === "free" ? linkInfo(driven("end_frame") ? "end_frame" : "end_seconds") : null;
|
||||
const room = roomFrames();
|
||||
const fills = durationFills(current);
|
||||
|
||||
showField("start", {
|
||||
value: seconds ? frameTime(first, clip) : first,
|
||||
unit: seconds ? "s" : "fr", decimals: seconds ? 2 : 0,
|
||||
sub: seconds ? `source frame ${first}` : `source frame · ${secondsText(frameTime(first, clip))} s`,
|
||||
lockedBy: startLink, readOnly: locked("start"),
|
||||
});
|
||||
|
||||
// Duration: what was asked for (a Duration longer than what is left
|
||||
// only caps the clip), in clip frames or seconds at the output rate.
|
||||
const asked = current.mode === "wired" ? current.requested ?? current.frames : current.mode === "length" ? current.requested : frames;
|
||||
let durationSub;
|
||||
if (current.mode === "wired" && current.frames == null && current.requested == null) {
|
||||
durationSub = room != null && rate > 0 ? `up to ${secondsText(room / rate)} s (${room} frames) fit after this start` : "set when the graph runs";
|
||||
} else if (current.mode === "wired" && wired !== "fixed_frames" && current.truncated) {
|
||||
durationSub = `the video ends first: ${frames} frames${rate > 0 ? ` (${secondsText(frames / rate)} s)` : ""}`;
|
||||
} else if (current.mode === "wired") {
|
||||
durationSub = wired === "fixed_frames" ? `${asked} clip frames required` : `up to ${asked} clip frames`;
|
||||
} else if (frames == null) {
|
||||
durationSub = "frame count at run time";
|
||||
} else if (current.truncated) {
|
||||
durationSub = `the video ends first: ${frames} frames${rate > 0 ? ` (${secondsText(frames / rate)} s)` : ""}`;
|
||||
} else {
|
||||
durationSub = seconds ? `${frames} clip frames at ${rateText}` : rate > 0 ? `clip frames · ${secondsText(frames / rate)} s at ${rateText}` : "seconds known at run time";
|
||||
}
|
||||
const durationInSeconds = seconds && rate > 0;
|
||||
showField("duration", {
|
||||
value: asked == null ? null : durationInSeconds ? asked / rate : asked,
|
||||
unit: durationInSeconds ? "s" : "fr", decimals: durationInSeconds ? 2 : 0,
|
||||
sub: durationSub, lockedBy: lengthLink,
|
||||
readOnly: current.mode === "wired" || Boolean(endLink) || asked == null,
|
||||
placeholder: current.mode === "wired" && asked == null ? "at run time" : "",
|
||||
});
|
||||
|
||||
const lastSource = endSec == null ? null : Math.min(clip.count - 1, Math.max(first, Math.ceil(endSec * clip.fps - 1e-7) - 1));
|
||||
showField("end", {
|
||||
value: endSec == null ? null : seconds ? endSec : lastSource,
|
||||
unit: seconds ? "s" : "fr", decimals: seconds ? 2 : 0,
|
||||
sub: endSec == null ? "known after the run" : seconds ? `last source frame ${lastSource}` : `last source frame · ${secondsText(endSec)} s`,
|
||||
lockedBy: endLink, readOnly: current.mode === "wired" || Boolean(endLink) || endSec == null,
|
||||
placeholder: endSec == null ? "after Run" : "",
|
||||
});
|
||||
|
||||
// The clip in words.
|
||||
if (endSec == null) {
|
||||
clipLine.textContent = `Clip from ${seconds ? `${secondsText(frameTime(first, clip))} s` : `src fr ${first}`} · length set when the graph runs`;
|
||||
} else {
|
||||
const span = seconds
|
||||
? `Clip ${secondsText(frameTime(first, clip))} → ${secondsText(endSec)} s`
|
||||
: `Clip src fr ${first} → ${lastSource}`;
|
||||
clipLine.textContent = `${span} · ${frames} frames out`;
|
||||
}
|
||||
|
||||
// What needs saying: a start that leaves too little video, a linked
|
||||
// exact count that does not fit, or a drag that stopped at the end.
|
||||
const latest = current.mode === "length" ? latestIn(current) : current.latestStart;
|
||||
if (current.mode === "wired" && wired === "fixed_frames" && current.tooLong) {
|
||||
showNote(`Needs ${asked} frames; the video is shorter.`, "warn");
|
||||
} else if (current.mode === "wired" && wired === "fixed_frames" && latest != null && first > latest && !unknown) {
|
||||
showNote(`Needs ${asked} frames; only ${room ?? "fewer"} fit from here. The run will fail.`, "warn",
|
||||
locked("start") ? [] : [{ label: `Move start to ${secondsText(latest / clip.fps)} s`, run: () => writeStart(latest) }]);
|
||||
} else if (current.mode === "length" && current.truncated && !fills) {
|
||||
const actions = [{ label: `Shorten to ${rate > 0 ? `${secondsText(frames / rate)} s` : `${frames} frames`}`, run: () => { set("max_frames", Math.max(1, frames)); set("end_seconds", 0); } }];
|
||||
if (!locked("start") && latest != null && latest < first) actions.push({ label: `Move start to ${secondsText(latest / clip.fps)} s`, run: () => writeStart(latest) });
|
||||
showNote(`Only ${frames} of ${current.requested} frames fit from here.`, "warn", actions);
|
||||
} else if (current.mode === "wired" && wired !== "fixed_frames" && current.truncated && frames != null) {
|
||||
showNote(`Video ends first: ${frames} of ${asked} frames.`);
|
||||
} else if (current.mode === "length" && fills && current.truncated) {
|
||||
showNote("Runs to the end of the video.");
|
||||
} else if (stopNote) {
|
||||
showNote(stopNote);
|
||||
} else {
|
||||
showNote("");
|
||||
}
|
||||
|
||||
publish?.(endSec == null || frames == null ? null : { frames, startSeconds: frameTime(first, clip), endSeconds: endSec, roomFrames: room });
|
||||
};
|
||||
|
||||
if (trim) {
|
||||
const sampling = el("div", "ausboss-transform-trim-line wrap");
|
||||
const nthLabel = el("label", "", "Every nth");
|
||||
controls.every_nth = makeScrubInput({ value: get("every_nth", 1), min: 1, max: 512, step: 1, decimals: 0,
|
||||
title: "Keep one frame in this many. Output fps is adjusted to keep real-time playback; a Length counts the frames kept.",
|
||||
title: "Keep one frame in this many. Output fps is adjusted to keep real-time playback; the Duration counts the frames kept.",
|
||||
onChange: (next) => { if (!driven("every_nth")) set("every_nth", next); sync(); }, onSettle: onCommit,
|
||||
});
|
||||
nthLabel.append(controls.every_nth.root);
|
||||
@@ -648,15 +864,20 @@ export function mountTransformTrim({ get, set, has = () => true, driven = () =>
|
||||
reset.type = "button";
|
||||
reset.addEventListener("click", () => {
|
||||
const current = plan();
|
||||
if (current.mode === "free") {
|
||||
if (current.mode !== "wired") {
|
||||
// The whole video: no start, no end, no Duration limit.
|
||||
if (!locked("start")) set("start_seconds", 0);
|
||||
if (!locked("end")) set("end_seconds", 0);
|
||||
sync(); onCommit?.();
|
||||
if (!endDriven()) set("end_seconds", 0);
|
||||
set("max_frames", 0);
|
||||
if (!driven("fixed_frames")) set("fixed_frames", 0);
|
||||
stopNote = "";
|
||||
seek(0, true); onCommit?.();
|
||||
} else {
|
||||
moveEdge("start", 0, true);
|
||||
}
|
||||
});
|
||||
footer.append(summary, reset);
|
||||
sampling.append(unitPill);
|
||||
footer.append(useStart, el("span", "spacer"), reset);
|
||||
root.append(sampling, footer);
|
||||
}
|
||||
root.addEventListener("pointerdown", (event) => {
|
||||
|
||||
@@ -10,6 +10,7 @@
|
||||
// panel's hidden storage widget stays hidden whichever renderer is on.
|
||||
|
||||
const STASH = "__ausbossVisibilityStash";
|
||||
const PAINT_KEYS = ["computeSize", "computeLayoutSize", "draw"];
|
||||
|
||||
// Collapse a widget to zero height. Returns true when this call changed
|
||||
// visibility, false when it was already hidden (callers use the return to
|
||||
@@ -27,6 +28,10 @@ export function hideWidget(widget) {
|
||||
computeSize: widget.computeSize,
|
||||
computeLayoutSize: widget.computeLayoutSize,
|
||||
draw: widget.draw,
|
||||
// Which of the three the widget carried itself. One it lacked is deleted
|
||||
// again on show: left behind as `undefined`, the frontend takes it for a
|
||||
// custom-drawn widget and App mode draws an empty row.
|
||||
own: Object.fromEntries(PAINT_KEYS.map((key) => [key, Object.prototype.hasOwnProperty.call(widget, key)])),
|
||||
hidden: widget.hidden,
|
||||
optionsHidden: options.hidden,
|
||||
optionsDescriptor: Object.getOwnPropertyDescriptor(widget, "options"),
|
||||
@@ -61,9 +66,10 @@ export function hideWidget(widget) {
|
||||
export function showWidget(widget) {
|
||||
const stash = widget?.[STASH];
|
||||
if (!stash) return false;
|
||||
widget.computeSize = stash.computeSize;
|
||||
widget.computeLayoutSize = stash.computeLayoutSize;
|
||||
widget.draw = stash.draw;
|
||||
for (const key of PAINT_KEYS) {
|
||||
if (stash.own && !stash.own[key]) delete widget[key];
|
||||
else widget[key] = stash[key];
|
||||
}
|
||||
widget.hidden = stash.hidden;
|
||||
const options = widget.options;
|
||||
try {
|
||||
@@ -84,6 +90,11 @@ export function showWidget(widget) {
|
||||
return true;
|
||||
}
|
||||
|
||||
// True while hideWidget holds the widget hidden.
|
||||
export function isWidgetHidden(widget) {
|
||||
return Boolean(widget?.[STASH]);
|
||||
}
|
||||
|
||||
// Drive a widget toward a target visibility; true when anything changed.
|
||||
export function setWidgetVisible(widget, visible) {
|
||||
return visible ? showWidget(widget) : hideWidget(widget);
|
||||
|
||||
@@ -213,6 +213,19 @@ export function matchInstalled(name, files) {
|
||||
return loose ? { found: true, path: loose } : { found: false, path: null };
|
||||
}
|
||||
|
||||
// The copy of `name` to pick without asking: the exact path when the folder
|
||||
// has it, else the only file of that name in any subfolder. Two copies in
|
||||
// different subfolders give null: which one to use is the user's choice.
|
||||
export function onlyCopy(name, files) {
|
||||
const match = matchInstalled(name, files);
|
||||
if (!match.found) return null;
|
||||
const wanted = str(name, 1000).trim().replace(/\\/g, "/").toLowerCase();
|
||||
if (match.path.toLowerCase() === wanted) return match.path;
|
||||
const base = baseName(name).toLowerCase();
|
||||
const copies = files.filter((file) => baseName(str(file, 1000).replace(/\\/g, "/")).toLowerCase() === base);
|
||||
return copies.length === 1 ? match.path : null;
|
||||
}
|
||||
|
||||
// Node packs present in a graph, from each node type's python_module
|
||||
// ("custom_nodes.ComfyUI-AusBoss"). Core modules ("nodes",
|
||||
// "comfy_extras.*") are skipped: nobody needs to install those. The first
|
||||
|
||||
@@ -11,6 +11,7 @@ import {
|
||||
unionOf,
|
||||
} from "../shared/snap_lines.mjs";
|
||||
import { boundsOf, nodeElements, outlineOf } from "../shared/outline.mjs";
|
||||
import { TILE_MOVE } from "../shared/tile_face.mjs";
|
||||
|
||||
// Snap into line: a node or group being dragged lines up with the nodes and
|
||||
// groups around it, and a teal line shows what it lines up with. A corner
|
||||
@@ -294,6 +295,13 @@ function onPointerMove(event) {
|
||||
});
|
||||
}
|
||||
|
||||
// A frameless tile moves itself and keeps its pointer moves to itself
|
||||
// (js/shared/tile_face.mjs), then reports each move here once placed.
|
||||
function onTileMove(event) {
|
||||
if (!state.before || !event.detail) return;
|
||||
step(event.detail);
|
||||
}
|
||||
|
||||
function end() {
|
||||
state.before = null;
|
||||
state.corner = null;
|
||||
@@ -348,6 +356,7 @@ function install() {
|
||||
window.addEventListener("pointerdown", onPointerDown, true);
|
||||
window.addEventListener("pointerdown", afterPointerDown);
|
||||
window.addEventListener("pointermove", onPointerMove);
|
||||
window.addEventListener(TILE_MOVE, onTileMove);
|
||||
// Capture phase: the classic canvas stops its pointerup from bubbling.
|
||||
window.addEventListener("pointerup", end, true);
|
||||
window.addEventListener("pointercancel", end, true);
|
||||
|
||||
@@ -12,6 +12,7 @@ import { formatWidgetVisibility } from "../shared/save_video_formats.mjs";
|
||||
import { hideInputsInDef } from "../shared/widget_visibility.mjs";
|
||||
import { castChips } from "../shared/h3_cast_chips.mjs";
|
||||
import { clipSetsDuration } from "../shared/h3_duration.mjs";
|
||||
import { loadImagePreview, loadImageStrip } from "../shared/load_image_strip.mjs";
|
||||
import { momentStrip } from "../shared/moment_strip.mjs";
|
||||
import { mountWidgetCard } from "../shared/widget_card.mjs";
|
||||
import { FONT_UI, THEME } from "../shared/theme.mjs";
|
||||
@@ -27,6 +28,9 @@ import { valueToTake, wiredLimits } from "../shared/wired_limits.mjs";
|
||||
const viewUrl = (value) => api.apiURL(`/view?${mediaViewQuery(value)}`);
|
||||
|
||||
const isMode = (name, ...modes) => (values) => modes.includes(values[name]);
|
||||
// Load Image: the choices that read the width and height boxes.
|
||||
const loadImageUsesBox = (values) => ["fit inside", "crop to fill", "crop"].includes(values.size)
|
||||
|| (values.size === "crop to ratio" && values.ratio === "width : height");
|
||||
// Stitch Inpaint: Seam lives in the card's gear menu (shared/stitch_seam.mjs
|
||||
// holds the menu and the rules). The gear in the card's corner opens the
|
||||
// menu; a "blend in" chip beside it shows the mode at a glance and opens the
|
||||
@@ -121,6 +125,21 @@ const CARDS = {
|
||||
{ widget: "interpolation", label: "Filter", kind: "select" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_EditLock: {
|
||||
minWidth: 300,
|
||||
rows: [
|
||||
{ widget: "lock", label: "Lock", kind: "segment",
|
||||
titles: { auto: "Lock the untouched parts; step aside when the whole picture changes", always: "Lock even when the whole picture changes", off: "Do nothing" } },
|
||||
{ widget: "tolerance", label: "Tolerance", decimals: 2, suffix: "×" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_EditLockRestore: {
|
||||
minWidth: 300,
|
||||
rows: [
|
||||
{ widget: "feather", label: "Feather", suffix: "px" },
|
||||
{ widget: "remove_tint", label: "Remove tint" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_ColorMatch: {
|
||||
minWidth: 300,
|
||||
rows: [
|
||||
@@ -281,6 +300,14 @@ const CARDS = {
|
||||
{ widget: "batch_size", label: "Batch" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_FlashVSRUpscale: {
|
||||
minWidth: 340,
|
||||
rows: [
|
||||
{ widget: "model", label: "Model", kind: "select" },
|
||||
{ widget: "scale", label: "Scale", suffix: "×", decimals: 2 },
|
||||
{ widget: "color_match", label: "Keep colors" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_AlignImage: {
|
||||
minWidth: 300,
|
||||
rows: [
|
||||
@@ -331,6 +358,23 @@ const CARDS = {
|
||||
rows: [{ widget: "text", kind: "textarea", placeholder: "Prompt, caption, or shared text", height: 100, grow: true,
|
||||
looks: { big: { height: 62, placeholder: "type here" } } }],
|
||||
},
|
||||
// One line for each run. The three settings under the box say where a run
|
||||
// ends, whether a leading "1." comes off, and how many lines run at most.
|
||||
AUSBOSS_NODES_TextLines: {
|
||||
minWidth: 320,
|
||||
rows: [
|
||||
{ widget: "text", kind: "textarea", placeholder: "One line for each run", height: 100, grow: true },
|
||||
{ widget: "split", label: "Split at", kind: "segment" },
|
||||
{ widget: "remove_numbering", label: "Numbering", onText: "remove", offText: "keep" },
|
||||
{ widget: "limit", label: "Limit" },
|
||||
],
|
||||
},
|
||||
// The seed arrives on a socket (forceInput), so the card is the prompt
|
||||
// alone, with its {a|b} choice sets.
|
||||
AUSBOSS_NODES_PromptVariations: {
|
||||
minWidth: 300,
|
||||
rows: [{ widget: "text", kind: "textarea", placeholder: "A woman in a {red|blue|green} {sweater|jacket}. One choice is picked from each set.", height: 100, grow: true }],
|
||||
},
|
||||
// All the words go in the box, and the strip under it shows when each
|
||||
// moment plays. The pictures, the video, the voice and the song arrive on
|
||||
// sockets, and each picture's job sits under the words.
|
||||
@@ -380,24 +424,50 @@ const CARDS = {
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_LoadVideo: {
|
||||
minWidth: 320, first: true, hide: ["upload"],
|
||||
// Every nth and the frame limit (as Duration) are in the timing card
|
||||
// under the player (js/load_video). The card hides them, so App mode can
|
||||
// still show them as fields.
|
||||
minWidth: 320, first: true, hide: ["upload", "every_nth", "max_frames"],
|
||||
rows: [
|
||||
{ widget: "video", label: "Source", kind: "select", preview: { kind: "video", url: viewUrl },
|
||||
button: { text: PANEL_LABELS.load, title: "Upload a video into ComfyUI's input folder", onClick: (node) => node.widgets?.find((w) => w.name === "upload")?.callback?.() } },
|
||||
{ widget: "custom_width", label: "Width", suffix: "px", title: "0 keeps the source width; set one side only to keep the aspect." },
|
||||
{ widget: "custom_height", label: "Height", suffix: "px", title: "0 keeps the source height; set one side only to keep the aspect." },
|
||||
{ widget: "every_nth", label: "Every nth" },
|
||||
{ widget: "max_frames", label: "Limit" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_LoadAudio: {
|
||||
// The wave, IN and OUT live in the panel under the card (js/load_audio).
|
||||
minWidth: 280, first: true, hide: ["upload"],
|
||||
// The wave and Start, Duration and Ends at live in the panel under the
|
||||
// card (js/load_audio); max_seconds is the Duration there. The card
|
||||
// hides it, so App mode can still show it as a field.
|
||||
minWidth: 280, first: true, hide: ["upload", "max_seconds"],
|
||||
rows: [
|
||||
{ widget: "audio", label: "Source", kind: "select",
|
||||
button: { text: PANEL_LABELS.load, title: "Upload a sound file into ComfyUI's input folder", onClick: (node) => node.widgets?.find((w) => w.name === "upload")?.callback?.() } },
|
||||
{ widget: "max_seconds", label: "Max length", suffix: "s", decimals: 1, trimZeros: true,
|
||||
title: "The longest the sound may be. Anything past it is cut off and shows in red. 0 means no limit." },
|
||||
],
|
||||
},
|
||||
// Load Image: the file, then where it is in the folder and the size it
|
||||
// goes in and comes out at, then the size settings. ComfyUI's own picture
|
||||
// preview sits under the card, so its mask editor and copy and paste work
|
||||
// as they do on the stock loader.
|
||||
AUSBOSS_NODES_LoadImage: {
|
||||
minWidth: 320, first: true, hide: ["upload"],
|
||||
rows: [
|
||||
{ widget: "image", label: "Source", kind: "select", preview: { kind: "image", url: viewUrl },
|
||||
button: { text: PANEL_LABELS.load, title: "Upload a picture into ComfyUI's input folder", onClick: (node) => node.widgets?.find((w) => w.name === "upload")?.callback?.() } },
|
||||
{ panel: loadImageStrip, height: 26 },
|
||||
{ panel: loadImagePreview, height: 72 },
|
||||
{ widget: "size", label: "Size", kind: "select" },
|
||||
{ widget: "megapixels", label: "Megapixels", suffix: "MP", when: isMode("size", "megapixels") },
|
||||
{ widget: "longest_side", label: "Longest side", suffix: "px", when: isMode("size", "longest side") },
|
||||
{ widget: "scale", label: "Scale", suffix: "\u00d7", when: isMode("size", "scale") },
|
||||
{ widget: "ratio", label: "Ratio", kind: "select", when: isMode("size", "crop to ratio") },
|
||||
{ widget: "width", label: "Width", suffix: "px", when: loadImageUsesBox },
|
||||
{ widget: "height", label: "Height", suffix: "px", when: loadImageUsesBox },
|
||||
{ widget: "crop_from", label: "Keep", kind: "select", when: isMode("size", "crop to fill", "crop", "crop to ratio") },
|
||||
{ widget: "multiple", label: "Round to", snap: true, step: 8, when: (values) => values.size !== "off" },
|
||||
{ group: "more", label: "More" },
|
||||
{ widget: "method", label: "Method", kind: "select", group: "more" },
|
||||
{ widget: "enlarge", label: "Enlarge", group: "more" },
|
||||
],
|
||||
},
|
||||
AUSBOSS_NODES_LoadImagePad: {
|
||||
|
||||
+63
-19
@@ -15,10 +15,11 @@ import { app } from "../../../scripts/app.js";
|
||||
import { setTileFace, TILE_FACE_CSS } from "../shared/tile_face.mjs";
|
||||
import { NOTE_TYPES, serializeNoteIdentity } from "../shared/note_identity.mjs";
|
||||
import { compactAction } from "../shared/panel_controls.mjs";
|
||||
import { BRAND, chainCallback, keepDomWidgetWidthAuto, notifyAusbossChange } from "../shared/index.mjs";
|
||||
import { BRAND, chainCallback, keepDomWidgetWidthAuto, notifyAusbossChange, showToast } from "../shared/index.mjs";
|
||||
import { collectGraphNodes } from "../shared/appearance.mjs";
|
||||
import { copyToClipboard } from "../shared/clipboard.mjs";
|
||||
import { confirmDiscard } from "../shared/discard_prompt.mjs";
|
||||
import { closeOnBackdropClick } from "../shared/dialog_backdrop.mjs";
|
||||
import { WIDGET_FRAME, fillNodeHeight, holdNodeMinHeight } from "../shared/panel_layout.mjs";
|
||||
import { hideInputsInDef, hideWidget } from "../shared/widget_visibility.mjs";
|
||||
import { suppressCoreImagePreview } from "../shared/core_preview.mjs";
|
||||
@@ -37,6 +38,7 @@ import {
|
||||
noteIsEmpty,
|
||||
noteShowsTileFace,
|
||||
normalizeNote,
|
||||
onlyCopy,
|
||||
packsFromGraph,
|
||||
parseMarkdown,
|
||||
rowFromCells,
|
||||
@@ -45,8 +47,36 @@ import {
|
||||
} from "../shared/workflow_note.mjs";
|
||||
import { comboValues, commitWidgetValue } from "../shared/widget_card_math.mjs";
|
||||
import { parseRows, serializeRows } from "../shared/lora_stack.mjs";
|
||||
import { isFileWidget } from "../shared/lora_file_widgets.mjs";
|
||||
|
||||
const CSS_ID = "ausboss-workflow-note-css";
|
||||
const PICK_FOUND_SETTING = "AusBoss.Models.PickFoundCopy";
|
||||
|
||||
function pickFoundCopies() {
|
||||
return app.ui?.settings?.getSettingValue?.(PICK_FOUND_SETTING) !== false;
|
||||
}
|
||||
|
||||
// Point the loaders (and LoRA Loader rows) that ask for a file by another
|
||||
// spelling at the copy the note found. Returns how many it changed.
|
||||
function useFoundCopy(path, loaders, loraRows, files) {
|
||||
for (const entry of loaders) {
|
||||
commitWidgetValue(entry.node, entry.widget, listedSpelling(path, entry.options), app.canvas);
|
||||
entry.node.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
const stacks = new Map();
|
||||
for (const entry of loraRows) stacks.set(entry.widget, [...(stacks.get(entry.widget) ?? []), entry]);
|
||||
for (const [widget, entries] of stacks) {
|
||||
const rows = parseRows(widget.value);
|
||||
for (const entry of entries) {
|
||||
if (rows[entry.index]) rows[entry.index] = { ...rows[entry.index], name: listedSpelling(path, files) };
|
||||
}
|
||||
commitWidgetValue(entries[0].node, widget, serializeRows(rows), app.canvas);
|
||||
entries[0].node.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
app.graph?.setDirtyCanvas?.(true, true);
|
||||
notifyAusbossChange();
|
||||
return loaders.length + loraRows.length;
|
||||
}
|
||||
const WIDGET_NAME = "ausboss_workflow_note_panel";
|
||||
const PANEL_MIN_WIDTH = 300;
|
||||
const CARD_MIN_HEIGHT = 160;
|
||||
@@ -416,22 +446,7 @@ function modelRow(state, row) {
|
||||
use.title = `Point ${names} at ${path}.`;
|
||||
use.addEventListener("pointerdown", (event) => event.stopPropagation());
|
||||
use.addEventListener("click", () => {
|
||||
for (const entry of loaders) {
|
||||
commitWidgetValue(entry.node, entry.widget, listedSpelling(path, entry.options), app.canvas);
|
||||
entry.node.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
const stacks = new Map();
|
||||
for (const entry of loraRows) stacks.set(entry.widget, [...(stacks.get(entry.widget) ?? []), entry]);
|
||||
for (const [widget, entries] of stacks) {
|
||||
const rows = parseRows(widget.value);
|
||||
for (const entry of entries) {
|
||||
if (rows[entry.index]) rows[entry.index] = { ...rows[entry.index], name: listedSpelling(path, files) };
|
||||
}
|
||||
commitWidgetValue(entries[0].node, widget, serializeRows(rows), app.canvas);
|
||||
entries[0].node.setDirtyCanvas?.(true, true);
|
||||
}
|
||||
app.graph?.setDirtyCanvas?.(true, true);
|
||||
notifyAusbossChange();
|
||||
useFoundCopy(path, loaders, loraRows, files);
|
||||
checkModels(state, false);
|
||||
});
|
||||
action.append(pill, use);
|
||||
@@ -551,6 +566,7 @@ function render(state) {
|
||||
if (note.author) header.append(el("p", "", `by ${note.author}`));
|
||||
dialog.append(header);
|
||||
renderSections({ note, rows: [] }, dialog);
|
||||
closeOnBackdropClick(dialog);
|
||||
document.body.append(dialog);
|
||||
state.reader = dialog;
|
||||
dialog.addEventListener("close", () => {
|
||||
@@ -626,6 +642,7 @@ function renderSections(state, scroll) {
|
||||
async function checkModels(state, fresh) {
|
||||
const generation = (state.checkGeneration = (state.checkGeneration || 0) + 1);
|
||||
const rows = state.rows;
|
||||
const picked = [];
|
||||
for (const entry of rows) setDot(entry.dot, "checking", "checking…");
|
||||
const byDir = new Map();
|
||||
for (const entry of rows) {
|
||||
@@ -664,7 +681,14 @@ async function checkModels(state, fresh) {
|
||||
const elsewhere = match.found && subfolderOf(match.path);
|
||||
const loaders = elsewhere ? loadersNeeding(entry.row.name, match.path, graphListWidgets()) : [];
|
||||
const loraRows = elsewhere ? loraRowsNeeding(entry.row.name, match.path, graphLoraLoaders(), files) : [];
|
||||
if (match.found && (loaders.length || loraRows.length)) {
|
||||
if (match.found && (loaders.length || loraRows.length) && pickFoundCopies() && onlyCopy(entry.row.name, files) === match.path) {
|
||||
// One copy, so nothing to choose: the loaders take it without a
|
||||
// click (App mode shows no Guide to click in).
|
||||
useFoundCopy(match.path, loaders, loraRows, files);
|
||||
picked.push(match.path);
|
||||
setDot(entry.dot, "ok", `Found in ${subfolderOf(match.path)}/ and picked: ${match.path}`);
|
||||
entry.showInstalled(match.path);
|
||||
} else if (match.found && (loaders.length || loraRows.length)) {
|
||||
setDot(entry.dot, "ok", `Found in ${subfolderOf(match.path)}/: ${match.path}`);
|
||||
entry.showElsewhere(match.path, loaders, loraRows, files);
|
||||
} else if (match.found) {
|
||||
@@ -676,6 +700,14 @@ async function checkModels(state, fresh) {
|
||||
}
|
||||
}
|
||||
}));
|
||||
if (picked.length && generation === state.checkGeneration) {
|
||||
const folders = [...new Set(picked.map(subfolderOf))].map((folder) => `${folder}/`).join(", ");
|
||||
showToast({
|
||||
summary: "Models found",
|
||||
detail: `${picked.length === 1 ? "1 model file" : `${picked.length} model files`} found in ${folders} and picked for this workflow.`,
|
||||
life: 4000,
|
||||
});
|
||||
}
|
||||
if (state.domWidget && generation === state.checkGeneration) fitNote(state);
|
||||
}
|
||||
|
||||
@@ -754,7 +786,7 @@ function graphListWidgets() {
|
||||
seen.add(graph);
|
||||
for (const node of graph._nodes ?? graph.nodes ?? []) {
|
||||
for (const widget of node.widgets ?? []) {
|
||||
if (widget?.type !== "combo" || typeof widget.value !== "string") continue;
|
||||
if (widget?.type !== "combo" || typeof widget.value !== "string" || isFileWidget(widget)) continue;
|
||||
out.push({ node, widget, value: widget.value, options: comboValues(widget) });
|
||||
}
|
||||
if (node.subgraph) walk(node.subgraph);
|
||||
@@ -1176,6 +1208,18 @@ function buildPanel(node) {
|
||||
|
||||
app.registerExtension({
|
||||
name: "AusBoss.WorkflowNote",
|
||||
settings: [
|
||||
{
|
||||
id: PICK_FOUND_SETTING,
|
||||
name: "Pick model files found in a subfolder",
|
||||
type: "boolean",
|
||||
defaultValue: true,
|
||||
category: ["🆎 AusBoss", "Models", "Pick found files"],
|
||||
tooltip:
|
||||
"When a workflow's Guide finds a model you already have in a subfolder, the workflow switches to that "
|
||||
+ "file by itself. Off: the Guide shows a Use it button instead.",
|
||||
},
|
||||
],
|
||||
beforeRegisterNodeDef(nodeType, nodeData) {
|
||||
if (!NOTE_TYPES.has(nodeData?.name)) return;
|
||||
hideInputsInDef(nodeData, ["note"]);
|
||||
|
||||
@@ -172,6 +172,64 @@
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_EditLock": {
|
||||
"display_name": "编辑锁定 🆎",
|
||||
"description": "让图片里编辑没有改动的部分保持在原位,颜色也不变。编辑模型会把整张图重画一遍,没改的部分常常略微放大、偏移或偏色。把这个节点放在模型和采样器之间,并接上你要编辑的图片。不需要遮罩:它在出图过程中看出编辑改了哪里。如果编辑改变的是整张图(新风格、新光线、新天气),它会自动让开。解码后再接上编辑锁定还原 🆎,未改动的部分就能换回你原图的像素。",
|
||||
"inputs": {
|
||||
"model": {
|
||||
"name": "模型"
|
||||
},
|
||||
"vae": {
|
||||
"name": "vae"
|
||||
},
|
||||
"image": {
|
||||
"name": "图像"
|
||||
},
|
||||
"lock": {
|
||||
"name": "锁定"
|
||||
},
|
||||
"tolerance": {
|
||||
"name": "容差"
|
||||
},
|
||||
"mask": {
|
||||
"name": "遮罩"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "模型"
|
||||
},
|
||||
"1": {
|
||||
"name": "编辑锁"
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_EditLockRestore": {
|
||||
"display_name": "编辑锁定还原 🆎",
|
||||
"description": "把编辑没有改动的部分换回你原图自己的像素。放在 VAE 解码之后,并接上编辑锁定 🆎 的 edit_lock 输出。未改动的部分与原图逐像素一致,只有编辑的部分是新的。如果锁定让开了,图片原样通过。",
|
||||
"inputs": {
|
||||
"image": {
|
||||
"name": "图像"
|
||||
},
|
||||
"edit_lock": {
|
||||
"name": "编辑锁"
|
||||
},
|
||||
"feather": {
|
||||
"name": "羽化"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "图像"
|
||||
},
|
||||
"1": {
|
||||
"name": "编辑遮罩"
|
||||
},
|
||||
"2": {
|
||||
"name": "报告"
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_Float": {
|
||||
"display_name": "浮点数 🆎",
|
||||
"description": "一个独立的 FLOAT 值。在多个节点之间共用同一个强度、缩放或 CFG,只需在一处修改。",
|
||||
@@ -593,6 +651,71 @@
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_LoadImage": {
|
||||
"display_name": "加载图像 🆎",
|
||||
"description": "在同一个节点里加载图片并设定尺寸。可以选择文件、上传,或用箭头逐张浏览输入文件夹。尺寸可以关闭,也可以按百万像素、最长边、缩放倍数、放入一个框内,或裁剪填满一个框来调整。也可以不缩放,直接从图片中裁出一块或裁成某个比例,并选择保留哪一边。运行之前,节点就会显示输入和输出的图片及两者的尺寸。",
|
||||
"inputs": {
|
||||
"image": {
|
||||
"name": "图像"
|
||||
},
|
||||
"size": {
|
||||
"name": "尺寸"
|
||||
},
|
||||
"megapixels": {
|
||||
"name": "百万像素"
|
||||
},
|
||||
"longest_side": {
|
||||
"name": "最长边"
|
||||
},
|
||||
"scale": {
|
||||
"name": "缩放倍数"
|
||||
},
|
||||
"width": {
|
||||
"name": "宽度"
|
||||
},
|
||||
"height": {
|
||||
"name": "高度"
|
||||
},
|
||||
"multiple": {
|
||||
"name": "取整倍数"
|
||||
},
|
||||
"method": {
|
||||
"name": "插值"
|
||||
},
|
||||
"enlarge": {
|
||||
"name": "放大"
|
||||
},
|
||||
"crop_from": {
|
||||
"name": "保留部分"
|
||||
},
|
||||
"ratio": {
|
||||
"name": "比例"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "图像"
|
||||
},
|
||||
"1": {
|
||||
"name": "遮罩"
|
||||
},
|
||||
"2": {
|
||||
"name": "宽度"
|
||||
},
|
||||
"3": {
|
||||
"name": "高度"
|
||||
},
|
||||
"4": {
|
||||
"name": "文件名"
|
||||
},
|
||||
"5": {
|
||||
"name": "原始宽度"
|
||||
},
|
||||
"6": {
|
||||
"name": "原始高度"
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_LoadImagePad": {
|
||||
"display_name": "加载图像 + 扩边 🆎",
|
||||
"description": "在一个节点里加载图像并扩边成扩图画布。拖动节点上画布的任意一边,直观地设置每边的扩边量;遮罩正好覆盖新扩出的部分,可选择向内羽化跨过接缝。画布会取整到整数倍,但不会给你没扩的边加条带;设置百万像素目标时会先缩放原图,让遮罩接缝保持清晰。",
|
||||
@@ -1052,6 +1175,9 @@
|
||||
},
|
||||
"step_preview": {
|
||||
"name": "步骤预览"
|
||||
},
|
||||
"autosave": {
|
||||
"name": "自动保存"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
@@ -1240,6 +1366,35 @@
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_TextLines": {
|
||||
"display_name": "逐行文本 🆎",
|
||||
"description": "文本的每一行都会让后面的节点运行一次。输入五条提示词,每行一条,运行一次就得到五张图。空行会跳过,行首的序号或项目符号可以去掉。",
|
||||
"inputs": {
|
||||
"text": {
|
||||
"name": "文本"
|
||||
},
|
||||
"split": {
|
||||
"name": "分隔方式"
|
||||
},
|
||||
"remove_numbering": {
|
||||
"name": "去掉序号"
|
||||
},
|
||||
"limit": {
|
||||
"name": "上限"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "行"
|
||||
},
|
||||
"1": {
|
||||
"name": "序号"
|
||||
},
|
||||
"2": {
|
||||
"name": "数量"
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_TiledUpscale": {
|
||||
"display_name": "分块放大 🆎",
|
||||
"description": "以百万像素设置成品尺寸,并把图片切成模型可以逐块重画的图块。一块就能装下的图片保持完整。把图块送进采样器,再和 stitcher 一起送进分块放大拼合 🆎。小图不会被重画到远超其细节的程度:最多重画到原百万像素的 max_growth 倍,再用普通方式放大。report 会说明对你的图片做了什么。",
|
||||
@@ -1591,5 +1746,48 @@
|
||||
"AUSBOSS_NODES_WorkflowNoteTile": {
|
||||
"display_name": "工作流说明卡片 🆎",
|
||||
"description": "点击打开完整说明的小按钮。点击铅笔编辑,拖动角落调整大小。旧版节点包会将它显示为缺失节点,请更新 ComfyUI-AusBoss。"
|
||||
},
|
||||
"AUSBOSS_NODES_FlashVSRUpscale": {
|
||||
"display_name": "FlashVSR 视频放大 🆎",
|
||||
"description": "用 FlashVSR 模型把视频放大并变清晰。选好模型文件,设定放大倍数,然后运行。节点每次只读取几帧,所以长视频和短视频占用的显存一样多,每一帧都会返回。接上 Wan 2.1 VAE 可以得到更慢但更细致的结果。",
|
||||
"inputs": {
|
||||
"frames": {
|
||||
"name": "帧"
|
||||
},
|
||||
"model": {
|
||||
"name": "模型"
|
||||
},
|
||||
"scale": {
|
||||
"name": "放大倍数"
|
||||
},
|
||||
"color_match": {
|
||||
"name": "保留颜色"
|
||||
},
|
||||
"vae": {
|
||||
"name": "vae"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "帧"
|
||||
}
|
||||
}
|
||||
},
|
||||
"AUSBOSS_NODES_PromptVariations": {
|
||||
"display_name": "提示词变体 🆎",
|
||||
"description": "每次运行都会变化的提示词框。把选项写在花括号里,例如 {红色|蓝色|绿色} 的 {毛衣|夹克},节点会从每组里选一个。种子决定选哪个:同一个种子得到同一个提示词,每次换新种子就得到新的组合。没有花括号的文字原样输出。",
|
||||
"inputs": {
|
||||
"text": {
|
||||
"name": "文本"
|
||||
},
|
||||
"seed": {
|
||||
"name": "种子"
|
||||
}
|
||||
},
|
||||
"outputs": {
|
||||
"0": {
|
||||
"name": "提示词"
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -0,0 +1,455 @@
|
||||
"""Edit Lock: keep what an edit did not change. Pure torch, no ComfyUI imports.
|
||||
|
||||
At every sampling step an edit model predicts the finished picture. Where
|
||||
that prediction sits close to the original, the edit is not happening
|
||||
there, so the original is put back and the model carries on from the real
|
||||
picture. This is the target TailSteer trains a model toward
|
||||
(original where nothing changed, the model's edit elsewhere), applied while
|
||||
sampling instead of through training.
|
||||
|
||||
What counts as changed is measured against the quiet part of the picture:
|
||||
the tenth of it where the prediction differs least, which a local edit
|
||||
leaves near the level of the model's own redraw noise. The node watches the
|
||||
first steps, where it only undoes the model's zoom and shift, and decides
|
||||
once the picture has formed. An edit of the whole picture (a style, the
|
||||
light) has no quiet part: the first look checks for that and the lock steps
|
||||
aside. A soft change over the whole picture (rain, snow, haze) has no clear
|
||||
peak: the decision checks for that, and the picture is only held in place.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
|
||||
LOCK_MODES = ("auto", "always", "off")
|
||||
|
||||
# A spot counts as changed when it differs from the original by STRONG times
|
||||
# the quiet level, or by WEAK times while it touches a spot that does. So a
|
||||
# faint change next to a clear one belongs to the edit, and a faint change
|
||||
# alone is redraw noise. Set against what 14 plain MiniMax-H3-Image edits
|
||||
# changed in the end: this frees about 80% of it at the first look and 3%
|
||||
# of what stayed.
|
||||
WEAK = 2.0
|
||||
STRONG = 3.0
|
||||
# The model often draws the whole picture a little larger or shifted. That
|
||||
# is measured and discounted before anything is compared, up to these
|
||||
# limits; a bigger move is treated as part of the edit.
|
||||
# The node watches the start of a run and decides once the noise is down to
|
||||
# this share of where it began. The first prediction is a blur: rain, snow,
|
||||
# smoke and the light a new hair color throws on a face only show a few
|
||||
# steps in. While it watches it only undoes the zoom and shift.
|
||||
DECIDE_AT = 0.8
|
||||
# A local edit has a peak: somewhere the picture clearly changes. Rain, snow
|
||||
# and haze change the whole picture a little and no part clearly. A lock
|
||||
# would hold them back, or let them through in patches where it hands a wide
|
||||
# part of the picture to the edit. So two things are read at the decision,
|
||||
# in units of the quiet level: the strongest spot, and the strongest
|
||||
# hundredth when the lock would free SOFT_WIDE of the picture or more. On
|
||||
# recorded MiniMax-H3-Image runs the strongest spot read 4.7 or more for
|
||||
# every local edit, glasses and makeup included, and 3.7 or less for light
|
||||
# rain, fog and snow; with a wide part freed, the strongest hundredth read
|
||||
# 4.4 or more for local edits and 3.6 or less for rain and snow.
|
||||
SOFT_WIDE = 0.10
|
||||
SOFT_PEAK = 4.0
|
||||
# Kept this much or more: the report adds what to do if the edit is missing.
|
||||
# A soft effect over the whole picture (rain, snow, smoke) looks like a small
|
||||
# edit at the first look, and under the lock the model then drops it.
|
||||
NEARLY_ALL = 0.95
|
||||
MAX_ZOOM = 0.05
|
||||
MAX_SHIFT = 4.0
|
||||
# The whole picture is changing when even its calmest fifth sits this far
|
||||
# from the original at the first look, in units of the original's own
|
||||
# spread. Measured on MiniMax-H3-Image: 28 local edits at 1, 4 and 16 MP came
|
||||
# in at 0.22 or less (zoomed ones included), 12 whole-picture edits (styles,
|
||||
# light, weather) at 0.27 or more.
|
||||
WHOLE_PICTURE = 0.243
|
||||
# Blur and growth of the change map, as a share of the short side.
|
||||
REACH = 1.0 / 54.0
|
||||
# A result this much larger or smaller than the picture is another picture.
|
||||
SIZE_SLACK = 0.04
|
||||
|
||||
|
||||
def box_blur(x: torch.Tensor, radius: int) -> torch.Tensor:
|
||||
if radius <= 0:
|
||||
return x
|
||||
size = 2 * radius + 1
|
||||
return F.avg_pool2d(F.pad(x, (radius,) * 4, mode="replicate"), size, stride=1)
|
||||
|
||||
|
||||
def reach(height: int, width: int) -> int:
|
||||
return max(1, round(min(height, width) * REACH))
|
||||
|
||||
|
||||
def _moved(source: torch.Tensor, move: torch.Tensor) -> torch.Tensor:
|
||||
"""The original sampled at (zoom_x * x + shift_x, zoom_y * y + shift_y), x and y in cells from the centre.
|
||||
|
||||
move = [zoom_x - 1, zoom_y - 1, shift_x, shift_y].
|
||||
"""
|
||||
height, width = source.shape[-2:]
|
||||
theta = torch.zeros(source.shape[0], 2, 3, device=source.device, dtype=source.dtype)
|
||||
theta[:, 0, 0] = 1.0 + move[0]
|
||||
theta[:, 1, 1] = 1.0 + move[1]
|
||||
theta[:, 0, 2] = 2.0 * move[2] / width
|
||||
theta[:, 1, 2] = 2.0 * move[3] / height
|
||||
grid = F.affine_grid(theta, list(source.shape), align_corners=False)
|
||||
return F.grid_sample(source, grid, mode="bilinear", padding_mode="border", align_corners=False)
|
||||
|
||||
|
||||
def measure_redraw(pred: torch.Tensor, source: torch.Tensor) -> torch.Tensor:
|
||||
"""How much larger and how far shifted the prediction is drawn, as the move that lays the original over it.
|
||||
|
||||
pred and source are [1, C, H, W]. Fitted on the calmer half of the
|
||||
picture, so the edit itself does not pull the answer, from a coarse grid
|
||||
down to the full one. Returns [zoom_x - 1, zoom_y - 1, shift_x, shift_y];
|
||||
zeros when nothing can be measured.
|
||||
"""
|
||||
pred, source = pred.float(), source.float()
|
||||
move = torch.zeros(4, device=pred.device, dtype=torch.float32)
|
||||
limit = torch.tensor([MAX_ZOOM, MAX_ZOOM, MAX_SHIFT, MAX_SHIFT], device=pred.device)
|
||||
for level in (4, 2, 1):
|
||||
if min(pred.shape[-2:]) // level < 16:
|
||||
continue
|
||||
a = box_blur(F.avg_pool2d(pred, level) if level > 1 else pred, 1)
|
||||
b = box_blur(F.avg_pool2d(source, level) if level > 1 else source, 1)
|
||||
height, width = a.shape[-2:]
|
||||
ys = (torch.arange(height, device=a.device, dtype=torch.float32) + 0.5 - height / 2)[:, None].expand(height, width)
|
||||
xs = (torch.arange(width, device=a.device, dtype=torch.float32) + 0.5 - width / 2)[None, :].expand(height, width)
|
||||
here = move.clone()
|
||||
here[2:] = here[2:] / level
|
||||
for _ in range(6):
|
||||
laid = _moved(b, here)
|
||||
rest = a - laid
|
||||
rest = rest - rest.flatten(2).median(dim=2).values[:, :, None, None]
|
||||
wide = F.pad(laid, (1, 1, 1, 1), mode="replicate")
|
||||
gx = (wide[..., 1:-1, 2:] - wide[..., 1:-1, :-2]) / 2.0
|
||||
gy = (wide[..., 2:, 1:-1] - wide[..., :-2, 1:-1]) / 2.0
|
||||
size = rest.abs().mean(dim=1, keepdim=True)
|
||||
calm = (size <= size.flatten().median()).float()
|
||||
calm[..., :2, :] = 0.0
|
||||
calm[..., -2:, :] = 0.0
|
||||
calm[..., :, :2] = 0.0
|
||||
calm[..., :, -2:] = 0.0
|
||||
jac = (torch.stack([gx * xs, gy * ys, gx, gy], dim=-1) * calm[..., None]).reshape(-1, 4)
|
||||
normal = jac.T @ jac + 1e-6 * torch.eye(4, device=a.device)
|
||||
try:
|
||||
change = torch.linalg.solve(normal, jac.T @ (rest * calm).reshape(-1))
|
||||
except RuntimeError:
|
||||
return torch.zeros(4, device=pred.device, dtype=torch.float32)
|
||||
if not bool(torch.isfinite(change).all()):
|
||||
return torch.zeros(4, device=pred.device, dtype=torch.float32)
|
||||
here = here + change
|
||||
if float(change.abs().max()) < 1e-4:
|
||||
break
|
||||
move = here.clone()
|
||||
move[2:] = move[2:] * level
|
||||
move = torch.maximum(torch.minimum(move, limit), -limit)
|
||||
return move
|
||||
|
||||
|
||||
def redraws(pred: torch.Tensor, source: torch.Tensor) -> list[torch.Tensor]:
|
||||
"""measure_redraw for each picture of a batch."""
|
||||
return [measure_redraw(pred[index:index + 1], source[:1]) for index in range(pred.shape[0])]
|
||||
|
||||
|
||||
def difference(pred: torch.Tensor, source: torch.Tensor, moves: list[torch.Tensor] | None = None) -> torch.Tensor:
|
||||
"""How far each spot of the prediction is from the original, [B, 1, H, W].
|
||||
|
||||
Two things are not a change. A redraw of the whole picture slightly
|
||||
larger or shifted: the original is laid over the prediction first. And
|
||||
an even tint: each channel's middle difference is taken out. The result
|
||||
is in units of the original's own spread, which keeps the numbers alike
|
||||
between models.
|
||||
"""
|
||||
pred = pred.float()
|
||||
source = source.float()
|
||||
moves = moves if moves is not None else redraws(pred, source)
|
||||
out = []
|
||||
for index in range(pred.shape[0]):
|
||||
delta = pred[index:index + 1] - _moved(source[:1], moves[index])
|
||||
delta = delta - delta.flatten(2).median(dim=2).values[:, :, None, None]
|
||||
out.append(delta.abs().mean(dim=1, keepdim=True))
|
||||
magnitude = torch.cat(out) / source.std().clamp_min(1e-6)
|
||||
return box_blur(magnitude, reach(*magnitude.shape[-2:]))
|
||||
|
||||
|
||||
def moved_back(pred: torch.Tensor, moves: list[torch.Tensor]) -> torch.Tensor:
|
||||
"""The prediction with its zoom and shift undone, so it sits where the original does.
|
||||
|
||||
Resampling softens a latent, so a move under a tenth of a cell at the
|
||||
corner is left alone, and the rest is resampled bicubic: on recorded
|
||||
runs that kept 85% of the fine detail where bilinear kept 65%.
|
||||
"""
|
||||
out = []
|
||||
for index in range(pred.shape[0]):
|
||||
one = pred[index:index + 1]
|
||||
move = moves[index]
|
||||
height, width = one.shape[-2:]
|
||||
corner = (move[0].abs() * width / 2 + move[2].abs()).maximum(move[1].abs() * height / 2 + move[3].abs())
|
||||
if float(corner) < 0.1:
|
||||
out.append(one)
|
||||
continue
|
||||
theta = torch.zeros(1, 2, 3, device=one.device, dtype=torch.float32)
|
||||
theta[:, 0, 0] = 1.0 / (1.0 + move[0])
|
||||
theta[:, 1, 1] = 1.0 / (1.0 + move[1])
|
||||
theta[:, 0, 2] = -2.0 * move[2] / (1.0 + move[0]) / width
|
||||
theta[:, 1, 2] = -2.0 * move[3] / (1.0 + move[1]) / height
|
||||
grid = F.affine_grid(theta, list(one.shape), align_corners=False)
|
||||
back = F.grid_sample(one.float(), grid, mode="bicubic", padding_mode="border", align_corners=False)
|
||||
out.append(back.to(one.dtype))
|
||||
return torch.cat(out)
|
||||
|
||||
|
||||
def broad_change(pred: torch.Tensor, source: torch.Tensor) -> torch.Tensor:
|
||||
"""How far the calmest fifth of each picture sits from the original, [B].
|
||||
|
||||
Read on a smoothed copy, where a redraw one percent larger nearly
|
||||
vanishes and a new style, new light or new weather does not. An even
|
||||
tint is taken out first: the model adds one by itself, a stronger one
|
||||
on large pictures, and that is no edit.
|
||||
"""
|
||||
delta = (pred - source).float()
|
||||
delta = delta - delta.flatten(2).median(dim=2).values[:, :, None, None]
|
||||
delta = box_blur(delta, max(1, round(min(delta.shape[-2:]) / 30.0)))
|
||||
level = (delta * delta).mean(dim=1).sqrt().flatten(1)
|
||||
stride = max(1, level.shape[1] // 200000)
|
||||
spread = source.float().std().clamp_min(1e-6)
|
||||
return torch.quantile(level[:, ::stride], 0.2, dim=1) / spread
|
||||
|
||||
|
||||
def quiet_level(magnitude: torch.Tensor) -> torch.Tensor:
|
||||
"""The tenth of each picture that differs least, [B]."""
|
||||
flat = magnitude.flatten(1)
|
||||
stride = max(1, flat.shape[1] // 200000)
|
||||
return torch.quantile(flat[:, ::stride], 0.1, dim=1)
|
||||
|
||||
|
||||
def peaks(magnitude: torch.Tensor, quiet: torch.Tensor) -> tuple[float, float]:
|
||||
"""How clear the clearest change is, in units of the quiet level: (strongest spot, strongest hundredth)."""
|
||||
ratio = (magnitude / quiet.clamp_min(1e-6)[:, None, None, None]).flatten(1)
|
||||
stride = max(1, ratio.shape[1] // 200000)
|
||||
return float(ratio.max(dim=1).values.min()), float(torch.quantile(ratio[:, ::stride], 0.99, dim=1).min())
|
||||
|
||||
|
||||
def change_core(magnitude: torch.Tensor, quiet: torch.Tensor, tolerance: float = 1.0,
|
||||
held: torch.Tensor | None = None) -> torch.Tensor:
|
||||
"""1 where the edit changed the picture, 0 where it did not, hard-edged, [B, 1, H, W].
|
||||
|
||||
`held` is what earlier steps called changed: it stays changed, and
|
||||
faint change that touches it joins it.
|
||||
"""
|
||||
ratio = magnitude / quiet.clamp_min(1e-6)[:, None, None, None]
|
||||
weak = (ratio > WEAK * tolerance).float()
|
||||
core = (ratio > STRONG * tolerance).float()
|
||||
if held is not None:
|
||||
weak = torch.maximum(weak, held)
|
||||
core = torch.maximum(core, held)
|
||||
for _ in range(sum(core.shape[-2:])):
|
||||
grown = F.max_pool2d(core, 3, stride=1, padding=1) * weak
|
||||
if torch.equal(grown, core):
|
||||
break
|
||||
core = grown
|
||||
return core
|
||||
|
||||
|
||||
def soften(core: torch.Tensor) -> torch.Tensor:
|
||||
"""The hard change map with a margin and a soft edge, so the edit has room to blend in."""
|
||||
grow = reach(*core.shape[-2:])
|
||||
return box_blur(F.max_pool2d(core, 2 * grow + 1, stride=1, padding=grow), grow)
|
||||
|
||||
|
||||
def fit_latent(latent: torch.Tensor, height: int, width: int) -> torch.Tensor | None:
|
||||
"""The original's latent on the result's grid, or None when they are different pictures."""
|
||||
h, w = latent.shape[-2:]
|
||||
if (h, w) == (height, width):
|
||||
return latent
|
||||
if abs(h / height - 1.0) > SIZE_SLACK or abs(w / width - 1.0) > SIZE_SLACK:
|
||||
return None
|
||||
return F.interpolate(latent, size=(height, width), mode="bilinear", align_corners=False)
|
||||
|
||||
|
||||
class EditLock:
|
||||
"""One lock: the original, the settings, and what the last run kept."""
|
||||
|
||||
def __init__(self, mode: str = "auto", tolerance: float = 1.0):
|
||||
self.mode = mode
|
||||
self.tolerance = tolerance
|
||||
self.source: torch.Tensor | None = None # IMAGE, [1, H, W, 3]
|
||||
self.latent: torch.Tensor | None = None # [1, C, H, W] or [1, C, 1, H, W], model space
|
||||
self.user_mask: torch.Tensor | None = None # MASK, white = the edit may happen
|
||||
self.status = "waiting"
|
||||
self.mask: torch.Tensor | None = None # [B, 1, h, w], 1 = changed, soft-edged
|
||||
self.core: torch.Tensor | None = None # the same before its margin, hard-edged
|
||||
self.quiet: torch.Tensor | None = None # [B], the quiet level at the decision
|
||||
self.broad: float | None = None # the first look's whole-picture reading
|
||||
self.wide: float | None = None # share of the picture the lock would free, at the decision
|
||||
self.top: float | None = None # the strongest spot then, in units of the quiet level
|
||||
self.peak: float | None = None # and the strongest hundredth
|
||||
self._sigma: float | None = None
|
||||
self._first = True
|
||||
self._first_sigma = 1.0
|
||||
self._decided = False
|
||||
|
||||
def _begin(self) -> None:
|
||||
self.status = "waiting"
|
||||
self.mask = None
|
||||
self.core = None
|
||||
self.quiet = None
|
||||
self.broad = None
|
||||
self.wide = None
|
||||
self.top = None
|
||||
self.peak = None
|
||||
self._first = True
|
||||
self._decided = False
|
||||
|
||||
def _decides_now(self, args: dict, sigma: float) -> bool:
|
||||
"""True once the noise is down far enough, or this is the run's last step."""
|
||||
steps = (args.get("model_options") or {}).get("transformer_options", {}).get("sample_sigmas")
|
||||
if steps is None or len(steps) < 2:
|
||||
return True # the sampler does not say where the run ends: decide at once
|
||||
if sigma <= float(steps[-2]) * 1.0001:
|
||||
return True
|
||||
return sigma <= DECIDE_AT * self._first_sigma
|
||||
|
||||
def step(self, args: dict) -> torch.Tensor:
|
||||
"""ComfyUI's post-CFG hook: the prediction comes in, the locked prediction goes out."""
|
||||
pred = args["denoised"]
|
||||
sigma = float(args["sigma"].flatten()[0])
|
||||
if self._sigma is None or sigma >= self._sigma:
|
||||
self._begin() # the noise level went up: a new run
|
||||
self._sigma = sigma
|
||||
if self.status in ("whole picture", "size", "video", "vae") or self.latent is None:
|
||||
return pred
|
||||
video = pred.ndim == 5
|
||||
if video and pred.shape[2] != 1:
|
||||
self.status = "video"
|
||||
return pred
|
||||
flat = pred[:, :, 0] if video else pred
|
||||
latent = self.latent[:, :, 0] if self.latent.ndim == 5 else self.latent
|
||||
if latent.shape[1] != flat.shape[1]:
|
||||
self.status = "vae"
|
||||
return pred
|
||||
source = fit_latent(latent.to(device=flat.device, dtype=torch.float32), *flat.shape[-2:])
|
||||
if source is None:
|
||||
self.status = "size"
|
||||
return pred
|
||||
if self.status == "soft":
|
||||
# nothing is put back, the picture is only held where the original sits
|
||||
back = moved_back(flat, redraws(flat, source))
|
||||
return back[:, :, None] if video else back
|
||||
if self.user_mask is not None:
|
||||
mask = self._given_mask(flat)
|
||||
else:
|
||||
if self._first:
|
||||
self._first_sigma = sigma
|
||||
self.broad = float(broad_change(flat, source).max())
|
||||
if self.mode == "auto" and self.broad > WHOLE_PICTURE:
|
||||
self.status = "whole picture"
|
||||
return pred
|
||||
moves = redraws(flat, source)
|
||||
magnitude = difference(flat, source, moves)
|
||||
if not self._decided and not self._decides_now(args, sigma):
|
||||
# still watching: the model draws its own picture, only its zoom and shift are undone
|
||||
self._first = False
|
||||
self.status = "watching"
|
||||
if self.broad > WHOLE_PICTURE:
|
||||
return pred # too much is changing to trust the measured move
|
||||
back = moved_back(flat, moves)
|
||||
return back[:, :, None] if video else back
|
||||
if not self._decided:
|
||||
self._decided = True
|
||||
# the yardstick is set once, here: later steps are read against it
|
||||
self.quiet = quiet_level(magnitude)
|
||||
self.top, self.peak = peaks(magnitude, self.quiet)
|
||||
self.wide = float(change_core(magnitude, self.quiet, self.tolerance).flatten(1).mean(dim=1).max())
|
||||
thin = self.top < SOFT_PEAK or (self.wide >= SOFT_WIDE and self.peak < SOFT_PEAK)
|
||||
if self.mode == "auto" and not self._first and thin:
|
||||
# a soft change over the whole picture: a lock would hold it back or let it through in patches
|
||||
self.status = "soft"
|
||||
back = moved_back(flat, moves)
|
||||
return back[:, :, None] if video else back
|
||||
held = self.core if self.core is not None and self.core.shape[0] == flat.shape[0] else None
|
||||
self.core = change_core(magnitude, self.quiet.to(magnitude.device), self.tolerance,
|
||||
held.to(magnitude.device) if held is not None else None).detach()
|
||||
mask = soften(self.core)
|
||||
self._first = False
|
||||
self.mask = mask.detach()
|
||||
self.status = "locked"
|
||||
keep = (1.0 - mask).to(flat.dtype)
|
||||
locked = flat * (1.0 - keep) + source.to(flat.dtype) * keep
|
||||
return locked[:, :, None] if video else locked
|
||||
|
||||
def _given_mask(self, flat: torch.Tensor) -> torch.Tensor:
|
||||
mask = self.user_mask
|
||||
if mask.ndim == 2:
|
||||
mask = mask[None]
|
||||
mask = mask[:1, None].to(device=flat.device, dtype=torch.float32)
|
||||
mask = F.interpolate(mask, size=flat.shape[-2:], mode="bilinear", align_corners=False).clamp(0.0, 1.0)
|
||||
return soften(mask).expand(flat.shape[0], -1, -1, -1)
|
||||
|
||||
def kept_share(self) -> float | None:
|
||||
if self.status != "locked" or self.mask is None:
|
||||
return None
|
||||
return float((self.mask < 0.5).float().mean())
|
||||
|
||||
def pixel_mask(self, height: int, width: int) -> torch.Tensor | None:
|
||||
"""The change mask at the picture's size, [B, H, W], or None when nothing was locked."""
|
||||
if self.status != "locked" or self.mask is None:
|
||||
return None
|
||||
mask = F.interpolate(self.mask.float().cpu(), size=(height, width), mode="bilinear", align_corners=False)
|
||||
return mask[:, 0].clamp(0.0, 1.0)
|
||||
|
||||
|
||||
def restore_pixels(result: torch.Tensor, source: torch.Tensor, mask: torch.Tensor,
|
||||
feather: int) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""The original's own pixels where the mask is black. result [B, H, W, C], source [1, h, w, 3], mask [b, H, W].
|
||||
|
||||
The fade lies outside the edit: the mask is widened by the feather first.
|
||||
"""
|
||||
frames, height, width = int(result.shape[0]), int(result.shape[1]), int(result.shape[2])
|
||||
edit = mask[:, None].float()
|
||||
if edit.shape[0] not in (1, frames):
|
||||
edit = edit[:1]
|
||||
if feather > 0:
|
||||
size = 2 * feather + 1
|
||||
edit = F.max_pool2d(edit, size, stride=1, padding=feather)
|
||||
edit = box_blur(edit, feather)
|
||||
original = source[..., :3].float()
|
||||
if tuple(original.shape[1:3]) != (height, width):
|
||||
original = F.interpolate(original.movedim(-1, 1), size=(height, width), mode="bicubic",
|
||||
align_corners=False, antialias=True).movedim(1, -1).clamp(0.0, 1.0)
|
||||
weight = edit.movedim(1, -1).to(result.device, result.dtype)
|
||||
out = result.clone()
|
||||
out[..., :3] = result[..., :3] * weight + original.to(result.device, result.dtype) * (1.0 - weight)
|
||||
edit_mask = edit[:, 0].expand(frames, -1, -1).clone()
|
||||
return out, edit_mask
|
||||
|
||||
|
||||
def describe(lock: EditLock) -> str:
|
||||
if lock.status == "locked":
|
||||
kept = lock.kept_share() or 0.0
|
||||
how = "your mask" if lock.user_mask is not None else "what the edit changed"
|
||||
text = (f"Edit Lock kept {kept * 100:.0f}% of your picture as it was and left "
|
||||
f"{(1.0 - kept) * 100:.0f}% to the edit ({how}).")
|
||||
if kept >= NEARLY_ALL and lock.user_mask is None:
|
||||
# rain, snow or smoke over the whole picture reads as a small edit and is held back
|
||||
text += (" If your edit is missing, it is one that spreads thinly over the whole picture "
|
||||
"(rain, snow, smoke): set lock to off.")
|
||||
return text
|
||||
if lock.status == "whole picture":
|
||||
return ("Edit Lock stepped aside: the edit changes the whole picture, so there was nothing to keep. "
|
||||
"Set lock to always to lock anyway.")
|
||||
if lock.status == "soft":
|
||||
return ("Edit Lock stepped aside: the edit spreads thinly over the whole picture (rain, snow, haze), "
|
||||
"so no part of it stayed as it was. It only held the picture in place. "
|
||||
"Set lock to always to lock anyway.")
|
||||
if lock.status == "off":
|
||||
return "Edit Lock is off."
|
||||
if lock.status == "size":
|
||||
return ("Edit Lock did nothing: the picture and the result are different sizes. "
|
||||
"Give it the picture at the size of the result.")
|
||||
if lock.status == "video":
|
||||
return "Edit Lock did nothing: it works on single pictures, not video."
|
||||
if lock.status == "vae":
|
||||
return "Edit Lock did nothing: this VAE does not belong to this model."
|
||||
return "Edit Lock has not run yet: plug its model output into the sampler."
|
||||
@@ -0,0 +1,153 @@
|
||||
"""Edit Lock Restore's tint fix: take out the even tint an edit model adds to the whole picture.
|
||||
|
||||
MiniMax H3 Image gives every edit one even color shift: a little lighter,
|
||||
pinker and bluer (Lab L +1 to +3, a +2 to +7, b -1 to -6 at 16 MP). Edit
|
||||
Lock removes it wherever it locks, because those parts are your picture
|
||||
again. When the lock steps aside (new hair color, new background, rain) the
|
||||
shift stays. Here it is read from the part of the picture the edit left
|
||||
alone and subtracted from every pixel.
|
||||
|
||||
A shift is only taken out when it looks like that tint. An edit that is
|
||||
meant to recolor everything (golden hour, red light, a restyle) moves the
|
||||
colors much further or the other way, or leaves too little of the picture
|
||||
alone to read anything, and passes through untouched.
|
||||
|
||||
Pure torch, no ComfyUI imports.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import math
|
||||
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
|
||||
from ._realign_helpers import measure, warp_to_source
|
||||
|
||||
# The pictures are compared at this long side; the fit is sub-pixel there.
|
||||
WORK = 1536
|
||||
# Cells closer than this to the running estimate count as "left alone".
|
||||
NEAR = 4.0
|
||||
# The tint as measured on 57 edits: lighter, pinker, bluer, within these
|
||||
# bounds (Lab units). Smaller than MIN_A is not worth touching.
|
||||
MIN_A = 0.5
|
||||
MAX_A = 10.0
|
||||
MAX_B = 0.0
|
||||
MIN_B = -8.0
|
||||
MIN_L = -1.0
|
||||
MAX_L = 6.0
|
||||
# At least this share of the picture must be left alone to read the tint.
|
||||
MIN_SHARE = 0.15
|
||||
|
||||
_D65 = (0.95047, 1.0, 1.08883)
|
||||
_TO_XYZ = ((0.4124564, 0.3575761, 0.1804375),
|
||||
(0.2126729, 0.7151522, 0.0721750),
|
||||
(0.0193339, 0.1191920, 0.9503041))
|
||||
_TO_RGB = ((3.2404542, -1.5371385, -0.4985314),
|
||||
(-0.9692660, 1.8760108, 0.0415560),
|
||||
(0.0556434, -0.2040259, 1.0572252))
|
||||
|
||||
|
||||
def rgb_to_lab(rgb: torch.Tensor) -> torch.Tensor:
|
||||
"""sRGB in 0..1, [..., 3] -> CIE Lab, float64."""
|
||||
x = rgb.double().clamp(0.0, 1.0)
|
||||
linear = torch.where(x > 0.04045, ((x + 0.055) / 1.055) ** 2.4, x / 12.92)
|
||||
xyz = linear @ torch.tensor(_TO_XYZ, dtype=torch.float64).T / torch.tensor(_D65, dtype=torch.float64)
|
||||
f = torch.where(xyz > 0.008856, xyz.clamp(min=1e-9) ** (1.0 / 3.0), 7.787 * xyz + 16.0 / 116.0)
|
||||
return torch.stack([116.0 * f[..., 1] - 16.0, 500.0 * (f[..., 0] - f[..., 1]),
|
||||
200.0 * (f[..., 1] - f[..., 2])], -1)
|
||||
|
||||
|
||||
def lab_to_rgb(lab: torch.Tensor) -> torch.Tensor:
|
||||
"""CIE Lab, [..., 3] -> sRGB in 0..1, float64."""
|
||||
fy = (lab[..., 0] + 16.0) / 116.0
|
||||
f = torch.stack([fy + lab[..., 1] / 500.0, fy, fy - lab[..., 2] / 200.0], -1)
|
||||
xyz = torch.where(f > 0.206893, f ** 3, (f - 16.0 / 116.0) / 7.787) * torch.tensor(_D65, dtype=torch.float64)
|
||||
linear = (xyz @ torch.tensor(_TO_RGB, dtype=torch.float64).T).clamp(0.0, 1.0)
|
||||
return torch.where(linear > 0.0031308, 1.055 * linear ** (1.0 / 2.4) - 0.055, 12.92 * linear).clamp(0.0, 1.0)
|
||||
|
||||
|
||||
def _fit(image: torch.Tensor, long_side: int) -> torch.Tensor:
|
||||
"""HWC float at most long_side px on its long side."""
|
||||
height, width = int(image.shape[0]), int(image.shape[1])
|
||||
scale = long_side / max(height, width)
|
||||
if scale >= 1.0:
|
||||
return image.float()
|
||||
size = (max(1, round(height * scale)), max(1, round(width * scale)))
|
||||
return F.interpolate(image.movedim(-1, 0)[None].float(), size=size, mode="bilinear",
|
||||
align_corners=False, antialias=True)[0].movedim(0, -1)
|
||||
|
||||
|
||||
def measure_tint(result: torch.Tensor, source: torch.Tensor) -> tuple[list[float] | None, float]:
|
||||
"""The even shift in (L, a, b) from source to result, read where the edit left the picture alone.
|
||||
|
||||
result, source: HWC float in 0..1, any sizes. Returns (shift or None, the
|
||||
share of the picture it was read on). The result is lined up with the
|
||||
source first, since edits often come back slightly zoomed or moved.
|
||||
"""
|
||||
src = _fit(source[..., :3], WORK)
|
||||
edit = _fit(result[..., :3], WORK)
|
||||
fit = measure(src, edit, mode="zoom + shift")
|
||||
if not fit.reliable:
|
||||
return None, 0.0
|
||||
height, width = int(src.shape[0]), int(src.shape[1])
|
||||
scaled = F.interpolate(edit.movedim(-1, 0)[None], size=(height, width), mode="bilinear",
|
||||
align_corners=False)[0].movedim(0, -1)
|
||||
try:
|
||||
warped = warp_to_source(scaled, fit.matrix, width, height)
|
||||
warped = warped[0] if isinstance(warped, tuple) else warped
|
||||
except Exception:
|
||||
warped = scaled
|
||||
# coarse cells of about 8 px at 1 MP, so grain and small detail drop out
|
||||
cell = max(4, round(8 / math.sqrt(1e6 / (width * height))))
|
||||
|
||||
def cells(img: torch.Tensor) -> torch.Tensor:
|
||||
return F.avg_pool2d(img.movedim(-1, 0)[None], cell, stride=cell)[0].movedim(0, -1)
|
||||
|
||||
diff = rgb_to_lab(cells(warped)) - rgb_to_lab(cells(src))
|
||||
margin = max(1, int(diff.shape[0]) // 12) # the edges carry the warp's filled strip
|
||||
diff = diff[margin:-margin, margin:-margin].reshape(-1, 3)
|
||||
if diff.shape[0] < 50:
|
||||
return None, 0.0
|
||||
centre = diff.median(0).values
|
||||
share = 1.0
|
||||
# start at the median change, then keep only the cells near it: a new
|
||||
# background or new hair is a big change, but not the same one everywhere
|
||||
for _ in range(5):
|
||||
near = (diff - centre).norm(dim=1) < NEAR
|
||||
if int(near.sum()) < 50:
|
||||
break
|
||||
centre = diff[near].mean(0)
|
||||
share = float(near.double().mean())
|
||||
return [float(v) for v in centre], share
|
||||
|
||||
|
||||
def looks_like_tint(shift: list[float] | None, share: float) -> bool:
|
||||
"""True for the model's own lighter-pinker-bluer shift; False for an intended recolor."""
|
||||
if shift is None or share < MIN_SHARE:
|
||||
return False
|
||||
light, a, b = shift
|
||||
return MIN_A <= a <= MAX_A and MIN_B <= b <= MAX_B and MIN_L <= light <= MAX_L
|
||||
|
||||
|
||||
def remove_tint(image: torch.Tensor, shift: list[float]) -> torch.Tensor:
|
||||
"""Subtract one (L, a, b) shift from every pixel. image: [B, H, W, C] in 0..1."""
|
||||
lab = rgb_to_lab(image[..., :3]) - torch.tensor(shift, dtype=torch.float64)
|
||||
out = image.clone()
|
||||
out[..., :3] = lab_to_rgb(lab).to(image.dtype)
|
||||
return out
|
||||
|
||||
|
||||
def take_out_tint(images: torch.Tensor, source: torch.Tensor) -> tuple[torch.Tensor, list[float] | None]:
|
||||
"""Each picture of the batch with its tint taken out when it looks like the model's tint.
|
||||
|
||||
Returns the pictures and the largest shift taken out (None when none was)."""
|
||||
out = images.clone()
|
||||
taken: list[float] | None = None
|
||||
for index in range(int(images.shape[0])):
|
||||
shift, share = measure_tint(images[index], source[0])
|
||||
if looks_like_tint(shift, share):
|
||||
out[index:index + 1] = remove_tint(images[index:index + 1], shift)
|
||||
if taken is None or abs(shift[1]) > abs(taken[1]):
|
||||
taken = shift
|
||||
return out, taken
|
||||
@@ -0,0 +1,946 @@
|
||||
"""FlashVSR video upscaling: the model's parts and the loop that runs a clip.
|
||||
|
||||
FlashVSR (OpenImagingLab, Apache-2.0) sharpens a clip in one model step.
|
||||
It is a Wan 2.1 1.3B transformer that is also shown the low-quality frames,
|
||||
a small projector that turns those frames into tokens, and a small decoder
|
||||
that reads the frames again while it draws the result.
|
||||
|
||||
Everything here is plain PyTorch. The model looks at a clip a few frames at
|
||||
a time and keeps a short memory of the frames before, so a long clip costs
|
||||
no more graphics memory than a short one. Module and parameter names follow
|
||||
the names in the published weights, which is what lets them load.
|
||||
|
||||
Shapes used below: a token is 16x16 output pixels of one latent frame, a
|
||||
latent frame is 4 video frames, and an attention block is 2 latent frames
|
||||
by 8x8 tokens (128 tokens, 128x128 pixels over 8 frames).
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import math
|
||||
import os
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
PIXEL_STEP = 128 # the working size is a multiple of this in both directions
|
||||
BLOCK_TOKENS = 128
|
||||
KEPT_GROUPS = 3 # earlier 2-frame groups each layer remembers
|
||||
LOCAL_REACH = 11 # a block looks at blocks inside an 11x11 neighbourhood
|
||||
SPARSE_RATIO = 2.0
|
||||
SPARSE_REFERENCE_AREA = 768 * 1280
|
||||
FIRST_PASS_LATENTS = 6
|
||||
PASS_LATENTS = 2
|
||||
MIN_GROUPS = 3 # the first pass alone needs 25 frames
|
||||
TIMESTEP = 1000.0
|
||||
DEFAULT_SLOT_BUDGET = 12000
|
||||
|
||||
|
||||
def _cast(parameter: torch.Tensor, like: torch.Tensor) -> torch.Tensor:
|
||||
return parameter.to(device=like.device, dtype=like.dtype)
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# Sizes and frame counts
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
def _even(value: float) -> int:
|
||||
return max(2, int(round(value / 2.0)) * 2)
|
||||
|
||||
|
||||
def plan_sizes(height: int, width: int, scale: float) -> tuple[int, int, int, int]:
|
||||
"""Output size (even numbers) and the padded size the model works at."""
|
||||
out_h, out_w = _even(height * scale), _even(width * scale)
|
||||
work_h = math.ceil(out_h / PIXEL_STEP) * PIXEL_STEP
|
||||
work_w = math.ceil(out_w / PIXEL_STEP) * PIXEL_STEP
|
||||
return out_h, out_w, work_h, work_w
|
||||
|
||||
|
||||
def plan_groups(frame_count: int) -> int:
|
||||
"""How many 8-frame groups the padded clip has.
|
||||
|
||||
The model reads 8n+1 frames and returns 8n-3, so the clip is padded by
|
||||
repeating its last frame until 8n-3 covers every frame it has.
|
||||
"""
|
||||
return max(MIN_GROUPS, math.ceil((int(frame_count) + 3) / 8))
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# Position encoding and attention blocks
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
def _axis_angles(dim: int, positions: torch.Tensor) -> torch.Tensor:
|
||||
steps = torch.arange(0, dim, 2, dtype=torch.float64)[: dim // 2] / dim
|
||||
return torch.outer(positions.double(), 1.0 / (10000.0**steps))
|
||||
|
||||
|
||||
def rope_table(head_dim: int, first_frame: int, frames: int, rows: int, cols: int, device):
|
||||
"""Rotation angles for every token, as (cos, sin) of shape [tokens, 1, head_dim / 2]."""
|
||||
side = head_dim // 3
|
||||
t = _axis_angles(head_dim - 2 * side, torch.arange(first_frame, first_frame + frames))
|
||||
r = _axis_angles(side, torch.arange(rows))
|
||||
c = _axis_angles(side, torch.arange(cols))
|
||||
angles = torch.cat(
|
||||
[
|
||||
t[:, None, None, :].expand(frames, rows, cols, -1),
|
||||
r[None, :, None, :].expand(frames, rows, cols, -1),
|
||||
c[None, None, :, :].expand(frames, rows, cols, -1),
|
||||
],
|
||||
dim=-1,
|
||||
).reshape(frames * rows * cols, 1, head_dim // 2)
|
||||
return angles.cos().float().to(device), angles.sin().float().to(device)
|
||||
|
||||
|
||||
def apply_rope(x: torch.Tensor, cos: torch.Tensor, sin: torch.Tensor, heads: int) -> torch.Tensor:
|
||||
"""Rotate each pair of channels of x [tokens, dim] by its token's angle."""
|
||||
tokens, dim = x.shape
|
||||
a, b = x.float().view(tokens, heads, dim // heads // 2, 2).unbind(-1)
|
||||
return torch.stack((a * cos - b * sin, a * sin + b * cos), dim=-1).view(tokens, dim).to(x.dtype)
|
||||
|
||||
|
||||
def to_blocks(x: torch.Tensor, frames: int, rows: int, cols: int) -> torch.Tensor:
|
||||
"""[tokens, dim] in (frame, row, column) order -> [blocks, 128, dim]."""
|
||||
dim = x.shape[-1]
|
||||
x = x.view(frames // 2, 2, rows // 8, 8, cols // 8, 8, dim).permute(0, 2, 4, 1, 3, 5, 6)
|
||||
return x.reshape(-1, BLOCK_TOKENS, dim)
|
||||
|
||||
|
||||
def from_blocks(blocks: torch.Tensor, frames: int, rows: int, cols: int) -> torch.Tensor:
|
||||
dim = blocks.shape[-1]
|
||||
x = blocks.view(frames // 2, rows // 8, cols // 8, 2, 8, 8, dim).permute(0, 3, 1, 4, 2, 5, 6)
|
||||
return x.reshape(frames * rows * cols, dim)
|
||||
|
||||
|
||||
def local_block_mask(block_rows: int, block_cols: int, reach: int, device) -> torch.Tensor:
|
||||
"""[blocks, blocks] True where a key block is inside a query block's neighbourhood."""
|
||||
r = torch.arange(block_rows, device=device).repeat_interleave(block_cols)
|
||||
c = torch.arange(block_cols, device=device).repeat(block_rows)
|
||||
dr, dc = r[None, :] - r[:, None], c[None, :] - c[:, None]
|
||||
low, high = -(reach // 2), reach - 1 - reach // 2
|
||||
return (dr >= low) & (dr <= high) & (dc >= low) & (dc <= high)
|
||||
|
||||
|
||||
def choose_key_blocks(
|
||||
q_blocks: torch.Tensor,
|
||||
k_blocks: torch.Tensor,
|
||||
heads: int,
|
||||
local: torch.Tensor,
|
||||
query_groups: int,
|
||||
keep: int,
|
||||
) -> torch.Tensor:
|
||||
"""Pick, for every head, which key blocks each query block reads.
|
||||
|
||||
A block's mean query is scored against every block's mean key. Blocks
|
||||
outside the neighbourhood are ruled out, the scores of a query block are
|
||||
turned into shares that sum to one, and for each head and each 2-frame
|
||||
group of queries the `keep` largest shares win. Returns two
|
||||
[heads, query blocks, key blocks] tensors: True where a pair is kept,
|
||||
and the shares themselves (0 outside the neighbourhood).
|
||||
"""
|
||||
nq, _, dim = q_blocks.shape
|
||||
nk = k_blocks.shape[0]
|
||||
head_dim = dim // heads
|
||||
q_mean = q_blocks.float().mean(1).view(nq, heads, head_dim).transpose(0, 1)
|
||||
k_mean = k_blocks.float().mean(1).view(nk, heads, head_dim).transpose(0, 1)
|
||||
scores = q_mean @ k_mean.transpose(1, 2) / math.sqrt(head_dim)
|
||||
per_group = local.shape[0]
|
||||
allowed = local.repeat(nq // per_group, nk // per_group)
|
||||
shares = scores.masked_fill(~allowed, float("-inf")).softmax(dim=-1)
|
||||
flat = shares.view(heads, query_groups, per_group * nk)
|
||||
keep = min(flat.shape[-1] - 1, int(keep))
|
||||
threshold = flat.topk(keep + 1, dim=-1).values[..., -1:]
|
||||
return (flat > threshold).view(heads, nq, nk), shares
|
||||
|
||||
|
||||
def sparse_block_attention(
|
||||
q: torch.Tensor,
|
||||
k: torch.Tensor,
|
||||
v: torch.Tensor,
|
||||
chosen: torch.Tensor,
|
||||
shares: torch.Tensor | None = None,
|
||||
slot_budget: int = DEFAULT_SLOT_BUDGET,
|
||||
) -> torch.Tensor:
|
||||
"""Attention where each query block only reads its chosen key blocks.
|
||||
|
||||
q is [heads, query blocks, 128, head_dim], k and v are
|
||||
[heads, key blocks, 128, head_dim] and chosen is
|
||||
[heads, query blocks, key blocks]. Each (head, query block) is a row
|
||||
that gathers the keys it chose. Rows with a similar number of chosen
|
||||
blocks run together, all at the width of the row that chose most. A
|
||||
row that chose nothing returns zeros. `slot_budget` caps
|
||||
(rows x chosen blocks) per call, which caps memory.
|
||||
|
||||
With `shares` (from choose_key_blocks) a narrower row is filled up
|
||||
with its next-best blocks from its neighbourhood, so a call needs no
|
||||
padding mask, which is what makes it fast: a row then reads a few
|
||||
blocks more than it chose, never fewer. Without `shares` the padding
|
||||
is masked and every row reads exactly what it chose.
|
||||
"""
|
||||
heads, nq, tokens, head_dim = q.shape
|
||||
nk = k.shape[1]
|
||||
rows = chosen.reshape(heads * nq, nk)
|
||||
counts = rows.sum(dim=1)
|
||||
if shares is None:
|
||||
rank, usable = rows.float(), counts
|
||||
else:
|
||||
flat_shares = shares.reshape(heads * nq, nk).float()
|
||||
rank, usable = rows.float() * 2 + flat_shares, (flat_shares > 0).sum(dim=1)
|
||||
order = torch.argsort(counts)
|
||||
picks = torch.argsort(rank, dim=1, descending=True, stable=True)
|
||||
picks = picks + (torch.arange(heads, device=q.device).repeat_interleave(nq) * nk)[:, None]
|
||||
q_rows = q.reshape(heads * nq, tokens, head_dim)
|
||||
k_rows = k.reshape(heads * nk, tokens, head_dim)
|
||||
v_rows = v.reshape(heads * nk, tokens, head_dim)
|
||||
out = torch.zeros_like(q_rows)
|
||||
sorted_counts = counts[order].tolist()
|
||||
total = len(sorted_counts)
|
||||
start = 0
|
||||
while start < total and sorted_counts[start] == 0:
|
||||
start += 1
|
||||
while start < total:
|
||||
end = start + 1
|
||||
while end < total and (end - start + 1) * sorted_counts[end] <= slot_budget:
|
||||
end += 1
|
||||
rows_here = order[start:end]
|
||||
widest = sorted_counts[end - 1]
|
||||
index = picks[rows_here, :widest]
|
||||
keys = k_rows[index].view(end - start, widest * tokens, head_dim)
|
||||
values = v_rows[index].view(end - start, widest * tokens, head_dim)
|
||||
mask = None
|
||||
real = torch.arange(widest, device=q.device)[None, :] < usable[rows_here][:, None]
|
||||
if not bool(real.all()):
|
||||
mask = real.repeat_interleave(tokens, dim=1)[:, None, :]
|
||||
out[rows_here] = F.scaled_dot_product_attention(q_rows[rows_here], keys, values, attn_mask=mask)
|
||||
start = end
|
||||
return out.view(heads, nq, tokens, head_dim)
|
||||
|
||||
|
||||
class AttentionPlan:
|
||||
"""What stays the same for every layer and pass at one working size."""
|
||||
|
||||
def __init__(self, rows: int, cols: int, device, slot_budget: int = DEFAULT_SLOT_BUDGET):
|
||||
self.rows, self.cols = rows, cols
|
||||
self.per_group = (rows // 8) * (cols // 8)
|
||||
self.local = local_block_mask(rows // 8, cols // 8, LOCAL_REACH, device)
|
||||
ratio = SPARSE_RATIO * SPARSE_REFERENCE_AREA / (rows * 16 * cols * 16)
|
||||
self.keep = int(self.per_group * self.per_group * ratio) - 1
|
||||
self.slot_budget = slot_budget
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# The transformer
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
class _RMSNorm(nn.Module):
|
||||
def __init__(self, dim: int, eps: float, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.eps = eps
|
||||
self.weight = nn.Parameter(torch.ones(dim, device=device, dtype=dtype))
|
||||
|
||||
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
||||
y = x.float()
|
||||
y = y * torch.rsqrt(y.pow(2).mean(dim=-1, keepdim=True) + self.eps)
|
||||
return y.to(x.dtype) * _cast(self.weight, x)
|
||||
|
||||
|
||||
def _heads_first(blocks: torch.Tensor, heads: int) -> torch.Tensor:
|
||||
count, tokens, dim = blocks.shape
|
||||
return blocks.view(count, tokens, heads, dim // heads).permute(2, 0, 1, 3).contiguous()
|
||||
|
||||
|
||||
class _SelfAttention(nn.Module):
|
||||
def __init__(self, dim, heads, eps, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.heads = heads
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.q = operations.Linear(dim, dim, **kw)
|
||||
self.k = operations.Linear(dim, dim, **kw)
|
||||
self.v = operations.Linear(dim, dim, **kw)
|
||||
self.o = operations.Linear(dim, dim, **kw)
|
||||
self.norm_q = _RMSNorm(dim, eps, **kw)
|
||||
self.norm_k = _RMSNorm(dim, eps, **kw)
|
||||
|
||||
def forward(self, x, rope, grid, memory: dict, plan: AttentionPlan):
|
||||
frames, rows, cols = grid
|
||||
cos, sin = rope
|
||||
q = to_blocks(apply_rope(self.norm_q(self.q(x))[0], cos, sin, self.heads), *grid)
|
||||
k = to_blocks(apply_rope(self.norm_k(self.k(x))[0], cos, sin, self.heads), *grid)
|
||||
v = to_blocks(self.v(x)[0], *grid)
|
||||
if memory.get("k") is not None:
|
||||
k = torch.cat([memory["k"], k], dim=0)
|
||||
v = torch.cat([memory["v"], v], dim=0)
|
||||
chosen, shares = choose_key_blocks(q, k, self.heads, plan.local, frames // 2, plan.keep)
|
||||
out = sparse_block_attention(
|
||||
_heads_first(q, self.heads),
|
||||
_heads_first(k, self.heads),
|
||||
_heads_first(v, self.heads),
|
||||
chosen,
|
||||
shares,
|
||||
plan.slot_budget,
|
||||
)
|
||||
if k.shape[0] // plan.per_group > KEPT_GROUPS:
|
||||
k, v = k[plan.per_group :], v[plan.per_group :]
|
||||
memory["k"], memory["v"] = k, v
|
||||
out = out.permute(1, 2, 0, 3).reshape(-1, BLOCK_TOKENS, x.shape[-1])
|
||||
return self.o(from_blocks(out, *grid)[None])
|
||||
|
||||
|
||||
class _CrossAttention(nn.Module):
|
||||
def __init__(self, dim, heads, eps, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.heads = heads
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.q = operations.Linear(dim, dim, **kw)
|
||||
self.k = operations.Linear(dim, dim, **kw)
|
||||
self.v = operations.Linear(dim, dim, **kw)
|
||||
self.o = operations.Linear(dim, dim, **kw)
|
||||
self.norm_q = _RMSNorm(dim, eps, **kw)
|
||||
self.norm_k = _RMSNorm(dim, eps, **kw)
|
||||
|
||||
def _split(self, x: torch.Tensor) -> torch.Tensor:
|
||||
batch, tokens, dim = x.shape
|
||||
return x.view(batch, tokens, self.heads, dim // self.heads).transpose(1, 2)
|
||||
|
||||
def forward(self, x, context, memory: dict):
|
||||
if memory.get("k") is None: # the prompt never changes during a run
|
||||
memory["k"] = self._split(self.norm_k(self.k(context)))
|
||||
memory["v"] = self._split(self.v(context))
|
||||
out = F.scaled_dot_product_attention(self._split(self.norm_q(self.q(x))), memory["k"], memory["v"])
|
||||
return self.o(out.transpose(1, 2).reshape(x.shape))
|
||||
|
||||
|
||||
class _Block(nn.Module):
|
||||
def __init__(self, dim, heads, ffn_dim, eps, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.dim, self.eps = dim, eps
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.self_attn = _SelfAttention(dim, heads, eps, operations, **kw)
|
||||
self.cross_attn = _CrossAttention(dim, heads, eps, operations, **kw)
|
||||
self.norm3 = operations.LayerNorm(dim, eps=eps, elementwise_affine=True, **kw)
|
||||
self.ffn = nn.Sequential(
|
||||
operations.Linear(dim, ffn_dim, **kw), nn.GELU(approximate="tanh"), operations.Linear(ffn_dim, dim, **kw)
|
||||
)
|
||||
self.modulation = nn.Parameter(torch.zeros(1, 6, dim, **kw))
|
||||
|
||||
def _plain_norm(self, x: torch.Tensor) -> torch.Tensor:
|
||||
return F.layer_norm(x, (self.dim,), eps=self.eps)
|
||||
|
||||
def forward(self, x, context, time_mod, rope, grid, memory: dict, plan: AttentionPlan):
|
||||
shift_a, scale_a, gate_a, shift_m, scale_m, gate_m = (_cast(self.modulation, x) + time_mod).chunk(6, dim=1)
|
||||
x = x + gate_a * self.self_attn(self._plain_norm(x) * (1 + scale_a) + shift_a, rope, grid, memory["self"], plan)
|
||||
x = x + self.cross_attn(self.norm3(x), context, memory["cross"])
|
||||
return x + gate_m * self.ffn(self._plain_norm(x) * (1 + scale_m) + shift_m)
|
||||
|
||||
|
||||
class _Head(nn.Module):
|
||||
def __init__(self, dim, out_dim, eps, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.dim, self.eps = dim, eps
|
||||
self.head = operations.Linear(dim, out_dim * 4, device=device, dtype=dtype)
|
||||
self.modulation = nn.Parameter(torch.zeros(1, 2, dim, device=device, dtype=dtype))
|
||||
|
||||
def forward(self, x, time_embed):
|
||||
shift, scale = (_cast(self.modulation, x) + time_embed).chunk(2, dim=1)
|
||||
return self.head(F.layer_norm(x, (self.dim,), eps=self.eps) * (1 + scale) + shift)
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# The low-quality projector
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
class _ChannelNorm(nn.Module):
|
||||
def __init__(self, dim: int, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.scale = dim**0.5
|
||||
self.gamma = nn.Parameter(torch.ones(dim, 1, 1, 1, device=device, dtype=dtype))
|
||||
|
||||
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
||||
return F.normalize(x.float(), dim=1).to(x.dtype) * self.scale * _cast(self.gamma, x)
|
||||
|
||||
|
||||
class LowQualityProjector(nn.Module):
|
||||
"""Turns enlarged low-quality frames into one set of tokens per 4 frames.
|
||||
|
||||
Two convolutions that only look backwards in time, each halving the
|
||||
frame rate, then one linear layer. `push` is fed the first frame alone,
|
||||
then 4 frames at a time, and remembers what the next call needs.
|
||||
"""
|
||||
|
||||
def __init__(self, out_dim=1536, hidden=(2048, 3072), patch=16, operations=nn, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.patch = patch
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.conv1 = operations.Conv3d(3 * patch * patch, hidden[0], (4, 3, 3), stride=(2, 1, 1), **kw)
|
||||
self.norm1 = _ChannelNorm(hidden[0], **kw)
|
||||
self.conv2 = operations.Conv3d(hidden[0], hidden[1], (4, 3, 3), stride=(2, 1, 1), **kw)
|
||||
self.norm2 = _ChannelNorm(hidden[1], **kw)
|
||||
self.linear_layers = nn.ModuleList([operations.Linear(hidden[1], out_dim, **kw)])
|
||||
|
||||
def _fold(self, clip: torch.Tensor) -> torch.Tensor:
|
||||
batch, channels, frames, height, width = clip.shape
|
||||
p = self.patch
|
||||
x = clip.view(batch, channels, frames, height // p, p, width // p, p).permute(0, 1, 4, 6, 2, 3, 5)
|
||||
return x.reshape(batch, channels * p * p, frames, height // p, width // p)
|
||||
|
||||
@staticmethod
|
||||
def _causal(conv, x: torch.Tensor, earlier) -> torch.Tensor:
|
||||
front = 2
|
||||
if earlier is not None:
|
||||
x = torch.cat([earlier, x], dim=2)
|
||||
front -= earlier.shape[2]
|
||||
return conv(F.pad(x, (1, 1, 1, 1, front, 0), mode="replicate"))
|
||||
|
||||
def push(self, memory: dict, clip: torch.Tensor):
|
||||
"""clip is [1, 3, frames, height, width] in -1..1. Returns tokens, or None for the first frame."""
|
||||
first = not memory
|
||||
if first:
|
||||
clip = torch.cat([clip[:, :, :1].repeat(1, 1, 3, 1, 1), clip], dim=2)
|
||||
x = self._fold(clip)
|
||||
tail = x[:, :, -2:].clone()
|
||||
x = F.silu(self.norm1(self._causal(self.conv1, x, memory.get("conv1"))))
|
||||
memory["conv1"] = tail
|
||||
tail = x[:, :, -2:].clone()
|
||||
if first:
|
||||
memory["conv2"] = tail
|
||||
return None
|
||||
x = F.silu(self.norm2(self._causal(self.conv2, x, memory["conv2"])))
|
||||
memory["conv2"] = tail
|
||||
return self.linear_layers[0](x.flatten(2).transpose(1, 2))
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# The fast decoder
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
class _SoftClamp(nn.Module):
|
||||
def forward(self, x):
|
||||
return torch.tanh(x / 3) * 3
|
||||
|
||||
|
||||
class _MemoryBlock(nn.Module):
|
||||
"""A residual block that also sees the frame before."""
|
||||
|
||||
def __init__(self, width: int, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.conv = nn.Sequential(
|
||||
operations.Conv2d(width * 2, width, 3, padding=1, **kw),
|
||||
nn.ReLU(),
|
||||
operations.Conv2d(width, width, 3, padding=1, **kw),
|
||||
nn.ReLU(),
|
||||
operations.Conv2d(width, width, 3, padding=1, **kw),
|
||||
)
|
||||
|
||||
def forward(self, x, past):
|
||||
return F.relu(self.conv(torch.cat([x, past], dim=1)) + x)
|
||||
|
||||
|
||||
class _TimeGrow(nn.Module):
|
||||
"""Makes `stride` frames out of each frame."""
|
||||
|
||||
def __init__(self, width: int, stride: int, operations, device=None, dtype=None):
|
||||
super().__init__()
|
||||
self.conv = operations.Conv2d(width, width * stride, 1, bias=False, device=device, dtype=dtype)
|
||||
|
||||
def forward(self, x):
|
||||
_, channels, height, width = x.shape
|
||||
return self.conv(x).reshape(-1, channels, height, width)
|
||||
|
||||
|
||||
class _Upsample(nn.Module):
|
||||
def forward(self, x):
|
||||
return F.interpolate(x, scale_factor=2, mode="nearest")
|
||||
|
||||
|
||||
class ConditionedDecoder(nn.Module):
|
||||
"""Latent frames plus the low-quality frames in, 4 video frames per latent frame out."""
|
||||
|
||||
def __init__(self, latent_channels=16, widths=(512, 256, 128, 128), operations=nn, device=None, dtype=None):
|
||||
super().__init__()
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
a, b, c, d = widths
|
||||
|
||||
def conv(n_in, n_out, bias=True):
|
||||
return operations.Conv2d(n_in, n_out, 3, padding=1, bias=bias, **kw)
|
||||
|
||||
def memory_blocks(width):
|
||||
return [_MemoryBlock(width, operations, **kw) for _ in range(3)]
|
||||
|
||||
self.latent_channels = latent_channels
|
||||
self.decoder = nn.Sequential(
|
||||
_SoftClamp(),
|
||||
conv(latent_channels + 3 * 4 * 8 * 8, a),
|
||||
nn.ReLU(),
|
||||
conv(a, a, bias=False),
|
||||
nn.ReLU(),
|
||||
*memory_blocks(a),
|
||||
_Upsample(),
|
||||
_TimeGrow(a, 1, operations, **kw),
|
||||
conv(a, b, bias=False),
|
||||
*memory_blocks(b),
|
||||
_Upsample(),
|
||||
_TimeGrow(b, 2, operations, **kw),
|
||||
conv(b, c, bias=False),
|
||||
*memory_blocks(c),
|
||||
_Upsample(),
|
||||
_TimeGrow(c, 2, operations, **kw),
|
||||
conv(c, d, bias=False),
|
||||
nn.ReLU(),
|
||||
conv(d, d, bias=False),
|
||||
nn.ReLU(),
|
||||
conv(d, 3),
|
||||
)
|
||||
|
||||
@staticmethod
|
||||
def fold_frames(clip: torch.Tensor) -> torch.Tensor:
|
||||
"""[3, 4, height, width] -> [768, height / 8, width / 8]."""
|
||||
channels, frames, height, width = clip.shape
|
||||
x = clip.view(channels, frames, height // 8, 8, width // 8, 8).permute(0, 1, 3, 5, 2, 4)
|
||||
return x.reshape(channels * frames * 64, height // 8, width // 8)
|
||||
|
||||
def decode(self, memory: dict, x: torch.Tensor) -> torch.Tensor:
|
||||
"""x is [latent frames, 768 + latent channels, h, w]. Returns [4 x latent frames, 3, 8h, 8w] in 0..1.
|
||||
|
||||
The first 3 frames of the first call are lead-in and are dropped.
|
||||
"""
|
||||
first = not memory
|
||||
memory["started"] = True
|
||||
for index, layer in enumerate(self.decoder):
|
||||
if isinstance(layer, _MemoryBlock):
|
||||
before = memory.get(index)
|
||||
if before is None:
|
||||
before = torch.zeros_like(x[:1])
|
||||
memory[index] = x[-1:].clone()
|
||||
x = layer(x, torch.cat([before, x[:-1]], dim=0))
|
||||
else:
|
||||
x = layer(x)
|
||||
return x[3:] if first else x
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# The whole model
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
|
||||
class FlashVSRModel(nn.Module):
|
||||
def __init__(
|
||||
self,
|
||||
dim=1536,
|
||||
heads=12,
|
||||
ffn_dim=8960,
|
||||
layers=30,
|
||||
latent_channels=16,
|
||||
text_dim=4096,
|
||||
text_len=512,
|
||||
freq_dim=256,
|
||||
eps=1e-6,
|
||||
projector_hidden=(2048, 3072),
|
||||
decoder_widths=(512, 256, 128, 128),
|
||||
operations=nn,
|
||||
device=None,
|
||||
dtype=None,
|
||||
):
|
||||
super().__init__()
|
||||
self.dim, self.heads, self.freq_dim, self.latent_channels = dim, heads, freq_dim, latent_channels
|
||||
kw = {"device": device, "dtype": dtype}
|
||||
self.patch_embedding = operations.Conv3d(latent_channels, dim, (1, 2, 2), stride=(1, 2, 2), **kw)
|
||||
self.text_embedding = nn.Sequential(
|
||||
operations.Linear(text_dim, dim, **kw), nn.GELU(approximate="tanh"), operations.Linear(dim, dim, **kw)
|
||||
)
|
||||
self.time_embedding = nn.Sequential(
|
||||
operations.Linear(freq_dim, dim, **kw), nn.SiLU(), operations.Linear(dim, dim, **kw)
|
||||
)
|
||||
self.time_projection = nn.Sequential(nn.SiLU(), operations.Linear(dim, dim * 6, **kw))
|
||||
self.blocks = nn.ModuleList([_Block(dim, heads, ffn_dim, eps, operations, **kw) for _ in range(layers)])
|
||||
self.head = _Head(dim, latent_channels, eps, operations, **kw)
|
||||
self.lq_proj_in = LowQualityProjector(dim, projector_hidden, operations=operations, **kw)
|
||||
self.tc_decoder = ConditionedDecoder(latent_channels, decoder_widths, operations=operations, **kw)
|
||||
self.register_buffer("positive_conditioning", torch.zeros(text_len, text_dim, **kw))
|
||||
|
||||
def start(self, like: torch.Tensor) -> dict:
|
||||
"""The values that never change during one clip."""
|
||||
half = self.freq_dim // 2
|
||||
steps = torch.pow(10000.0, -torch.arange(half, dtype=torch.float64) / half) * TIMESTEP
|
||||
wave = torch.cat([torch.cos(steps), torch.sin(steps)])[None].to(device=like.device, dtype=like.dtype)
|
||||
time_embed = self.time_embedding(wave)
|
||||
return {
|
||||
"time_embed": time_embed[:, None],
|
||||
"time_mod": self.time_projection(time_embed).unflatten(1, (6, self.dim)),
|
||||
"context": self.text_embedding(_cast(self.positive_conditioning, like)[None]),
|
||||
"layers": [{"self": {}, "cross": {}} for _ in self.blocks],
|
||||
"projector": {},
|
||||
"decoder": {},
|
||||
}
|
||||
|
||||
def denoise(self, state: dict, noise: torch.Tensor, lq_tokens: torch.Tensor, first_frame: int, plan: AttentionPlan):
|
||||
"""One pass over a few latent frames. Returns the clean latent frames."""
|
||||
x = self.patch_embedding(noise)
|
||||
frames, rows, cols = x.shape[2:]
|
||||
grid = (frames, rows, cols)
|
||||
x = x.flatten(2).transpose(1, 2) + lq_tokens
|
||||
rope = rope_table(self.dim // self.heads, first_frame, frames, rows, cols, x.device)
|
||||
for block, memory in zip(self.blocks, state["layers"]):
|
||||
x = block(x, state["context"], state["time_mod"], rope, grid, memory, plan)
|
||||
x = self.head(x, state["time_embed"])
|
||||
x = x.view(1, frames, rows, cols, 2, 2, self.latent_channels).permute(0, 6, 1, 2, 4, 3, 5)
|
||||
return noise - x.reshape(1, self.latent_channels, frames, rows * 2, cols * 2)
|
||||
|
||||
|
||||
def pass_layout(groups: int):
|
||||
"""(first latent frame, latent frames) of each pass over a clip of `groups` 8-frame groups."""
|
||||
yield 0, FIRST_PASS_LATENTS
|
||||
for index in range(1, groups - 2):
|
||||
yield 4 + 2 * index, PASS_LATENTS
|
||||
|
||||
|
||||
def match_colors(frames: torch.Tensor, source: torch.Tensor) -> torch.Tensor:
|
||||
"""Give each frame its source frame's mean and spread, per colour channel. Both [n, 3, h, w]."""
|
||||
f, s = frames.float(), source.float()
|
||||
f_mean, s_mean = f.mean(dim=(2, 3), keepdim=True), s.mean(dim=(2, 3), keepdim=True)
|
||||
f_std, s_std = f.std(dim=(2, 3), keepdim=True) + 1e-5, s.std(dim=(2, 3), keepdim=True) + 1e-5
|
||||
return (f - f_mean) / f_std * s_std + s_mean
|
||||
|
||||
|
||||
TILE_OVERLAP = 128
|
||||
|
||||
|
||||
def plan_tiles(work_h: int, work_w: int, max_pixels: int | None) -> list[tuple[int, int, int, int]]:
|
||||
"""Cut the working size into as few overlapping tiles as fit `max_pixels` each.
|
||||
|
||||
Returns (top, left, height, width) boxes. Tile sides are multiples of
|
||||
128 and corners sit on the 16-pixel token grid. One box, the whole
|
||||
picture, when it fits or when no limit is given.
|
||||
"""
|
||||
if not max_pixels or work_h * work_w <= max_pixels:
|
||||
return [(0, 0, work_h, work_w)]
|
||||
|
||||
def side(total: int, parts: int) -> int:
|
||||
if parts == 1:
|
||||
return total
|
||||
return min(total, math.ceil((total + (parts - 1) * TILE_OVERLAP) / parts / PIXEL_STEP) * PIXEL_STEP)
|
||||
|
||||
best = None
|
||||
for rows in range(1, work_h // PIXEL_STEP + 1):
|
||||
for cols in range(1, work_w // PIXEL_STEP + 1):
|
||||
tile_h, tile_w = side(work_h, rows), side(work_w, cols)
|
||||
if tile_h * tile_w <= max_pixels and (best is None or rows * cols < best[0] * best[1]):
|
||||
best = (rows, cols, tile_h, tile_w)
|
||||
if best is None: # even one 128x128 tile is over the limit: use the smallest tiles there are
|
||||
best = (work_h // PIXEL_STEP, work_w // PIXEL_STEP, PIXEL_STEP, PIXEL_STEP)
|
||||
rows, cols, tile_h, tile_w = best
|
||||
|
||||
def starts(total: int, tile: int, parts: int) -> list[int]:
|
||||
if parts == 1:
|
||||
return [0]
|
||||
return [round(i * (total - tile) / (parts - 1) / 16) * 16 for i in range(parts)]
|
||||
|
||||
return [(top, left, tile_h, tile_w) for top in starts(work_h, tile_h, rows) for left in starts(work_w, tile_w, cols)]
|
||||
|
||||
|
||||
def tile_weight(box: tuple[int, int, int, int], work_h: int, work_w: int) -> torch.Tensor:
|
||||
"""[height, width] blend weights for a tile: 1 inside, fading over the overlap on edges that meet another tile."""
|
||||
top, left, height, width = box
|
||||
|
||||
def ramp(length: int, fade_start: bool, fade_end: bool) -> torch.Tensor:
|
||||
w = torch.ones(length)
|
||||
fade = (torch.arange(TILE_OVERLAP) + 0.5) / TILE_OVERLAP
|
||||
if fade_start:
|
||||
w[:TILE_OVERLAP] = fade
|
||||
if fade_end:
|
||||
w[-TILE_OVERLAP:] = torch.minimum(w[-TILE_OVERLAP:], fade.flip(0))
|
||||
return w
|
||||
|
||||
rows = ramp(height, top > 0, top + height < work_h)
|
||||
cols = ramp(width, left > 0, left + width < work_w)
|
||||
return rows[:, None] * cols[None, :]
|
||||
|
||||
|
||||
@torch.no_grad()
|
||||
def upscale_frames(
|
||||
model: FlashVSRModel,
|
||||
frames: torch.Tensor,
|
||||
scale: float,
|
||||
*,
|
||||
device,
|
||||
dtype,
|
||||
color_match: bool = True,
|
||||
seed: int = 0,
|
||||
slot_budget: int = DEFAULT_SLOT_BUDGET,
|
||||
max_tile_pixels: int | None = None,
|
||||
decode_latents=None,
|
||||
on_pass=None,
|
||||
) -> torch.Tensor:
|
||||
"""Upscale a clip. `frames` is [count, height, width, 3] in 0..1; so is the result, at the new size.
|
||||
|
||||
A picture larger than `max_tile_pixels` is done in overlapping tiles,
|
||||
one after the other, and blended; that is what bounds graphics memory.
|
||||
`decode_latents`, when given, takes the clean latent frames of a tile
|
||||
([1, 16, frames, h, w]) and returns [frames, height, width, 3]; it is how
|
||||
another decoder (the Wan 2.1 VAE) is used in place of the fast one.
|
||||
`on_pass(done, total)` is called after each pass.
|
||||
"""
|
||||
count, height, width, _ = frames.shape
|
||||
out_h, out_w, work_h, work_w = plan_sizes(height, width, scale)
|
||||
groups = plan_groups(count)
|
||||
layout = list(pass_layout(groups))
|
||||
tiles = plan_tiles(work_h, work_w, max_tile_pixels)
|
||||
|
||||
def enlarged(start: int, stop: int) -> torch.Tensor:
|
||||
"""Frames start..stop-1 of the padded clip at the working size: [n, 3, work_h, work_w] in 0..1."""
|
||||
index = torch.arange(start, stop).clamp(max=count - 1)
|
||||
x = frames[index][..., :3].to(device).movedim(-1, 1).float()
|
||||
if (height, width) != (out_h, out_w):
|
||||
x = F.interpolate(x, size=(out_h, out_w), mode="bicubic", align_corners=False).clamp(0, 1)
|
||||
return F.pad(x, (0, work_w - out_w, 0, work_h - out_h), mode="replicate")
|
||||
|
||||
noise = torch.randn(
|
||||
(1, model.latent_channels, 2 * groups, work_h // 8, work_w // 8), generator=torch.Generator().manual_seed(seed)
|
||||
)
|
||||
result = torch.zeros((count, out_h, out_w, 3), dtype=torch.float32)
|
||||
total_weight = torch.zeros((work_h, work_w), dtype=torch.float32)
|
||||
like = torch.zeros(1, device=device, dtype=dtype)
|
||||
done = 0
|
||||
for box in tiles:
|
||||
top, left, tile_h, tile_w = box
|
||||
weight = tile_weight(box, work_h, work_w)
|
||||
total_weight[top : top + tile_h, left : left + tile_w] += weight
|
||||
weight = weight.to(device)
|
||||
|
||||
def crop(start: int, stop: int) -> torch.Tensor:
|
||||
return enlarged(start, stop)[..., top : top + tile_h, left : left + tile_w]
|
||||
|
||||
def signed_clip(start: int, stop: int) -> torch.Tensor:
|
||||
return (crop(start, stop) * 2 - 1).movedim(0, 1)[None].to(dtype)
|
||||
|
||||
def store(decoded: torch.Tensor, written: int) -> int:
|
||||
"""Add decoded frames [n, 3, tile_h, tile_w] to the result; returns how many frames the tile has given."""
|
||||
decoded = decoded[: count - written].float()
|
||||
if decoded.shape[0] == 0:
|
||||
return written
|
||||
stop = written + decoded.shape[0]
|
||||
if color_match:
|
||||
decoded = match_colors(decoded, crop(written, stop))
|
||||
bottom, right = min(top + tile_h, out_h), min(left + tile_w, out_w)
|
||||
if bottom > top and right > left:
|
||||
piece = (decoded.clamp(0, 1) * weight)[..., : bottom - top, : right - left]
|
||||
result[written:stop, top:bottom, left:right] += piece.movedim(1, -1).cpu()
|
||||
return stop
|
||||
|
||||
state = model.start(like)
|
||||
plan = AttentionPlan(tile_h // 16, tile_w // 16, device, slot_budget)
|
||||
tile_noise = noise[..., top // 8 : (top + tile_h) // 8, left // 8 : (left + tile_w) // 8]
|
||||
kept_latents = []
|
||||
written = 0
|
||||
model.lq_proj_in.push(state["projector"], signed_clip(0, 1))
|
||||
for first, latent_count in layout:
|
||||
tokens = torch.cat(
|
||||
[
|
||||
model.lq_proj_in.push(state["projector"], signed_clip(4 * i + 1, 4 * i + 5))
|
||||
for i in range(first, first + latent_count)
|
||||
],
|
||||
dim=1,
|
||||
)
|
||||
chunk = tile_noise[:, :, first : first + latent_count].to(device=device, dtype=dtype)
|
||||
latents = model.denoise(state, chunk, tokens, first, plan)
|
||||
if decode_latents is not None:
|
||||
kept_latents.append(latents.float().cpu())
|
||||
else:
|
||||
for i in range(first, first + latent_count):
|
||||
if written >= count:
|
||||
break
|
||||
source = signed_clip(0, 1)[0].repeat(1, 4, 1, 1) if i == 0 else signed_clip(4 * i - 3, 4 * i + 1)[0]
|
||||
both = torch.cat([ConditionedDecoder.fold_frames(source), latents[0, :, i - first]], dim=0)
|
||||
written = store(model.tc_decoder.decode(state["decoder"], both[None]), written)
|
||||
done += 1
|
||||
if on_pass is not None:
|
||||
on_pass(done, len(layout) * len(tiles))
|
||||
state.clear() # the layers' memory is no longer needed; the next tile or another decoder wants the room
|
||||
if decode_latents is not None:
|
||||
decoded = decode_latents(torch.cat(kept_latents, dim=2)).movedim(-1, 1)
|
||||
for start in range(0, min(count, decoded.shape[0]), 8):
|
||||
written = store(decoded[start : start + 8].to(device), written)
|
||||
if len(tiles) > 1:
|
||||
result /= total_weight[:out_h, :out_w, None]
|
||||
return result
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------
|
||||
# ComfyUI side: loading the file and running a clip. ComfyUI is imported
|
||||
# inside the functions so the code above runs without it.
|
||||
# --------------------------------------------------------------------------
|
||||
|
||||
MODEL_PARTS = {
|
||||
"blocks.0.self_attn.q.weight": "the transformer",
|
||||
"lq_proj_in.conv1.weight": "the low-quality projector",
|
||||
"tc_decoder.decoder.1.weight": "the fast decoder",
|
||||
"positive_conditioning": "the stored prompt",
|
||||
}
|
||||
NOT_A_FLASHVSR_FILE = (
|
||||
"This is not a FlashVSR model file for this node: it has no {missing}. "
|
||||
"Pick the FlashVSR file named in the node's help page."
|
||||
)
|
||||
|
||||
_loaded: dict = {}
|
||||
|
||||
|
||||
# Measured on one clip (640x1024, bf16): about 9.6 KB of graphics memory per working pixel on top of
|
||||
# the weights. Most of it is the keys and values every layer keeps of the last few frames.
|
||||
WORKING_BYTES_PER_PIXEL = 12_000
|
||||
MEMORY_TO_ASK_FOR = 16 * 1024**3
|
||||
MEMORY_HEADROOM = 1024**3
|
||||
|
||||
|
||||
def working_memory(pixels: int, bytes_per_value: int = 2) -> int:
|
||||
"""A rough count of the graphics memory a picture of `pixels` needs besides the weights, in bytes."""
|
||||
return int(pixels * WORKING_BYTES_PER_PIXEL * bytes_per_value / 2)
|
||||
|
||||
|
||||
def tile_pixels_that_fit(free_bytes: int, bytes_per_value: int = 2) -> int:
|
||||
"""The largest tile, in pixels, that `free_bytes` of graphics memory can hold."""
|
||||
return max(PIXEL_STEP * PIXEL_STEP, int((free_bytes - MEMORY_HEADROOM) / working_memory(1_000_000, bytes_per_value) * 1_000_000))
|
||||
|
||||
|
||||
def load_flashvsr(path: str, dtype: torch.dtype | None = None):
|
||||
"""The model behind ComfyUI's memory manager, and the number type it runs in. Kept between runs."""
|
||||
import comfy.model_management as mm
|
||||
import comfy.model_patcher
|
||||
import comfy.ops
|
||||
import comfy.utils
|
||||
|
||||
stamp = (path, os.path.getmtime(path), dtype)
|
||||
if _loaded.get("stamp") == stamp:
|
||||
return _loaded["patcher"], _loaded["dtype"]
|
||||
_loaded.clear()
|
||||
weights = comfy.utils.load_torch_file(path, safe_load=True)
|
||||
missing = [what for key, what in MODEL_PARTS.items() if key not in weights]
|
||||
if missing:
|
||||
raise ValueError(NOT_A_FLASHVSR_FILE.format(missing=" or ".join(missing)))
|
||||
device, offload = mm.get_torch_device(), mm.unet_offload_device()
|
||||
if dtype is None:
|
||||
# The file is 16-bit, so that is the first choice: no weight has to be converted.
|
||||
if mm.should_use_fp16(device):
|
||||
dtype = torch.float16
|
||||
elif mm.should_use_bf16(device):
|
||||
dtype = torch.bfloat16
|
||||
else:
|
||||
dtype = torch.float32
|
||||
# manual_cast: every layer brings its weights to the device and number type of what it is given,
|
||||
# so a model ComfyUI keeps partly, or wholly, outside graphics memory still runs.
|
||||
model = FlashVSRModel(operations=comfy.ops.manual_cast, device=offload, dtype=dtype)
|
||||
model.eval()
|
||||
# ComfyUI's own wrapper for a model: the one that loads weights as they are needed, where it has one.
|
||||
wrapper = getattr(comfy.model_patcher, "CoreModelPatcher", comfy.model_patcher.ModelPatcher)
|
||||
patcher = wrapper(model, load_device=device, offload_device=offload)
|
||||
assign = bool(getattr(patcher, "is_dynamic", lambda: False)())
|
||||
model.load_state_dict(weights, strict=True, assign=assign)
|
||||
model.requires_grad_(False)
|
||||
_loaded.update(stamp=stamp, patcher=patcher, dtype=dtype)
|
||||
return patcher, dtype
|
||||
|
||||
|
||||
def _memory_for_work(patcher, device) -> int:
|
||||
"""Graphics memory left for the picture once all of the model's weights are in it."""
|
||||
import comfy.model_management as mm
|
||||
|
||||
free = mm.get_free_memory(device)
|
||||
try:
|
||||
still_to_load = max(0, int(patcher.model_size()) - int(patcher.loaded_size()))
|
||||
except Exception:
|
||||
still_to_load = 0
|
||||
return free - still_to_load
|
||||
|
||||
|
||||
def run_flashvsr(
|
||||
path: str,
|
||||
frames: torch.Tensor,
|
||||
scale: float,
|
||||
color_match: bool = True,
|
||||
vae=None,
|
||||
seed: int = 0,
|
||||
max_tile_pixels: int | None = None,
|
||||
):
|
||||
"""Upscale `frames` with the model file at `path`. `vae`, when given, decodes in place of the fast decoder."""
|
||||
import comfy.latent_formats
|
||||
import comfy.model_management as mm
|
||||
|
||||
from ._execution_helpers import progress_bar, raise_if_interrupted
|
||||
|
||||
patcher, dtype = load_flashvsr(path)
|
||||
device = patcher.load_device
|
||||
_, _, work_h, work_w = plan_sizes(frames.shape[1], frames.shape[2], scale)
|
||||
wanted = working_memory(work_h * work_w, dtype.itemsize)
|
||||
mm.load_models_gpu([patcher], memory_required=min(wanted, MEMORY_TO_ASK_FOR))
|
||||
if max_tile_pixels is None and device.type != "cpu":
|
||||
max_tile_pixels = tile_pixels_that_fit(_memory_for_work(patcher, device), dtype.itemsize)
|
||||
passes = len(list(pass_layout(plan_groups(frames.shape[0]))))
|
||||
|
||||
decode = None
|
||||
if vae is not None:
|
||||
latent_format = comfy.latent_formats.Wan21()
|
||||
|
||||
def decode(latents: torch.Tensor) -> torch.Tensor:
|
||||
images = vae.decode(latent_format.process_out(latents))
|
||||
return images.reshape(-1, *images.shape[-3:])
|
||||
|
||||
while True:
|
||||
progress = progress_bar(passes * len(plan_tiles(work_h, work_w, max_tile_pixels)))
|
||||
|
||||
def on_pass(done: int, total: int) -> None:
|
||||
raise_if_interrupted()
|
||||
if progress is not None:
|
||||
progress.update_absolute(done, total)
|
||||
|
||||
try:
|
||||
return upscale_frames(
|
||||
patcher.model,
|
||||
frames,
|
||||
scale,
|
||||
device=device,
|
||||
dtype=dtype,
|
||||
color_match=color_match,
|
||||
seed=seed,
|
||||
max_tile_pixels=max_tile_pixels,
|
||||
decode_latents=decode,
|
||||
on_pass=on_pass,
|
||||
)
|
||||
except torch.cuda.OutOfMemoryError:
|
||||
# Out of graphics memory: start again in tiles half the size, down to the smallest tile.
|
||||
tried = max_tile_pixels if max_tile_pixels is not None else work_h * work_w
|
||||
if tried <= PIXEL_STEP * PIXEL_STEP:
|
||||
raise
|
||||
max_tile_pixels = max(PIXEL_STEP * PIXEL_STEP, min(tried, work_h * work_w) // 2)
|
||||
mm.soft_empty_cache()
|
||||
|
||||
|
||||
MODEL_FOLDER = "diffusion_models"
|
||||
EXPECTED_MODEL_NAME = "flashvsr_v1.1_fp16.safetensors"
|
||||
MODEL_NOT_FOUND = (
|
||||
"FlashVSR Upscale could not find {name} in ComfyUI/models/diffusion_models. "
|
||||
"Download the FlashVSR model file named in the node's help page and put it there."
|
||||
)
|
||||
|
||||
|
||||
def list_flashvsr_models() -> list[str]:
|
||||
"""FlashVSR files in the diffusion models folder; every file there when none is named FlashVSR."""
|
||||
try:
|
||||
import folder_paths
|
||||
except ImportError: # Offline tests import this module without ComfyUI.
|
||||
return [EXPECTED_MODEL_NAME]
|
||||
names = list(folder_paths.get_filename_list(MODEL_FOLDER))
|
||||
ours = [name for name in names if "flashvsr" in name.lower().replace("_", "").replace("-", "")]
|
||||
return ours or names or [EXPECTED_MODEL_NAME]
|
||||
|
||||
|
||||
def resolve_flashvsr_model(name: str) -> str:
|
||||
import folder_paths
|
||||
|
||||
path = folder_paths.get_full_path(MODEL_FOLDER, name)
|
||||
if not path:
|
||||
raise FileNotFoundError(MODEL_NOT_FOUND.format(name=name))
|
||||
return path
|
||||
@@ -60,6 +60,26 @@ def add_module_version(source, token):
|
||||
return _MODULE_IMPORT.sub(lambda m: f"{m.group(1)}{m.group(2)}{m.group(3)}?v={token}{m.group(2)}", source)
|
||||
|
||||
|
||||
def list_served_scripts(js_root):
|
||||
"""The pack's own scripts by their address below the js folder:
|
||||
{"shared/index.mjs": <path>}. Made once at startup, so a request is
|
||||
only looked up in it and the address a browser sends never reaches the
|
||||
disk: no whole path, network share or ".." is ever opened. Links are
|
||||
left out, and anything that resolves outside the js folder."""
|
||||
root = Path(js_root).resolve()
|
||||
scripts = {}
|
||||
for path in sorted(root.rglob("*")):
|
||||
try:
|
||||
if path.suffix not in (".js", ".mjs") or path.is_symlink() or not path.is_file():
|
||||
continue
|
||||
if not path.resolve().is_relative_to(root):
|
||||
continue
|
||||
except OSError:
|
||||
continue
|
||||
scripts[path.relative_to(root).as_posix()] = path
|
||||
return scripts
|
||||
|
||||
|
||||
def make_revalidate_middleware(pack_name, js_root=None, token=None):
|
||||
"""aiohttp middleware for this pack's scripts, nothing else touched.
|
||||
|
||||
@@ -69,7 +89,7 @@ def make_revalidate_middleware(pack_name, js_root=None, token=None):
|
||||
from aiohttp import web
|
||||
|
||||
marker = f"/extensions/{pack_name}/"
|
||||
root = Path(js_root).resolve() if js_root and token else None
|
||||
scripts = list_served_scripts(js_root) if js_root and token else None
|
||||
served = {}
|
||||
|
||||
def versioned(path):
|
||||
@@ -90,15 +110,15 @@ def make_revalidate_middleware(pack_name, js_root=None, token=None):
|
||||
|
||||
def tagged_response(request):
|
||||
"""The tagged script for this request, or None to let ComfyUI serve it."""
|
||||
if root is None or request.method not in ("GET", "HEAD"):
|
||||
if not scripts or request.method not in ("GET", "HEAD"):
|
||||
return None
|
||||
start = request.path.find(marker)
|
||||
if start < 0 or not request.path.endswith((".js", ".mjs")):
|
||||
return None
|
||||
target = scripts.get(request.path[start + len(marker):])
|
||||
if target is None:
|
||||
return None
|
||||
try:
|
||||
target = (root / request.path[start + len(marker):]).resolve()
|
||||
if not target.is_relative_to(root) or not target.is_file():
|
||||
return None
|
||||
found = versioned(target)
|
||||
except (OSError, ValueError, UnicodeDecodeError):
|
||||
return None
|
||||
|
||||
+230
-40
@@ -19,6 +19,10 @@ STYLE = "style"
|
||||
START = "start on it"
|
||||
END = "end on it"
|
||||
JOBS = (PERSON, SHEET, PLACE, OUTFIT, THING, STYLE, START, END)
|
||||
# Not a pick on the node: a thing picture that does the action of a move copy or a swap (a cartoon hand, a toy,
|
||||
# a mascot). It takes the place a person would, with its own shape and look.
|
||||
PERFORMER = "thing that moves"
|
||||
ACTORS = (PERSON, SHEET, PERFORMER)
|
||||
|
||||
MOVES = "copy its moves"
|
||||
SWAP = "swap the person"
|
||||
@@ -26,7 +30,7 @@ KEEP = "keep going"
|
||||
REVOICE = "new voice"
|
||||
EDIT = "change something"
|
||||
VIDEO_JOBS = (MOVES, SWAP, KEEP, REVOICE, EDIT)
|
||||
KEEPS_ITS_LENGTH = (SWAP, REVOICE, EDIT) # These change a clip, so the new one runs as long as the old one.
|
||||
KEEPS_ITS_LENGTH = (MOVES, SWAP, REVOICE, EDIT) # These follow a clip, so the new one runs as long as the old one.
|
||||
NEEDS_NO_WORDS = (SWAP, KEEP)
|
||||
|
||||
TOKEN = re.compile(r"@(\d)\b")
|
||||
@@ -35,14 +39,31 @@ MAIN = "@0" # The main person, for lines the node writes itself.
|
||||
BACKGROUND = "The background of {picture} is not kept."
|
||||
NOTHING_IN = "Connect a picture or a video. H3 needs one to start from."
|
||||
NEEDS_A_PERSON = {
|
||||
MOVES: "To copy the moves, connect a picture of your person and pick person for it.",
|
||||
SWAP: "To swap the person, connect a picture of the new person and pick person for it.",
|
||||
}
|
||||
# A hat or a shirt with no person picture: the user wants the video's own person to wear it.
|
||||
DRESS_THE_VIDEO_PERSON = 'To put this on the person in your video, pick the video job "change something" and type "He wears @1".'
|
||||
|
||||
SWAP_NEEDS_PERSON_THING = (
|
||||
'A swap needs a picture of the new person, and pick person for it. To show a thing in the clip, '
|
||||
'pick the video job "change something" and type "He holds @1".'
|
||||
)
|
||||
SWAP_NEEDS_PERSON_PLACE = (
|
||||
'A swap keeps the place of the video. Connect a picture of the new person and pick person for it. '
|
||||
'To change the place, pick the video job "change something" and type "It looks like @1".'
|
||||
)
|
||||
REVOICE_KEEPS_PERSON = (
|
||||
"New voice keeps the person in the clip, so a person picture can't change who is in it. "
|
||||
'Take the picture off, or pick the video job "swap the person".'
|
||||
)
|
||||
HE_SHE = re.compile(r"\b(?:he|she|they)\b", re.I)
|
||||
|
||||
# (what the label is, how it is kept). {picture} is H3's number for the picture.
|
||||
JOB_WORDS = {
|
||||
PERFORMER: (
|
||||
"{subject} is the character in {picture}, with the same shape, colours and drawing style.",
|
||||
"fully_preserved - the shape, colours and drawing style are retained.",
|
||||
),
|
||||
PERSON: (
|
||||
"{subject} is the person in {picture}, with the same face, hair and clothes.",
|
||||
"fully_preserved - the face, hair and clothes are retained.",
|
||||
@@ -91,6 +112,18 @@ MOVES_KEPT = (
|
||||
"keep the source performer's face, body, clothes, props, background and colours out of the target.\n"
|
||||
"<Video 1>: attribute_transfer - apply only its camera movement to the target scene."
|
||||
)
|
||||
# Copy its moves with no person picture: the person in the video keeps their own moves and goes where the
|
||||
# script or a place picture says.
|
||||
OWN_MOVES_ARE = (
|
||||
"{main} is the person in <Video 1>, with the same face, hair, body{clothes} and the same moves, gestures and "
|
||||
"timing.\n<Video 1> provides the camera movement."
|
||||
)
|
||||
OWN_MOVES_KEPT = (
|
||||
"{main} (appears in {shots}): {kept}\n"
|
||||
"<Video 1>: attribute_transfer - apply its person, their moves and its camera movement to the target scene; {place}"
|
||||
)
|
||||
OWN_PLACE_FROM = "the background of <Video 1> is not kept, and the place is {place}."
|
||||
OWN_PLACE_KEPT = "keep its place unless the script names a new one."
|
||||
SWAP_VIDEO = "<Video 1> supplies the place, the action, the timing, the camera movement and the look."
|
||||
SWAP_IS = "{replaced} is {who} in <Video 1>, selected for replacement."
|
||||
SWAP_VIDEO_KEPT = (
|
||||
@@ -99,7 +132,28 @@ SWAP_VIDEO_KEPT = (
|
||||
)
|
||||
SWAP_KEPT = "{replaced}: removed - replace {who} with {main} throughout."
|
||||
ONE_GOES = "the original person"
|
||||
THE_PERSON = "the person in the shot"
|
||||
APPEARANCE_LINE = re.compile(r'^\s*@([1-4])\s+looks like\s+([^"\n]+?)\s*$', re.I)
|
||||
# Two ways people type a swap that used to be dropped without a word, so the swap ran with no description and H3
|
||||
# often gave the clip back: the looks sentence on the same line as the swap sentence, and the clothes on a line of
|
||||
# their own ("@1 is wearing a black suit"). A looks sentence starts after a full stop, ! or ?; "Put @1 looks like a
|
||||
# friend here." is one sentence and stays whole.
|
||||
LOOKS_MID_LINE = re.compile(r"(?<=[.!?])[ \t]*(?=@[1-4]\s+looks like\b)", re.I)
|
||||
WEARS_LINE = re.compile(r'^\s*@([1-4])\s+(?:is\s+wearing|wears|is\s+dressed\s+in|has\s+on)\s+([^"@\n]+?)\s*$', re.I)
|
||||
QUOTE = re.compile(r'["\u201c\u201d]')
|
||||
|
||||
|
||||
def looks_on_own_lines(text):
|
||||
"""The script with each @N looks like sentence on a line of its own. A spoken line is left alone."""
|
||||
text = str(text or "")
|
||||
lines = []
|
||||
for line in text.splitlines():
|
||||
if QUOTE.search(line) or not LOOKS_MID_LINE.search(line):
|
||||
lines.append(line)
|
||||
continue
|
||||
first, *rest = LOOKS_MID_LINE.sub("\n", line).splitlines()
|
||||
lines += [re.sub(r"\s+([.,;:])$", r"\1", first), *rest]
|
||||
return "\n".join(lines) if len(lines) != len(text.splitlines()) else text
|
||||
|
||||
|
||||
# Hair and face only, no clothes, expression or background. When the line names clothes that differ from what the
|
||||
@@ -115,6 +169,7 @@ def add_looks(text, jobs, describe):
|
||||
"""A swap's script with an @N looks like line for each new person who has none, written by describe(slot), a
|
||||
small picture-reading model shown that person's picture. Without such a line H3 kept the original person for the
|
||||
first seconds (2026-11-07), so a connected model writes it when the user didn't."""
|
||||
text = looks_on_own_lines(text)
|
||||
typed = {int(match.group(1)) for match in map(APPEARANCE_LINE.match, text.splitlines()) if match}
|
||||
for slot in sorted(jobs):
|
||||
if jobs[slot] not in (PERSON, SHEET) or slot in typed:
|
||||
@@ -137,8 +192,12 @@ def swap_appearances(text, jobs, video):
|
||||
"""
|
||||
if video != SWAP:
|
||||
return text, {}
|
||||
kept, descriptions = [], {}
|
||||
for line in str(text or "").splitlines():
|
||||
kept, descriptions, clothes = [], {}, {}
|
||||
for line in looks_on_own_lines(text).splitlines():
|
||||
worn = WEARS_LINE.fullmatch(line)
|
||||
if worn and (jobs or {}).get(int(worn.group(1))) in (PERSON, SHEET):
|
||||
clothes.setdefault(int(worn.group(1)), []).append(worn.group(2).strip().rstrip(" ."))
|
||||
continue
|
||||
match = APPEARANCE_LINE.fullmatch(line)
|
||||
if match is None:
|
||||
kept.append(line)
|
||||
@@ -150,10 +209,18 @@ def swap_appearances(text, jobs, video):
|
||||
raise ValueError(f"Picture {slot} must be a person or character sheet for a swap's looks like line.")
|
||||
if slot in descriptions:
|
||||
raise ValueError(f"Describe Picture {slot} once, in one looks like line.")
|
||||
description = match.group(2).strip().rstrip(" .")
|
||||
# A sentence on the looks line that names an @ again ("The trees behind @1 stay the same.") is not how
|
||||
# the person looks: it goes back to the script as its own line, or H3 reads a raw @1 in the appearance.
|
||||
sentences = re.split(r"(?<=[.!?])\s+", match.group(2).strip())
|
||||
looks = [sentence for sentence in sentences if not TOKEN.search(sentence)]
|
||||
kept.extend(sentence for sentence in sentences if TOKEN.search(sentence))
|
||||
description = " ".join(looks).strip().rstrip(" .")
|
||||
if not description:
|
||||
raise ValueError(f"Say what Picture {slot}'s person looks like, or remove its looks like line.")
|
||||
descriptions[slot] = description
|
||||
for slot, worn in clothes.items():
|
||||
# Clothes typed on their own line join the description; with none typed, the picture still gives the hair.
|
||||
descriptions[slot] = f"{descriptions.get(slot, 'the same hair')}, wearing {' and '.join(worn)}"
|
||||
return ("\n".join(kept), descriptions) if descriptions else (text, {})
|
||||
|
||||
|
||||
@@ -204,23 +271,24 @@ CLOTHES_STAY = (
|
||||
"attribute_transfer - apply only the face and hair to the original performance; the clothes of {who} in <Video 1> "
|
||||
"stay as they are; the background of {picture} is not transferred."
|
||||
)
|
||||
KEEP_IS = (
|
||||
"<Video 1> is the video being continued, and the target video starts from its last moment.\n"
|
||||
"{main} is the main subject in <Video 1>, with the same look."
|
||||
)
|
||||
KEEP_KEPT = (
|
||||
"<Video 1> (continuation point): fully_preserved - the place, the light, the colours and the camera carry on from its last moment.\n"
|
||||
"{main} (appears in {shots}): fully_preserved - the look is retained."
|
||||
)
|
||||
KEEP_IS = "<Video 1> is the video being continued, and the target video starts from its last moment."
|
||||
KEEP_MAIN = "{main} is the main subject in <Video 1>, with the same look."
|
||||
KEEP_KEPT = "<Video 1> (continuation point): fully_preserved - the place, the light, the colours and the camera carry on from its last moment."
|
||||
KEEP_KEPT_PLACE = "<Video 1> (continuation point): fully_preserved - the light, the colours and the camera carry on from its last moment."
|
||||
KEEP_MAIN_KEPT = "{main} (appears in {shots}): fully_preserved - the look is retained."
|
||||
SOUND_GOES_ON = "<Audio 1> is the soundtrack of <Video 1>, whose sound carries on in the target video."
|
||||
SOUND_GOES_ON_KEPT = "<Audio 1>: reference - the sound continues without copying the source signal."
|
||||
EDIT_IS = (
|
||||
"<Video 1> supplies the place, the person, the action, the timing, the camera movement, the look and the sound.\n"
|
||||
"{main} is the person in <Video 1>, kept as they are: the same face, hair, body and movements, with only what the script names changed."
|
||||
)
|
||||
EDIT_KEPT = (
|
||||
"<Video 1>: partially_preserved - keep the place, the person, the action, the timing, the camera movement, the look and the sound; change only what the script names.\n"
|
||||
"{main} (appears in {shots}): fully_preserved - the face, hair and every movement are retained; only what the script names changes."
|
||||
EDIT_VIDEO = "<Video 1> supplies {supplies}."
|
||||
EDIT_IS = "{main} is the person in <Video 1>, kept as they are: the same face, hair, body and movements, with only what the script names changed."
|
||||
EDIT_VIDEO_KEPT = "<Video 1>: partially_preserved - keep {supplies}; change only what the script names."
|
||||
EDIT_KEPT = "{main} (appears in {shots}): fully_preserved - the face, hair and every movement are retained; only what the script names changes."
|
||||
# What a source video gives an edit, less what a connected picture takes over.
|
||||
EDIT_SUPPLIES = ("the place", "the person", "the action", "the timing", "the camera movement", "the look", "the sound")
|
||||
# A line that is about someone. An edit or a move copy of a clip with nobody in it (an empty hallway) names nobody.
|
||||
NAMES_SOMEONE = re.compile(
|
||||
r"@0|\b(?:he|she|him|her|his|hers|they|them|their|someone|somebody|anyone|everyone|person|people|man|men|woman|women|"
|
||||
r"guy|girl|boy|kid|child|baby|dancer|singer|actor|actress)\b",
|
||||
re.IGNORECASE,
|
||||
)
|
||||
REVOICE_IS = "<Video 1> is the source video for the target video edit.\n{main} is the person speaking in <Video 1>, with the same face, hair and clothing."
|
||||
REVOICE_KEPT = (
|
||||
@@ -229,6 +297,15 @@ REVOICE_KEPT = (
|
||||
)
|
||||
|
||||
|
||||
# A line only names who goes when it is a noun phrase like "the child in the seat", not an action or a pronoun.
|
||||
NOUN_PHRASE = re.compile(r"^(?:the|a|an|this|that|these|those|his|her|their|everyone|someone|somebody|both|each|all)\b", re.I)
|
||||
ACTION_WORD = re.compile(
|
||||
r"\b(?:dances|walks|runs|talks|says|sits|stands|smiles|waves|jumps|enters|leaves|laughs|sings|eats|drinks|waits|"
|
||||
r"is\s+\w+ing|are\s+\w+ing)\b",
|
||||
re.I,
|
||||
)
|
||||
|
||||
|
||||
def who_goes_in(text):
|
||||
"""Who a typed swap line says goes, without the rest of the line: "Elon is @1, he replaces the man at the
|
||||
counter" is the man at the counter, and so is "Replace the man at the counter with @1."."""
|
||||
@@ -244,7 +321,10 @@ def who_goes_in(text):
|
||||
text = re.split(r"[.;!?]", text.strip(" .;!?"), maxsplit=1)[0]
|
||||
text = NAMES_NEXT.sub("", text)
|
||||
text = SAYS_WHO.sub("", TOKEN.sub("", text).strip(" .,;")).strip(" .,;")
|
||||
return re.sub(r"\s+and$", "", text).strip(" .,;")
|
||||
text = re.sub(r"\s+and$", "", text).strip(" .,;")
|
||||
if not (replaces or plain) and not (NOUN_PHRASE.match(text) and not ACTION_WORD.search(text)):
|
||||
return "" # "@1 dances" or "He dances" says what happens, not who goes.
|
||||
return text
|
||||
|
||||
|
||||
def lower_first(text):
|
||||
@@ -274,7 +354,30 @@ class Picture:
|
||||
|
||||
def swapped_in(jobs):
|
||||
"""How many people a swap puts in: one for each person or sheet picture."""
|
||||
return max(1, sum(job in (PERSON, SHEET) for job in jobs.values()))
|
||||
return max(1, sum(job in ACTORS for job in jobs.values()))
|
||||
|
||||
|
||||
# A thing a line holds, wears or uses stays a prop; any other thing named in a line is the one doing the action.
|
||||
HELD = re.compile(
|
||||
r"\b(?:holds?|holding|carr(?:y|ies|ying)|wears?|wearing|puts? on|picks? up|uses?|using|grabs?|throws?|"
|
||||
r"eats?|drinks?|with)\s+(?:\w+\s+){0,2}@(\d)\b",
|
||||
re.IGNORECASE,
|
||||
)
|
||||
|
||||
|
||||
def things_that_perform(script, jobs, video):
|
||||
"""A move copy or a swap with no person picture: a thing picture the lines make do the action (a cartoon hand
|
||||
waving, a toy dancing) takes the moves, or comes in for the one in the video. With nothing typed, a swap's thing
|
||||
comes in. A thing that is held or worn stays a thing."""
|
||||
if video not in (MOVES, SWAP) or any(job in (PERSON, SHEET) for job in (jobs or {}).values()):
|
||||
return jobs
|
||||
held = {int(slot) for slot in HELD.findall(script or "")}
|
||||
named = {int(slot) for slot in TOKEN.findall(script or "")}
|
||||
jobs = dict(jobs)
|
||||
for slot, job in jobs.items():
|
||||
if job == THING and slot not in held and (slot in named or (video == SWAP and not named)):
|
||||
jobs[slot] = PERFORMER
|
||||
return jobs
|
||||
|
||||
|
||||
def number_pictures(jobs, video=""):
|
||||
@@ -295,15 +398,20 @@ class Cast:
|
||||
self.pictures = pictures
|
||||
self.video = video
|
||||
last = max((picture.subject for picture in pictures), default=0)
|
||||
self.people = [picture for picture in pictures if picture.job in (*ACTORS, START)]
|
||||
self.moves = subject_label(last + 1) if video == MOVES else ""
|
||||
# Nobody to give the moves to: the person in the video keeps them and is the main person.
|
||||
self.own_moves = video == MOVES and not self.people
|
||||
self.video_person = True # The video's own person is in the result; notice_people can say there is none.
|
||||
# In a swap each person or sheet picture takes one person's place in the video, in picture order. The script
|
||||
# can name who goes: "the child in the seat", or "@1 the man on the left. @2 the woman in red".
|
||||
self.swappers = [picture for picture in pictures if picture.job in (PERSON, SHEET)] if video == SWAP else []
|
||||
self.swappers = [picture for picture in pictures if picture.job in ACTORS] if video == SWAP else []
|
||||
goes = len(self.swappers)
|
||||
self.whos = [f"the person on the {spot}" for spot in SPOTS[goes]] if goes > 1 else [ONE_GOES]
|
||||
self.replaced = subject_label(1) if video == SWAP else ""
|
||||
self.in_video = subject_label(last + 1) if video in (KEEP, REVOICE, EDIT) else ""
|
||||
self.people = [picture for picture in pictures if picture.job in (PERSON, SHEET, START)]
|
||||
self.in_video = subject_label(last + 1) if video in (KEEP, REVOICE, EDIT) or self.own_moves else ""
|
||||
if video == KEEP and self.people: # A person picture is the main person; the video adds no second subject.
|
||||
self.in_video = ""
|
||||
self.subjects = {picture.slot: subject_label(picture.subject) for picture in pictures if picture.subject}
|
||||
self.clip_clothes = False # A swap that keeps the clothes of whoever goes.
|
||||
self.appearances = {} # Explicit looks-like lines, by the visible replacement picture slot.
|
||||
@@ -311,6 +419,29 @@ class Cast:
|
||||
self.speaks_as_video = False # With that sound and no words typed: the new person says what the video says.
|
||||
self.typed_lines = False # The script typed words in quotes for someone to say.
|
||||
|
||||
def notice_people(self, text):
|
||||
"""An edit or a move copy with no person picture keeps the video's person only when the lines are about
|
||||
someone, or something is put on them. A clip of an empty hallway gets no made-up person."""
|
||||
if self.video == KEEP:
|
||||
if not self.people and (text or "").strip():
|
||||
self.video_person = bool(NAMES_SOMEONE.search(text))
|
||||
elif self.video == EDIT or self.own_moves:
|
||||
self.video_person = bool(self.of(OUTFIT) or NAMES_SOMEONE.search(text or ""))
|
||||
|
||||
@property
|
||||
def supplies(self):
|
||||
"""What the source video of an edit still gives: a place picture brings the place and its look, a style
|
||||
picture the look."""
|
||||
gone = set()
|
||||
if self.of(PLACE):
|
||||
gone |= {"the place", "the look"}
|
||||
if self.of(STYLE):
|
||||
gone.add("the look")
|
||||
if not self.video_person:
|
||||
gone.add("the person")
|
||||
items = [item for item in EDIT_SUPPLIES if item not in gone]
|
||||
return ", ".join(items[:-1]) + " and " + items[-1]
|
||||
|
||||
def of(self, *jobs):
|
||||
return [picture for picture in self.pictures if picture.job in jobs]
|
||||
|
||||
@@ -356,12 +487,32 @@ class Cast:
|
||||
@property
|
||||
def main(self):
|
||||
"""Who a line is about when it names nobody."""
|
||||
if self.in_video:
|
||||
if self.in_video and self.video_person:
|
||||
return self.in_video
|
||||
if self.people:
|
||||
return subject_label(self.people[0].subject)
|
||||
return next(iter(self.subjects.values()), subject_label(1))
|
||||
|
||||
@property
|
||||
def main_in_shot(self):
|
||||
"""Whether main is a person or thing the clip defines, so "<Subject N> is in the shot" can name it."""
|
||||
if self.in_video and self.video_person and self.main == self.in_video:
|
||||
return True
|
||||
defined = {subject_label(picture.subject) for picture in self.of(*ACTORS, THING, START)}
|
||||
return self.main in defined
|
||||
|
||||
def unclear_speaker(self, moments):
|
||||
"""A refusal when two or more people are connected and a spoken line does not say which one speaks."""
|
||||
if self.video not in ("", MOVES, KEEP) or len(self.people) < 2:
|
||||
return ""
|
||||
people = {picture.slot for picture in self.people}
|
||||
for moment in moments:
|
||||
if moment.speech and not moment.voice and not moment.voiceover:
|
||||
if not any(int(slot) in people for slot in TOKEN.findall(moment.text)):
|
||||
count = "two people" if len(self.people) == 2 else "more than one person"
|
||||
return f'With {count}, name who speaks: @{self.people[1].slot} says "hello".'
|
||||
return ""
|
||||
|
||||
@property
|
||||
def opens(self):
|
||||
return next(iter(self.of(START)), None)
|
||||
@@ -378,9 +529,15 @@ class Cast:
|
||||
if not self.pictures and not self.video:
|
||||
return NOTHING_IN
|
||||
if self.video in NEEDS_A_PERSON and not self.people:
|
||||
if self.of(OUTFIT, THING):
|
||||
if self.of(OUTFIT):
|
||||
return DRESS_THE_VIDEO_PERSON
|
||||
if self.of(THING):
|
||||
return SWAP_NEEDS_PERSON_THING
|
||||
if self.of(PLACE):
|
||||
return SWAP_NEEDS_PERSON_PLACE
|
||||
return NEEDS_A_PERSON[self.video]
|
||||
if self.video == REVOICE and self.of(PERSON, SHEET):
|
||||
return REVOICE_KEEPS_PERSON
|
||||
typed = sorted({int(number) for number in TOKEN.findall(text)} - {0})
|
||||
slots = {picture.slot: picture for picture in self.pictures}
|
||||
for slot in typed:
|
||||
@@ -389,16 +546,23 @@ class Cast:
|
||||
if slots[slot].job == END:
|
||||
return f"Picture {slot} is where the clip ends, so @{slot} can't be in a line. Say what happens on the way there."
|
||||
if typed:
|
||||
for picture in self.of(PERSON, SHEET, THING, OUTFIT):
|
||||
if picture.slot not in typed:
|
||||
return f"Picture {picture.slot} is on, but no line names @{picture.slot}."
|
||||
for picture in self.of(*ACTORS, THING, OUTFIT):
|
||||
if picture.slot in typed:
|
||||
continue
|
||||
# The only person picture is who "He" or "She" means.
|
||||
if picture.job in (PERSON, SHEET) and len(self.of(PERSON, SHEET)) == 1 and HE_SHE.search(text):
|
||||
continue
|
||||
hint = f' Type who wears it, like "He wears @{picture.slot}."' if picture.job == OUTFIT else ""
|
||||
return f"Picture {picture.slot} is on, but no line names @{picture.slot}.{hint}"
|
||||
return ""
|
||||
|
||||
def wearers(self, moments):
|
||||
"""Who wears each outfit: the person named just before it, or else the main person."""
|
||||
found = {}
|
||||
outfits = {subject_label(picture.subject) for picture in self.of(OUTFIT)}
|
||||
default = THE_PERSON if self.main in outfits else self.main # An outfit is never worn by itself.
|
||||
for picture in self.of(OUTFIT):
|
||||
found[picture.slot] = self.main
|
||||
found[picture.slot] = default
|
||||
token = f"@{picture.slot}"
|
||||
for moment in moments:
|
||||
before = moment.text.partition(token)[0]
|
||||
@@ -414,7 +578,9 @@ class Cast:
|
||||
ends = self.ends
|
||||
lines = []
|
||||
if self.video == EDIT:
|
||||
lines.append(EDIT_IS.format(main=self.in_video))
|
||||
lines.append(EDIT_VIDEO.format(supplies=self.supplies))
|
||||
if self.video_person:
|
||||
lines.append(EDIT_IS.format(main=self.in_video))
|
||||
if self.video == SWAP:
|
||||
lines.append(SWAP_VIDEO)
|
||||
lines += [SWAP_IS.format(replaced=gone, who=who) for gone, who, _ in self.swaps]
|
||||
@@ -445,10 +611,17 @@ class Cast:
|
||||
lines.append(FIRST_FRAME.format(picture=picture_label(self.opens.picture)))
|
||||
if ends:
|
||||
lines.append(LAST_FRAME.format(picture=picture_label(ends.picture), last=last))
|
||||
if self.moves:
|
||||
if self.own_moves and not self.video_person:
|
||||
lines.append("<Video 1> provides the camera movement.")
|
||||
elif self.own_moves:
|
||||
clothes = "" if self.of(OUTFIT) else " and clothes,"
|
||||
lines.append(OWN_MOVES_ARE.format(main=self.in_video, clothes=clothes or ","))
|
||||
elif self.moves:
|
||||
lines.append(MOVES_ARE.format(moves=self.moves))
|
||||
if self.video == KEEP:
|
||||
lines.append(KEEP_IS.format(main=self.in_video))
|
||||
lines.append(KEEP_IS)
|
||||
if self.in_video and self.video_person:
|
||||
lines.append(KEEP_MAIN.format(main=self.in_video))
|
||||
if self.video == REVOICE:
|
||||
lines.append(REVOICE_IS.format(main=self.in_video))
|
||||
return "\n".join(lines)
|
||||
@@ -458,7 +631,7 @@ class Cast:
|
||||
subject = subject_label(picture.subject)
|
||||
label = picture_label(picture.picture)
|
||||
wearing = [slot for slot, wearer in wearers.items() if wearer == subject]
|
||||
if self.video == SWAP and picture.job in (PERSON, SHEET):
|
||||
if self.video == SWAP and picture.job in ACTORS:
|
||||
if self.clip_clothes:
|
||||
return CLOTHES_STAY.format(who=self.goes_for(picture), picture=label)
|
||||
return SWAPPED.format(picture=label)
|
||||
@@ -467,9 +640,9 @@ class Cast:
|
||||
elif picture.job == PERSON and wearing:
|
||||
words = DRESSED[1].format(outfit=self.subjects[wearing[0]])
|
||||
else:
|
||||
wearer = wearers.get(picture.slot, self.main)
|
||||
wearer = wearers.get(picture.slot, THE_PERSON if picture.job == OUTFIT else self.main)
|
||||
words = JOB_WORDS[picture.job][1].format(wearer=wearer)
|
||||
if picture.job in (PERSON, SHEET, THING):
|
||||
if picture.job in (*ACTORS, THING):
|
||||
words += " " + BACKGROUND.format(picture=label)
|
||||
return words
|
||||
|
||||
@@ -482,7 +655,9 @@ class Cast:
|
||||
|
||||
lines = []
|
||||
if self.video == EDIT:
|
||||
lines.append(EDIT_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video))))
|
||||
lines.append(EDIT_VIDEO_KEPT.format(supplies=self.supplies))
|
||||
if self.video_person:
|
||||
lines.append(EDIT_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video))))
|
||||
if self.video == SWAP:
|
||||
lines.append(SWAP_VIDEO_KEPT.format(people="people" if len(self.swaps) > 1 else "person"))
|
||||
lines += [SWAP_KEPT.format(replaced=gone, who=who, main=new) for gone, who, new in self.swaps]
|
||||
@@ -496,10 +671,25 @@ class Cast:
|
||||
lines.append(FIRST_KEPT.format(picture=picture_label(self.opens.picture)))
|
||||
if self.ends:
|
||||
lines.append(LAST_KEPT.format(picture=picture_label(self.ends.picture), last=len(shots)))
|
||||
if self.moves:
|
||||
if self.own_moves and not self.video_person:
|
||||
places = [self.subjects[picture.slot] for picture in self.of(PLACE)]
|
||||
place = OWN_PLACE_FROM.format(place=places[0]) if places else OWN_PLACE_KEPT
|
||||
lines.append(f"<Video 1>: attribute_transfer - apply only its camera movement to the target scene; {place}")
|
||||
elif self.own_moves:
|
||||
outfits = [self.subjects[picture.slot] for picture in self.of(OUTFIT)]
|
||||
kept = (
|
||||
f"partially_preserved - the face, hair, body, moves and timing are retained, and the clothes change to {outfits[0]}."
|
||||
if outfits else "fully_preserved - the face, hair, body, clothes, moves and timing are retained."
|
||||
)
|
||||
places = [self.subjects[picture.slot] for picture in self.of(PLACE)]
|
||||
place = OWN_PLACE_FROM.format(place=places[0]) if places else OWN_PLACE_KEPT
|
||||
lines.append(OWN_MOVES_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video)), kept=kept, place=place))
|
||||
elif self.moves:
|
||||
lines.append(MOVES_KEPT.format(moves=self.moves, shots=shot_list(seen(self.main)), main=self.main))
|
||||
if self.video == KEEP:
|
||||
lines.append(KEEP_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video))))
|
||||
lines.append(KEEP_KEPT_PLACE if self.of(PLACE) else KEEP_KEPT)
|
||||
if self.in_video and self.video_person:
|
||||
lines.append(KEEP_MAIN_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video))))
|
||||
if self.video == REVOICE:
|
||||
lines.append(REVOICE_KEPT.format(main=self.in_video, shots=shot_list(seen(self.in_video))))
|
||||
return "\n".join(lines)
|
||||
|
||||
@@ -13,7 +13,7 @@ import re
|
||||
import numpy as np
|
||||
|
||||
from ._h3_cast_helpers import EDIT, KEEP, NEEDS_NO_WORDS, ONE_GOES, REVOICE, SWAP, Cast, number_pictures, swap_appearances
|
||||
from ._h3_cast_helpers import TOKEN
|
||||
from ._h3_cast_helpers import TOKEN, things_that_perform
|
||||
from ._h3_script_helpers import (
|
||||
ACTION_GOES_ON,
|
||||
PRONOUN,
|
||||
@@ -78,9 +78,10 @@ def one_line(text):
|
||||
|
||||
def clean_words(text):
|
||||
"""Spoken words: one line, no tags, ending the way H3 wants a line to end."""
|
||||
text = one_line(text).replace("<", "").replace(">", "")
|
||||
if text and text[-1] not in ".?!":
|
||||
text += "."
|
||||
# '"Hi," she says' is "Hi.", not "Hi,.". Chinese and Japanese lines end in their own stops: "。", not "。.".
|
||||
text = one_line(text).replace("<", "").replace(">", "").rstrip(" ,;:,、;:")
|
||||
if text and text[-1] not in ".?!。?!…":
|
||||
text += "。" if re.search(r"[\u3040-\u30ff\u4e00-\u9fff]", text) else "."
|
||||
return text
|
||||
|
||||
|
||||
@@ -312,6 +313,7 @@ def write_prompt(script, seconds=15.0, camera="", look="", sound="", language="E
|
||||
longest) puts a question to a small model, if one is connected: it fills in
|
||||
a one-line idea and the sound of the place."""
|
||||
parts = split_sections(script)
|
||||
written = parts is None # Words someone typed in six parts go through as they are.
|
||||
audio = next_label(script, "Audio")
|
||||
speaker = None
|
||||
|
||||
@@ -319,10 +321,13 @@ def write_prompt(script, seconds=15.0, camera="", look="", sound="", language="E
|
||||
jobs = jobs if jobs is not None else {1: "person"}
|
||||
if video == EDIT:
|
||||
script, jobs = put_on_from_pictures(script, jobs)
|
||||
jobs = things_that_perform(script, jobs, video)
|
||||
cast = Cast(number_pictures(jobs, video), video)
|
||||
script, cast.appearances = swap_appearances(script, jobs, video)
|
||||
scene = read_script(script)
|
||||
cast.notice_people(" ".join(moment.text for moment in scene.moments))
|
||||
problem = cast.problem(" ".join(moment.text for moment in scene.moments))
|
||||
problem = problem or cast.unclear_speaker(scene.moments)
|
||||
if problem:
|
||||
raise ValueError(problem)
|
||||
if not scene.moments and video == EDIT:
|
||||
@@ -370,6 +375,8 @@ def write_prompt(script, seconds=15.0, camera="", look="", sound="", language="E
|
||||
scene, seconds, camera=camera, look=look, sound=sound, say=say, cast=cast, sounds=not keeps_song, soundtrack=soundtrack
|
||||
)
|
||||
|
||||
if written and (video == EDIT or (video == SWAP and clip_sound)): # The clip's own sound stays in the result.
|
||||
parts["summary"] = add_task(parts["summary"], "audio reuse")
|
||||
if voice:
|
||||
add_voice(parts, voice_label, speaker)
|
||||
if song_beats is not None and voice:
|
||||
|
||||
+69
-21
@@ -17,15 +17,19 @@ from ._h3_cast_helpers import (
|
||||
MAIN,
|
||||
OUTFIT,
|
||||
PERSON,
|
||||
PERFORMER,
|
||||
PLACE,
|
||||
REVOICE,
|
||||
SHEET,
|
||||
START as START_JOB,
|
||||
END as END_JOB,
|
||||
STYLE,
|
||||
SWAP,
|
||||
SOUND_GOES_ON,
|
||||
SOUND_GOES_ON_KEPT,
|
||||
THING,
|
||||
TOKEN,
|
||||
things_that_perform,
|
||||
ONE_GOES,
|
||||
SPOTS,
|
||||
Cast,
|
||||
@@ -43,7 +47,9 @@ THEN = r"(?:,?\s*(?:and\s+)?then\b|,?\s*and\s+after\s+that\b|,?\s*after\s+that\b
|
||||
# A word that starts a new moment and stays in it.
|
||||
TURN = r"(?=,?\s*\b(?:suddenly|all\s+of\s+a\s+sudden|out\s+of\s+nowhere|finally|at\s+last|(?:hard\s+)?cut\s+to|(?:it\s+|and\s+|(?<!\bit\s))ends?\s+(?:with|on|in|as)|ending\s+(?:with|on|in))\b)"
|
||||
# Where a line breaks into sentences: a full stop that is not a title, the end of a quote, or a semicolon.
|
||||
SENTENCE = r"(?<!\bMr\.)(?<!\bMs\.)(?<!\bDr\.)(?<!\bSt\.)(?<!\bMrs\.)(?<=[.!?])\s+(?=\S)|(?<=\x00)\s+(?=[A-Z])|\s*;\s*"
|
||||
# After a quote, a new sentence starts with a capital or an @: in '"Get a job." @1 looks up' the @1 is a new moment,
|
||||
# or its own words are lost behind the first quote's. So is '"Hi" and @2 says "Bye"': one moment holds one quote.
|
||||
SENTENCE = r"(?<!\bMr\.)(?<!\bMs\.)(?<!\bDr\.)(?<!\bSt\.)(?<!\bMrs\.)(?<=[.!?])\s+(?=\S)|(?<=\x00)\s+(?=[A-Z@])|(?<=\x00),?\s+(?:and|but|while)\s+(?=@\d)|\s*;\s*"
|
||||
CUT_WORDS = re.compile(r"^(?:hard\s+)?(?:cut|cuts|smash\s+cut|jump\s+cut)\s+to\b[:,]?\s*", re.I)
|
||||
NEW_SHOT = re.compile(r"^(?:new\s+shot|next\s+shot|shot\s+\d+)\b[:,.\-]?\s*", re.I)
|
||||
START = re.compile(r"^(?:it\s+|the\s+(?:clip|video|scene|shot)\s+)?(?:starts?|begins?|opens?)(?:\s+(?:with|on|in|at|as)\b[:,]?|\s*:)\s*", re.I)
|
||||
@@ -110,6 +116,30 @@ SOUND_WORDS = 22
|
||||
NAMED = {PLACE: "the place", OUTFIT: "the outfit", THING: "the thing"} # What @2 reads as to the small model.
|
||||
|
||||
WORDS_A_SECOND = 3 # How fast H3 speaks.
|
||||
# Words that trail off ("I never thought...") keep going: across the next cut, or past the end of the clip.
|
||||
TRAILS_OFF = re.compile(r"(?:\.\.\.|\u2026|\u2014|--)\s*$")
|
||||
LINE_CARRIES = "<scenetrans> The line from the previous shot continues seamlessly across the cut, and the words are not repeated."
|
||||
CUT_OFF = "The line is cut off by the end of the video. <cutoff>"
|
||||
|
||||
|
||||
def speech_runs_on(script, seconds):
|
||||
"""(carried, cut_off): the moments whose words go on across the next cut, and the last moment's words if the
|
||||
clip ends mid-line. Words carry on when they trail off with "..." or need more time than their shot has."""
|
||||
starts = timeline(script, seconds)
|
||||
ends = [*starts[1:], seconds]
|
||||
carried, cut_off = set(), None
|
||||
for index, moment in enumerate(script.moments):
|
||||
if not moment.speech:
|
||||
continue
|
||||
trails = bool(TRAILS_OFF.search(moment.speech))
|
||||
long = len(moment.speech.split()) / WORDS_A_SECOND > ends[index] - starts[index] + 0.5
|
||||
following = script.moments[index + 1] if index + 1 < len(script.moments) else None
|
||||
if following is not None and following.cut and not script.one_take and (trails or long):
|
||||
carried.add(index)
|
||||
elif following is None and trails:
|
||||
cut_off = index
|
||||
return carried, cut_off
|
||||
|
||||
HOLD = f"{MAIN} holds still" # What fills the time before a first moment set later.
|
||||
KEEP_LOOK = "The target video keeps the look, the light and the colours of {picture}."
|
||||
KEEP_COLOURS = "The target video keeps the colours and the film grain of {picture}."
|
||||
@@ -130,6 +160,7 @@ OPENS_ON = "The shot begins from {picture}."
|
||||
ENDS_ON = "The shot ends on {picture}."
|
||||
ON_SHEET = "{subject} appears once, without the panels, labels or other views of {picture}."
|
||||
THE_MOVES = "{main} performs the same moves as {moves}, at the same timing."
|
||||
OWN_MOVES = "{main} moves exactly as in <Video 1>, at the same timing."
|
||||
IN_THE_PLACE = "The shot takes place in {place}."
|
||||
WEARS = "{wearer} wears {outfit}."
|
||||
IN_THE_SHOT = "{main} is in the shot."
|
||||
@@ -142,10 +173,10 @@ SAYS_NOTHING = "{main} keeps the lips closed and says nothing the whole time."
|
||||
SPEAKS_AS_VIDEO = "{main} speaks as in <Video 1>, each word landing exactly on the lip movement of <Video 1>."
|
||||
CLIP_SOUND = "The original sound of <Video 1> continues throughout."
|
||||
FOLLOWS_CAMERA = "The camera follows the movement of <Video 1>."
|
||||
REST_AS_VIDEO = "Everything else stays as it is in <Video 1>."
|
||||
SAME_AS_VIDEO = "{main} has the same words, mouth movements and moves as in <Video 1>. Everything else stays as it is in <Video 1>."
|
||||
CARRIES_ON = "The shot carries straight on from the end of <Video 1>."
|
||||
ACTION_GOES_ON = "The action keeps going the way it was heading, and the camera keeps moving the same way."
|
||||
FACES_US = "{main} faces the camera as in <Video 1>."
|
||||
VIDEO_LOOKS = {
|
||||
SWAP: "Retain the look of <Video 1>.",
|
||||
KEEP: "The target video keeps the look, colours and camera of <Video 1>.",
|
||||
@@ -557,12 +588,13 @@ def script_warning(script, seconds, cast=None):
|
||||
if any((moment.at or 0.0) >= seconds for moment in script.moments):
|
||||
return f"A moment starts after the clip ends, at {seconds:g} seconds."
|
||||
starts = timeline(script, seconds)
|
||||
for moment, start, end in zip(script.moments, starts, [*starts[1:], seconds]):
|
||||
carried, cut_off = speech_runs_on(script, seconds)
|
||||
for index, (moment, start, end) in enumerate(zip(script.moments, starts, [*starts[1:], seconds])):
|
||||
unseen = UNSEEN.search(without_speech(moment))
|
||||
if unseen:
|
||||
return f'A camera can\'t see "{unseen.group(0).lower()}". Say what we would see, like "she smiles".'
|
||||
takes = len(moment.speech.split()) / WORDS_A_SECOND
|
||||
if takes > end - start + 0.5:
|
||||
if takes > end - start + 0.5 and index not in carried and index != cut_off:
|
||||
return f"A spoken line takes about {takes:.0f} seconds to say, and its moment has {end - start:.0f}. Give it more room."
|
||||
return ""
|
||||
|
||||
@@ -585,6 +617,7 @@ def script_report(text, seconds, jobs=None, video=""):
|
||||
}
|
||||
for moment, start, end in zip(script.moments, starts, ends)
|
||||
]
|
||||
jobs = things_that_perform(text, jobs, video) if jobs is not None else None
|
||||
cast = Cast(number_pictures(jobs, video), video) if jobs is not None else None
|
||||
return {"moments": moments, "seconds": seconds, "warning": script_warning(script, seconds, cast)}
|
||||
|
||||
@@ -732,6 +765,8 @@ def before_the_action(number, text, cast, wearers):
|
||||
# or for the whole clip at Fast preview; with it the new
|
||||
# person is there from the first frame. An @1 looks like line only adds its own words.
|
||||
looks = (f"has the defined appearance from {picture_label(picture.picture)}: {appearance}" if appearance
|
||||
else f"has the shape, colours and drawing style of the character in {picture_label(picture.picture)}"
|
||||
if picture.job == PERFORMER
|
||||
else f"has the face, hair and clothes of the person in {picture_label(picture.picture)}")
|
||||
notes.append(
|
||||
f"{new} occupies the starting position of {who} and follows the same route, movements, timing and pace. "
|
||||
@@ -749,8 +784,6 @@ def before_the_action(number, text, cast, wearers):
|
||||
notes.append((OWN_LINES_ONLY if f"{new} (S" in text else SAYS_NOTHING).format(main=new))
|
||||
if number == 1 and cast.video == KEEP:
|
||||
notes += [CARRIES_ON]
|
||||
if number == 1 and cast.video == REVOICE:
|
||||
notes += [FACES_US.format(main=cast.main)]
|
||||
places = [subject_label(picture.subject) for picture in cast.of(PLACE)]
|
||||
if places and not any(place in text for place in places):
|
||||
notes.append(IN_THE_PLACE.format(place=places[0]))
|
||||
@@ -760,7 +793,8 @@ def before_the_action(number, text, cast, wearers):
|
||||
notes.append(sentence(", ".join(spots[:-1]) + f" and {spots[-1]} of the frame"))
|
||||
for slot, wearer in wearers.items():
|
||||
outfit = cast.subjects[slot]
|
||||
if outfit not in text and (wearer in people or wearer in text):
|
||||
# An edit has no person picture: the person in the video wears an outfit that no line names.
|
||||
if outfit not in text and (wearer in people or wearer in text or (cast.video == EDIT and wearer == cast.in_video)):
|
||||
notes.append(WEARS.format(wearer=wearer, outfit=outfit))
|
||||
return notes
|
||||
|
||||
@@ -768,10 +802,10 @@ def before_the_action(number, text, cast, wearers):
|
||||
def after_the_action(text, cast):
|
||||
"""What H3 needs to hear about a shot once its action has been said."""
|
||||
notes = []
|
||||
if cast.moves and cast.main in text:
|
||||
notes.append(THE_MOVES.format(main=cast.main, moves=cast.moves))
|
||||
if cast.moves and cast.video_person and cast.main in text:
|
||||
notes.append(OWN_MOVES.format(main=cast.main) if cast.own_moves else THE_MOVES.format(main=cast.main, moves=cast.moves))
|
||||
if cast.video == EDIT:
|
||||
notes.append(SAME_AS_VIDEO.format(main=cast.main))
|
||||
notes.append(SAME_AS_VIDEO.format(main=cast.main) if cast.video_person else REST_AS_VIDEO)
|
||||
if cast.video in (SWAP, EDIT):
|
||||
notes.append(FOLLOWS_CAMERA)
|
||||
for picture in cast.of(SHEET):
|
||||
@@ -799,6 +833,7 @@ def shot_lines(script, seconds, say, cast, wearers, move="", hold="", sounds=Tru
|
||||
talking = speakers(script.moments, cast)
|
||||
quiet = " ".join(QUIET.format(speaker=speaker) for speaker in talking if speaker.startswith("<"))
|
||||
kept = still_true(script.moments)
|
||||
carried, cut_off = speech_runs_on(script, seconds)
|
||||
lines = []
|
||||
for number, shot in enumerate(shots_of(script), 1):
|
||||
shot_camera = script.moments[shot[0]].camera
|
||||
@@ -841,8 +876,17 @@ def shot_lines(script, seconds, say, cast, wearers, move="", hold="", sounds=Tru
|
||||
*([hold] if number > 1 else []),
|
||||
capital(script.moments[first].sound) if script.moments[first].sound and not script.silent else "",
|
||||
]
|
||||
lines.append(f"[Shot {number}] " + " ".join(part for part in body if part))
|
||||
if not any(cast.main in line for line in lines):
|
||||
line = " ".join(part for part in body if part)
|
||||
if shot[0] - 1 in carried:
|
||||
# The speaker is mid-line as the shot opens, so the mouth is not closed here.
|
||||
line = f"{LINE_CARRIES} " + re.sub(r" ?<Subject \d+> stays silent with the lips closed\.", "", line, count=1)
|
||||
if shot[-1] in carried:
|
||||
# The speaker is still talking at the cut, so the mouth does not close here.
|
||||
line = re.sub(r" <Subject \d+> closes the lips\.", "", line) + " <scenetrans>"
|
||||
if shot[-1] == cut_off:
|
||||
line = re.sub(r" <Subject \d+> closes the lips\.", "", line) + f" {CUT_OFF}"
|
||||
lines.append(f"[Shot {number}] " + line)
|
||||
if cast.video_person and cast.main_in_shot and not any(cast.main in line for line in lines):
|
||||
lines[0] = lines[0].replace("[Shot 1] ", f"[Shot 1] {IN_THE_SHOT.format(main=cast.main)} ", 1)
|
||||
return lines
|
||||
|
||||
@@ -881,16 +925,19 @@ def summary_of(script, cast, new_places, soundtrack):
|
||||
return f"[{task}] {changes} Everything else in <Video 1> is kept."
|
||||
if cast.video == KEEP:
|
||||
tag = f"{task} + audio reference" if soundtrack else task
|
||||
return f"[{tag}] The target video continues directly from the end of <Video 1> in the same place."
|
||||
opening = with_labels(without_speech(script.moments[0]), cast) or f"{cast.main} speaks"
|
||||
where = "" if cast.of(PLACE) else " in the same place"
|
||||
return f"[{tag}] The target video continues directly from the end of <Video 1>{where}."
|
||||
talking = [who for who in speakers(script.moments, cast) if who.startswith("<")]
|
||||
who_speaks = f"{' and '.join(talking)} speak" if len(talking) > 1 else f"{cast.main} speaks"
|
||||
opening = with_labels(without_speech(script.moments[0]), cast) or who_speaks
|
||||
later = [with_labels(without_speech(script.moments[shot[0]]), cast) for shot in shots_of(script)[1:]]
|
||||
if cast.opens and cast.ends:
|
||||
moving = f"moving from {picture_label(cast.opens.picture)} to {picture_label(cast.ends.picture)}"
|
||||
return f"[{task}] In the target video, {lower_first(opening.rstrip('.!? '))}, {moving}."
|
||||
summary = f"[{task}] In the target video, {lower_first(sentence(opening))}"
|
||||
summary += "".join(f" {AND_THEN.format(line=lower_first(sentence(line)))}" for line in later if line)
|
||||
if cast.moves:
|
||||
summary += " " + THE_MOVES.format(main=cast.main, moves=cast.moves)
|
||||
if cast.moves and cast.video_person:
|
||||
summary += " " + (OWN_MOVES.format(main=cast.main) if cast.own_moves else THE_MOVES.format(main=cast.main, moves=cast.moves))
|
||||
return f"{summary} {EVERY_PLACE}" if new_places else summary
|
||||
|
||||
|
||||
@@ -916,14 +963,15 @@ def script_parts(script, seconds, camera="", look="", sound="", say=None, cast=N
|
||||
cuts = [script.moments[shot[0]] for shot in shots_of(script)[1:]]
|
||||
moved = any(not FRAME_WORDS.match(without_speech(moment)) for moment in cuts)
|
||||
new_places = last > 1 and not cast.of(PLACE) and moved and all(not FRAME_WORDS.match(without_speech(moment)) for moment in cuts)
|
||||
first = next(iter(cast.of(PLACE) or cast.pictures), None)
|
||||
if not style and cast.video == EDIT:
|
||||
# Only a place, a start frame or a person sets the look; an outfit, a thing or a style never does.
|
||||
first = next(iter(cast.of(PLACE) or cast.of(START_JOB) or cast.of(PERSON, SHEET, PERFORMER, END_JOB)), None)
|
||||
if not style and cast.video == EDIT and not cast.of(PLACE):
|
||||
style = VIDEO_LOOKS[EDIT]
|
||||
if not style and len(cast.of(PLACE)) > 1:
|
||||
style = OWN_COLOURS
|
||||
elif not style and not (first and first.job == THING):
|
||||
elif not style and (first or cast.video in VIDEO_LOOKS):
|
||||
words = KEEP_LOOK if cast.opens else KEEP_COLOURS
|
||||
style = VIDEO_LOOKS.get(cast.video) or words.format(picture=picture_label(first.picture if first else 1))
|
||||
style = (cast.video != EDIT and VIDEO_LOOKS.get(cast.video)) or words.format(picture=picture_label(first.picture if first else 1))
|
||||
takes = [EVERY_PLACE] if new_places else []
|
||||
if last == 1 and (len(script.moments) > 1 or cast.moves or script.one_take):
|
||||
takes = [ONE_SHOT]
|
||||
@@ -947,7 +995,7 @@ def script_parts(script, seconds, camera="", look="", sound="", say=None, cast=N
|
||||
"subject_definitions": subjects,
|
||||
"summary": summary_of(script, cast, new_places, soundtrack),
|
||||
"retention_analysis": kept,
|
||||
"detailed_description": "\n".join([" ".join(part for part in (style, move, *takes) if part), *lines]),
|
||||
"detailed_description": "\n".join([*[" ".join(part for part in (style, move, *takes) if part)] * bool(style or move or takes), *lines]),
|
||||
"overall_soundscape": ambience,
|
||||
"non_diegetic_music": tune,
|
||||
}
|
||||
|
||||
@@ -40,6 +40,9 @@ OUTSIDE = (
|
||||
"Image Folder reads only folders inside ComfyUI's input or output folder. "
|
||||
"Drop your folder on the node, or press Add, and its pictures are copied in."
|
||||
)
|
||||
# A drive ("D:") in any part of a folder name, and control characters.
|
||||
_DRIVE = re.compile(r"^[A-Za-z]:")
|
||||
_CONTROL = re.compile(r"[\x00-\x1f\x7f]")
|
||||
NO_FOLDER = 'Image Folder: there is no folder "{folder}" in ComfyUI\'s {source} folder.'
|
||||
NO_PICTURES = 'Image Folder: no pictures in "{folder}". Add some, or choose another folder.'
|
||||
NONE_LEFT = "Image Folder: none of the picked pictures are in the folder anymore. Pick again on the node."
|
||||
@@ -64,7 +67,9 @@ def clean_folder(folder) -> str:
|
||||
if text.startswith("/") or text.startswith("~") or re.match(r"^[A-Za-z]:", text):
|
||||
raise ValueError(OUTSIDE)
|
||||
parts = [part for part in text.split("/") if part not in ("", ".")]
|
||||
if any(part == ".." for part in parts):
|
||||
# ".." climbs out; a drive in any part ("a/D:/x") jumps to another drive
|
||||
# on Windows; a control character is never part of a folder name.
|
||||
if any(part == ".." or _DRIVE.match(part) for part in parts) or _CONTROL.search(text):
|
||||
raise ValueError(OUTSIDE)
|
||||
return "/".join(parts)
|
||||
|
||||
@@ -82,9 +87,10 @@ def place_inside(folder, roots: dict) -> tuple[str, str] | None:
|
||||
This compares text only. A path that is outside those folders is never
|
||||
looked up on disk, and one on another machine is refused from its text."""
|
||||
text = str(folder or "").strip().strip('"').replace("\\", "/")
|
||||
if not looks_full(text) or text.startswith("//"):
|
||||
# A "~" path is not expanded: only a path written out in full can match.
|
||||
if not looks_full(text) or text.startswith("//") or text.startswith("~"):
|
||||
return None
|
||||
full = os.path.normpath(os.path.expanduser(text))
|
||||
full = os.path.normpath(text)
|
||||
for source, root in roots.items():
|
||||
for base in {os.path.normpath(str(root)), os.path.realpath(str(root))}:
|
||||
if full == base:
|
||||
@@ -94,14 +100,27 @@ def place_inside(folder, roots: dict) -> tuple[str, str] | None:
|
||||
return None
|
||||
|
||||
|
||||
def inside_root(top: Path, relative: str) -> Path:
|
||||
"""`relative` below the resolved root `top`, as a resolved path. It is
|
||||
checked as text before the disk is asked to resolve it, and again after,
|
||||
so a link that leads out of the root is refused too."""
|
||||
if not relative:
|
||||
return top
|
||||
lexical = top.joinpath(*relative.split("/"))
|
||||
if top not in lexical.parents:
|
||||
raise ValueError(OUTSIDE)
|
||||
target = lexical.resolve()
|
||||
if target != top and top not in target.parents:
|
||||
raise ValueError(OUTSIDE)
|
||||
return target
|
||||
|
||||
|
||||
def safe_folder(root, folder, source: str = "input") -> Path:
|
||||
"""The folder as a real path below `root`. A link that leads out of the
|
||||
root is refused like any other way out."""
|
||||
relative = clean_folder(folder)
|
||||
top = Path(root).resolve()
|
||||
target = top.joinpath(*relative.split("/")).resolve() if relative else top
|
||||
if target != top and top not in target.parents:
|
||||
raise ValueError(OUTSIDE)
|
||||
target = inside_root(top, relative)
|
||||
if not target.is_dir():
|
||||
raise ValueError(NO_FOLDER.format(folder=relative, source=source))
|
||||
return target
|
||||
|
||||
@@ -0,0 +1,203 @@
|
||||
"""Pictures Save Image holds until they are kept.
|
||||
|
||||
With Auto save off, a run writes its pictures to ComfyUI's temp folder and
|
||||
shows them on the node. A click on the node's Save button keeps the ones
|
||||
that are wanted: they are copied into the output folder under the names the
|
||||
node would have given them. Everything here works from names this module
|
||||
made itself; the only thing a caller sends is the run's token and which
|
||||
pictures to keep.
|
||||
|
||||
What a run holds (its settings, its count, what was kept) stays in this
|
||||
process's memory, never in a file. ComfyUI's temp folder takes uploads, so a
|
||||
list read back from there could be anyone's. The held pictures themselves
|
||||
are named from the token, their number and one of the node's own formats,
|
||||
and must still be in the held folder when they are copied.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from collections import OrderedDict
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
import re
|
||||
import secrets
|
||||
import shutil
|
||||
import threading
|
||||
|
||||
from ._image_save_helpers import (
|
||||
FORMAT_EXTENSIONS,
|
||||
IMAGE_SAVE_LOCK,
|
||||
existing_action,
|
||||
name_stem,
|
||||
next_free_counter,
|
||||
plan_exact_names,
|
||||
plan_local_names,
|
||||
require_output_path,
|
||||
resolve_output_root,
|
||||
sidecar_path,
|
||||
split_prefix,
|
||||
strip_image_extension,
|
||||
)
|
||||
|
||||
HELD_FOLDER = "ausboss_held"
|
||||
_TOKEN = re.compile(r"^[0-9a-f]{16}$")
|
||||
GONE = "These pictures are no longer held. Run the workflow again."
|
||||
# The held pictures' formats: the extensions Save Image writes, nothing else.
|
||||
HELD_EXTENSIONS = frozenset(FORMAT_EXTENSIONS.values())
|
||||
# A run holds at most this many pictures, and the server remembers this many
|
||||
# runs; the oldest is forgotten first and then says GONE.
|
||||
MAX_HELD_PICTURES = 10000
|
||||
MAX_HELD_RUNS = 256
|
||||
_RUNS: "OrderedDict[str, dict]" = OrderedDict()
|
||||
_RUNS_LOCK = threading.Lock()
|
||||
|
||||
|
||||
def plan_destinations(settings: dict, count: int, now: datetime, root: Path, output_root: Path, picks=None):
|
||||
"""Where each picture goes, as (path, what to do if it exists).
|
||||
|
||||
Call it with IMAGE_SAVE_LOCK held: the counter it reads stays free only
|
||||
until the lock is let go. `picks` keeps some of a run's pictures; an
|
||||
exact name then keeps the number it would have had in the whole run.
|
||||
"""
|
||||
wanted = list(range(count)) if picks is None else list(picks)
|
||||
extension = settings["extension"]
|
||||
exact = settings.get("exact") or ""
|
||||
if exact:
|
||||
names = plan_exact_names(strip_image_extension(exact), extension, count)
|
||||
return [(root / names[index], settings["on_existing"]) for index in wanted]
|
||||
subfolder, base = split_prefix(settings["filename_prefix"])
|
||||
folder = root / subfolder if subfolder else root
|
||||
require_output_path(folder, output_root)
|
||||
stem = name_stem(
|
||||
base,
|
||||
{"date": bool(settings["name_date"]), "time": bool(settings["name_time"]), "size": bool(settings["name_size"])},
|
||||
now,
|
||||
int(settings["width"]),
|
||||
int(settings["height"]),
|
||||
)
|
||||
counter = bool(settings["name_counter"])
|
||||
existing = [entry.name for entry in folder.iterdir()] if folder.is_dir() else []
|
||||
names = plan_local_names(
|
||||
stem, extension, len(wanted), counter=counter, batch=bool(settings["name_batch"]),
|
||||
next_counter=next_free_counter(existing, stem),
|
||||
)
|
||||
# With a counter the name is free by construction. Without one, the
|
||||
# legacy policy decides whether to overwrite.
|
||||
policy = "overwrite" if counter else settings["on_existing"]
|
||||
return [(folder / name, policy) for name in names]
|
||||
|
||||
|
||||
def settle_destinations(planned, has_caption: bool, output_root: Path) -> list[str]:
|
||||
"""Check every destination, then say 'write' or 'skip' for each.
|
||||
|
||||
The whole batch is settled before any write: an "error" on a later
|
||||
picture stops the save before the first one lands.
|
||||
"""
|
||||
for path, _policy in planned:
|
||||
require_output_path(path, output_root)
|
||||
if has_caption:
|
||||
require_output_path(sidecar_path(path), output_root)
|
||||
return [
|
||||
existing_action(path.exists() or (has_caption and sidecar_path(path).exists()), policy)
|
||||
for path, policy in planned
|
||||
]
|
||||
|
||||
|
||||
def _held_name(token: str, index: int, extension: str) -> str:
|
||||
if not _TOKEN.match(token) or extension not in HELD_EXTENSIONS or not 0 <= index < MAX_HELD_PICTURES:
|
||||
raise ValueError("Save Image: that is not a picture it is holding.")
|
||||
return f"{token}_{index:03d}.{extension}"
|
||||
|
||||
|
||||
def _held_source(folder: Path, token: str, index: int, extension: str) -> Path:
|
||||
"""A held picture's file, which must still sit directly in the held folder."""
|
||||
source = folder / _held_name(token, index, extension)
|
||||
if not source.is_file() or source.is_symlink() or source.resolve().parent != folder.resolve():
|
||||
raise LookupError(GONE)
|
||||
return source
|
||||
|
||||
|
||||
def hold_pictures(temp_root, frames, settings: dict, encode) -> tuple[str, list[dict]]:
|
||||
"""Write a run's pictures to the temp folder and remember how to name them.
|
||||
|
||||
`encode(path, frame)` writes one picture in the run's format. Returns
|
||||
the run's token and the entries the node shows them from.
|
||||
"""
|
||||
frames = list(frames)
|
||||
if len(frames) > MAX_HELD_PICTURES:
|
||||
raise ValueError(
|
||||
f"Save Image: {len(frames)} pictures are too many to hold (at most {MAX_HELD_PICTURES}). "
|
||||
"Turn Auto save on to save them straight away."
|
||||
)
|
||||
folder = Path(temp_root) / HELD_FOLDER
|
||||
folder.mkdir(parents=True, exist_ok=True)
|
||||
token = secrets.token_hex(8)
|
||||
previews = []
|
||||
for index, frame in enumerate(frames):
|
||||
name = _held_name(token, index, settings["extension"])
|
||||
encode(folder / name, frame)
|
||||
previews.append({"filename": name, "subfolder": HELD_FOLDER, "type": "temp"})
|
||||
with _RUNS_LOCK:
|
||||
_RUNS[token] = dict(settings, count=len(previews), kept={})
|
||||
while len(_RUNS) > MAX_HELD_RUNS:
|
||||
_RUNS.popitem(last=False)
|
||||
return token, previews
|
||||
|
||||
|
||||
def keep_pictures(temp_root, output_root, token, picks=None, now: datetime | None = None) -> dict:
|
||||
"""Copy held pictures into the output folder. Returns what was saved.
|
||||
|
||||
`picks` lists which pictures of the run to keep, counted from 0; None
|
||||
keeps them all. A picture kept before is not saved a second time.
|
||||
"""
|
||||
if not isinstance(token, str) or not _TOKEN.match(token):
|
||||
raise ValueError("Save Image: that is not a run it is holding.")
|
||||
folder = Path(temp_root) / HELD_FOLDER
|
||||
with _RUNS_LOCK:
|
||||
manifest = _RUNS.get(token)
|
||||
if manifest is None:
|
||||
raise LookupError(GONE)
|
||||
count = int(manifest["count"])
|
||||
kept = dict(manifest["kept"])
|
||||
|
||||
if picks is None:
|
||||
wanted = list(range(count))
|
||||
else:
|
||||
if len(picks) > MAX_HELD_PICTURES:
|
||||
raise ValueError("Save Image: that picture is not in this run.")
|
||||
wanted = []
|
||||
for pick in picks:
|
||||
if isinstance(pick, bool) or not isinstance(pick, int) or not 0 <= pick < count:
|
||||
raise ValueError("Save Image: that picture is not in this run.")
|
||||
if pick not in wanted:
|
||||
wanted.append(pick)
|
||||
todo = [index for index in wanted if index not in kept]
|
||||
|
||||
output_root = Path(output_root)
|
||||
caption = str(manifest.get("caption") or "")
|
||||
has_caption = bool(caption.strip())
|
||||
skipped: list[int] = []
|
||||
saved: list[int] = []
|
||||
with IMAGE_SAVE_LOCK:
|
||||
root = resolve_output_root(manifest.get("output_dir", ""), output_root)
|
||||
planned = plan_destinations(manifest, count, now or datetime.now(), root, output_root, picks=todo)
|
||||
actions = settle_destinations(planned, has_caption, output_root)
|
||||
resolved_root = output_root.resolve()
|
||||
for (path, _policy), action, index in zip(planned, actions, todo):
|
||||
if action == "skip":
|
||||
skipped.append(index)
|
||||
continue
|
||||
source = _held_source(folder, token, index, manifest["extension"])
|
||||
path.parent.mkdir(parents=True, exist_ok=True)
|
||||
shutil.copyfile(source, path)
|
||||
if has_caption:
|
||||
sidecar_path(path).write_text(caption, encoding="utf-8")
|
||||
kept[index] = path.resolve().relative_to(resolved_root).as_posix()
|
||||
saved.append(index)
|
||||
manifest["kept"] = kept
|
||||
return {
|
||||
"count": count,
|
||||
"saved": [{"index": index, "path": kept[index]} for index in saved],
|
||||
"kept": sorted(kept),
|
||||
"skipped": skipped,
|
||||
}
|
||||
@@ -10,6 +10,7 @@ from __future__ import annotations
|
||||
|
||||
import json
|
||||
import re
|
||||
import secrets
|
||||
import threading
|
||||
from datetime import datetime
|
||||
from pathlib import Path, PurePosixPath, PureWindowsPath
|
||||
@@ -271,6 +272,36 @@ def frame_to_pil(frame: torch.Tensor) -> Image.Image:
|
||||
return Image.fromarray(array)
|
||||
|
||||
|
||||
# A small copy of each picture, for the node's face. The saved file can be
|
||||
# tens of megabytes; this one is a few hundred kilobytes, so the picture
|
||||
# shows at once on a slow connection too. It lives in the temp folder, which
|
||||
# ComfyUI empties when it starts.
|
||||
PREVIEW_FOLDER = "ausboss_preview"
|
||||
PREVIEW_EDGE = 2048
|
||||
|
||||
|
||||
def write_light_previews(temp_root, frames) -> list[str]:
|
||||
"""Write one small WebP per frame; returns each place below the temp folder.
|
||||
|
||||
Never in the way of a save: if anything goes wrong the list is empty
|
||||
and the node shows the saved file itself, as it always did.
|
||||
"""
|
||||
try:
|
||||
folder = Path(temp_root) / PREVIEW_FOLDER
|
||||
folder.mkdir(parents=True, exist_ok=True)
|
||||
token = secrets.token_hex(8)
|
||||
places = []
|
||||
for index, frame in enumerate(frames):
|
||||
image = frame_to_pil(frame)
|
||||
image.thumbnail((PREVIEW_EDGE, PREVIEW_EDGE), Image.Resampling.LANCZOS)
|
||||
name = f"{token}_{index:03d}.webp"
|
||||
image.save(folder / name, format="WEBP", quality=88, method=2)
|
||||
places.append(f"{PREVIEW_FOLDER}/{name}")
|
||||
return places
|
||||
except Exception:
|
||||
return []
|
||||
|
||||
|
||||
def _png_info(metadata: dict | None):
|
||||
if not metadata:
|
||||
return None
|
||||
@@ -340,6 +371,8 @@ __all__ = [
|
||||
"IMAGE_FORMATS",
|
||||
"IMAGE_SAVE_LOCK",
|
||||
"OUTPUT_DIR_RULE",
|
||||
"PREVIEW_EDGE",
|
||||
"PREVIEW_FOLDER",
|
||||
"counter_pattern",
|
||||
"encode_image",
|
||||
"existing_action",
|
||||
@@ -357,4 +390,5 @@ __all__ = [
|
||||
"sidecar_path",
|
||||
"split_prefix",
|
||||
"strip_image_extension",
|
||||
"write_light_previews",
|
||||
]
|
||||
|
||||
@@ -653,10 +653,13 @@ def apply_stitch(
|
||||
exact with the toggle off, since the spread deliberately rewrites the
|
||||
feathered band.
|
||||
|
||||
``color_match`` (0..1) shifts the patch's tone by the offset measured
|
||||
in the feathered band (:func:`estimate_tone_offset`) before blending,
|
||||
so an outpaint whose new bands came out a touch lighter or warmer than
|
||||
the picture lands on the picture's own tone. 0 leaves the patch alone.
|
||||
``color_match`` (0..1) shifts the patch's tone by the measured drift
|
||||
before blending, so a new area that came out a touch lighter or warmer
|
||||
than the picture lands on the picture's own tone. An outpaint is read
|
||||
across its seams (:func:`tone_offset_field`); an inpaint, which has no
|
||||
source rectangle, inside its painted area (:func:`inner_tone_offset`),
|
||||
or in the feathered band (:func:`estimate_tone_offset`) when the
|
||||
painted area holds new content. 0 leaves the patch alone.
|
||||
|
||||
``seam`` "classic" is everything above. "blend in" replaces it on a
|
||||
stitcher that knows where the picture's edge is (Load Image + Pad and
|
||||
@@ -788,7 +791,13 @@ def apply_stitch(
|
||||
# Trimmed like the blend when fewer frames came back.
|
||||
generated = generated[:frames] if generated.shape[0] > frames else generated
|
||||
generated = generated[:, cy : cy + ch, cx : cx + cw].to(out.device)
|
||||
field = tone_offset_field(patch, out[:, cy : cy + ch, cx : cx + cw, :], alpha, bbox, generated)
|
||||
region = out[:, cy : cy + ch, cx : cx + cw, :]
|
||||
field = tone_offset_field(patch, region, alpha, bbox, generated)
|
||||
if bbox is None:
|
||||
# An inpaint has no source rectangle to read a seam across. Its
|
||||
# drift is read just inside the painted area, where it holds.
|
||||
inner, held = inner_tone_offset(patch, region, alpha, came_back)
|
||||
field = torch.where(held.view(-1, 1, 1, 1), inner.view(-1, 1, 1, 3).expand_as(field), field)
|
||||
# Weighted by the blend: a fully generated pixel drifted by the whole
|
||||
# field, a band pixel the sampler mixed at alpha drifted by alpha of
|
||||
# it, and the blend below scales the correction by alpha once more.
|
||||
@@ -841,6 +850,175 @@ def estimate_tone_offset(
|
||||
return torch.where(measured, offset, torch.zeros_like(offset))
|
||||
|
||||
|
||||
# An inpaint's drift is read in the painted area's own rim: this many
|
||||
# pixels in from where the paste becomes whole. The rim is where the new
|
||||
# part most often still shows what was there before (skin beside skin).
|
||||
INNER_RIM_PX = 32
|
||||
|
||||
# Rim pixels within this many LAB units of the rim's median change count as
|
||||
# the same content, drifted. The reading is only used when at least
|
||||
# INNER_MIN_SHARE of the rim agrees and there are INNER_MIN_PIXELS of it.
|
||||
INNER_SAME_CONTENT = 8.0
|
||||
INNER_MIN_SHARE = 0.5
|
||||
INNER_MIN_PIXELS = 64
|
||||
|
||||
# A larger change than this is new content (a jacket over skin), not a
|
||||
# drift: matching it would pull the new thing toward the old colour.
|
||||
INNER_MAX_SHIFT = 20.0
|
||||
|
||||
|
||||
# New content has no "before" in its own place (skin where a shirt was), so
|
||||
# it is read against the untouched picture beside it instead: the painted
|
||||
# area and the picture, each averaged within this many pixels of a point.
|
||||
NEIGHBOUR_REACH_PX = 48
|
||||
|
||||
# A point counts as on the seam when each side fills this much of its reach.
|
||||
NEIGHBOUR_MIN_SIDE = 0.1
|
||||
|
||||
# Seam points within INNER_MAX_SHIFT of each other show the same thing on
|
||||
# both sides (skin beside skin). Fewer than this share of the seam and the
|
||||
# new part is something else all round, so nothing is matched.
|
||||
NEIGHBOUR_MIN_SHARE = 0.25
|
||||
|
||||
|
||||
def _neighbour_offset(
|
||||
new: torch.Tensor, old: torch.Tensor, whole: torch.Tensor, outside: torch.Tensor
|
||||
) -> torch.Tensor | None:
|
||||
"""LAB offset of new content against the untouched picture beside it.
|
||||
|
||||
``new`` and ``old`` are one frame in LAB [H, W, 3]; ``whole`` marks the
|
||||
fully pasted pixels and ``outside`` the untouched ones. Along the seam
|
||||
the painted side's local mean is compared with the picture's. Where the
|
||||
two are alike the model drew more of what is already there, and the gap
|
||||
between them is its drift. None when too little of the seam is alike or
|
||||
the alike part does not agree on one gap.
|
||||
"""
|
||||
size = 2 * NEIGHBOUR_REACH_PX + 1
|
||||
|
||||
def local(values: torch.Tensor, mask: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
weight = mask.to(values.dtype).unsqueeze(-1)
|
||||
stack = torch.cat([values * weight, weight], dim=-1).movedim(-1, 0).unsqueeze(0)
|
||||
stack = functional.avg_pool2d(stack, (1, size), stride=1, padding=(0, NEIGHBOUR_REACH_PX))
|
||||
stack = functional.avg_pool2d(stack, (size, 1), stride=1, padding=(NEIGHBOUR_REACH_PX, 0))[0]
|
||||
return (stack[:3] / stack[3:].clamp_min(1e-6)).movedim(0, -1), stack[3]
|
||||
|
||||
painted, painted_share = local(new, whole)
|
||||
picture, picture_share = local(old, outside)
|
||||
seam = (painted_share >= NEIGHBOUR_MIN_SIDE) & (picture_share >= NEIGHBOUR_MIN_SIDE)
|
||||
gaps = (painted - picture)[seam]
|
||||
if gaps.shape[0] < INNER_MIN_PIXELS:
|
||||
return None
|
||||
alike = gaps.norm(dim=-1) <= INNER_MAX_SHIFT
|
||||
if int(alike.sum()) < INNER_MIN_PIXELS or float(alike.float().mean()) < NEIGHBOUR_MIN_SHARE:
|
||||
return None
|
||||
gaps = gaps[alike]
|
||||
agree = (gaps - gaps.median(dim=0).values).norm(dim=-1) < INNER_SAME_CONTENT
|
||||
if float(agree.float().mean()) < INNER_MIN_SHARE:
|
||||
return None
|
||||
return gaps[agree].median(dim=0).values
|
||||
|
||||
|
||||
def _area_resize(image: torch.Tensor, width: int, height: int) -> torch.Tensor:
|
||||
"""Shrink [B, H, W, C] by averaging the pixels each new one covers."""
|
||||
return functional.interpolate(image.movedim(-1, 1), size=(height, width), mode="area").movedim(1, -1)
|
||||
|
||||
|
||||
def inner_tone_offset(
|
||||
inpainted: torch.Tensor,
|
||||
canvas: torch.Tensor,
|
||||
blend: torch.Tensor,
|
||||
sampled: tuple[int, int] | None = None,
|
||||
) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""Per-frame LAB drift of an inpaint, read inside the painted area.
|
||||
|
||||
:func:`estimate_tone_offset` reads the feathered band and assumes the
|
||||
sampler mixed original and generated there by ``blend``. A workflow that
|
||||
samples with its own, wider or softer noise mask breaks that: the model
|
||||
leaves the band close to the original and drifts further in, so the band
|
||||
reads a fraction of the drift (issue #361: 0.8 of 2.5 LAB units).
|
||||
|
||||
This reads the same pixels before and after in the rim of the fully
|
||||
pasted area (``blend`` > 0.999, INNER_RIM_PX deep): the median change,
|
||||
then the median again over the pixels that agree with it. Returns the
|
||||
offsets [B, 3] and which frames hold [B]. A frame whose rim mostly
|
||||
disagrees, or changed by more than INNER_MAX_SHIFT, is new content
|
||||
rather than drift. That frame is read against the untouched picture
|
||||
beside it (:func:`_neighbour_offset`), and holds at zero, left as
|
||||
painted, when the picture beside it shows something else. Only a rim too
|
||||
small to judge does not hold, and the caller keeps the band reading for it.
|
||||
|
||||
``sampled`` is the (width, height) the model worked at. A crop from a
|
||||
large picture is pasted back several times that size, and a rim measured
|
||||
in pasted pixels would be a sliver at the very edge of what the model
|
||||
painted, where it drifts least. The reading is taken at the model's
|
||||
size instead, which also keeps its cost the same for any picture.
|
||||
"""
|
||||
from ._color_helpers import rgb_to_lab
|
||||
|
||||
if sampled is not None:
|
||||
width, height = int(sampled[0]), int(sampled[1])
|
||||
if 0 < width < inpainted.shape[2] and 0 < height < inpainted.shape[1]:
|
||||
inpainted = _area_resize(inpainted, width, height)
|
||||
canvas = _area_resize(canvas, width, height)
|
||||
blend = _area_resize(blend.unsqueeze(-1), width, height).squeeze(-1)
|
||||
new = rgb_to_lab(inpainted[..., :3])
|
||||
old = rgb_to_lab(canvas[..., :3])
|
||||
diff = new - old
|
||||
frames = diff.shape[0]
|
||||
whole = (blend > 0.999).to(diff.dtype)
|
||||
size = 2 * INNER_RIM_PX + 1
|
||||
# A pixel is deeper than the rim when nothing within INNER_RIM_PX of it
|
||||
# is outside the whole paste. The crop's own border does not count as
|
||||
# outside: the painted area simply runs off the crop there.
|
||||
# One pass along each axis: the same square reach at a fraction of the cost.
|
||||
outside = (1.0 - whole).unsqueeze(1)
|
||||
outside = functional.max_pool2d(outside, (1, size), stride=1, padding=(0, INNER_RIM_PX))
|
||||
outside = functional.max_pool2d(outside, (size, 1), stride=1, padding=(INNER_RIM_PX, 0))
|
||||
deep = 1.0 - outside.squeeze(1)
|
||||
rim = (whole > 0) & (deep < 0.5)
|
||||
rim = torch.where(rim.flatten(1).any(dim=1).view(-1, 1, 1), rim, whole > 0)
|
||||
pasted = whole > 0
|
||||
untouched = blend < 0.001
|
||||
if rim.shape[0] != frames:
|
||||
rim = rim.expand(frames, -1, -1)
|
||||
pasted = pasted.expand(frames, -1, -1)
|
||||
untouched = untouched.expand(frames, -1, -1)
|
||||
# See-through pixels hold whatever colour sits under the clear part.
|
||||
if inpainted.shape[-1] > 3:
|
||||
rim = rim & (inpainted[..., 3] >= 0.999)
|
||||
pasted = pasted & (inpainted[..., 3] >= 0.999)
|
||||
if canvas.shape[-1] > 3:
|
||||
rim = rim & (canvas[..., 3] >= 0.999)
|
||||
untouched = untouched & (canvas[..., 3] >= 0.999)
|
||||
|
||||
def beside(index: int) -> torch.Tensor:
|
||||
gap = _neighbour_offset(new[index], old[index], pasted[index], untouched[index])
|
||||
return offsets[index] if gap is None else gap
|
||||
|
||||
offsets = torch.zeros((frames, 3), dtype=diff.dtype, device=diff.device)
|
||||
held = torch.zeros((frames,), dtype=torch.bool, device=diff.device)
|
||||
for index in range(frames):
|
||||
changes = diff[index][rim[index]]
|
||||
if changes.shape[0] < INNER_MIN_PIXELS:
|
||||
continue
|
||||
# From here the rim is large enough to judge, so this reading stands
|
||||
# and the band is not asked. New content keeps a zero shift: the band
|
||||
# of a repainted edge reads the new colour itself as drift (a yellow
|
||||
# swimsuit over skin turned the whole patch blue).
|
||||
held[index] = True
|
||||
middle = changes.median(dim=0).values
|
||||
agree = (changes - middle).norm(dim=-1) < INNER_SAME_CONTENT
|
||||
if float(agree.float().mean()) < INNER_MIN_SHARE:
|
||||
offsets[index] = beside(index)
|
||||
continue
|
||||
offset = changes[agree].median(dim=0).values
|
||||
if float(offset.norm()) > INNER_MAX_SHIFT:
|
||||
offsets[index] = beside(index)
|
||||
continue
|
||||
offsets[index] = offset
|
||||
return offsets, held
|
||||
|
||||
|
||||
def _band_weights(blend: torch.Tensor) -> torch.Tensor:
|
||||
return torch.where((blend > 0.001) & (blend < 0.999), blend, torch.zeros_like(blend))
|
||||
|
||||
|
||||
@@ -61,11 +61,8 @@ def list_lama_models() -> list[str]:
|
||||
|
||||
def resolve_lama_model(model_name: str) -> Path:
|
||||
register_lama_model_folder()
|
||||
if folder_paths is None:
|
||||
path = Path(model_name).expanduser().resolve()
|
||||
if path.is_file():
|
||||
return path
|
||||
else:
|
||||
# Only a name from ComfyUI's lama model folder: a typed path is never opened.
|
||||
if folder_paths is not None:
|
||||
resolved = folder_paths.get_full_path(MODEL_FOLDER, model_name)
|
||||
if resolved:
|
||||
return Path(resolved).resolve()
|
||||
|
||||
@@ -0,0 +1,247 @@
|
||||
"""Load Image: the size a picture comes out at, as plain numbers.
|
||||
|
||||
``plan_load_size`` turns the node's size choice into what the picture is
|
||||
resized to and what is cut from it. It holds no tensors, so it is tested as
|
||||
numbers, and ``js/shared/load_size.mjs`` repeats it for the size the node
|
||||
shows before a run (``tests/fixtures/load_size_cases.json`` holds both to the
|
||||
same answers). ``apply_load_size`` runs a plan on a picture and its mask.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import math
|
||||
|
||||
import numpy as np
|
||||
import torch
|
||||
|
||||
from ._resize_helpers import plan_resize, resample_batch, resolve_target, snap_to_multiple
|
||||
|
||||
SIZE_MODES = (
|
||||
"off",
|
||||
"megapixels",
|
||||
"longest side",
|
||||
"scale",
|
||||
"fit inside",
|
||||
"crop to fill",
|
||||
"crop",
|
||||
"crop to ratio",
|
||||
)
|
||||
# Which part of the picture a crop keeps.
|
||||
CROP_SIDES = (
|
||||
"center",
|
||||
"top",
|
||||
"bottom",
|
||||
"left",
|
||||
"right",
|
||||
"top left",
|
||||
"top right",
|
||||
"bottom left",
|
||||
"bottom right",
|
||||
)
|
||||
# The last one takes its shape from the width and height boxes.
|
||||
RATIOS = (
|
||||
"1:1",
|
||||
"4:3",
|
||||
"3:4",
|
||||
"3:2",
|
||||
"2:3",
|
||||
"16:9",
|
||||
"9:16",
|
||||
"5:4",
|
||||
"4:5",
|
||||
"21:9",
|
||||
"9:21",
|
||||
"2:1",
|
||||
"1:2",
|
||||
"width : height",
|
||||
)
|
||||
|
||||
|
||||
def _round_half_up(value: float) -> int:
|
||||
return int(math.floor(value + 0.5))
|
||||
|
||||
|
||||
def _multiple_below(value: int, multiple: int) -> int:
|
||||
"""The largest multiple that is not above ``value``, never below one step."""
|
||||
step = max(1, int(multiple))
|
||||
return max(step, (int(value) // step) * step)
|
||||
|
||||
|
||||
def _multiple_inside(value: int, multiple: int) -> int:
|
||||
"""``value`` rounded down to a multiple, or left alone when it is smaller
|
||||
than one step: a cut never asks for more than the picture holds."""
|
||||
step = max(1, int(multiple))
|
||||
return (int(value) // step) * step if int(value) >= step else int(value)
|
||||
|
||||
|
||||
def _crop_start(spare: int, side: str, low: str, high: str) -> int:
|
||||
"""Where a cut starts along one edge with ``spare`` pixels to lose."""
|
||||
if low in side:
|
||||
return 0
|
||||
if high in side:
|
||||
return int(spare)
|
||||
return int(spare) // 2
|
||||
|
||||
|
||||
def _ratio_terms(ratio: str, width: int, height: int) -> tuple[int, int]:
|
||||
if ratio == "width : height":
|
||||
return max(1, int(width)), max(1, int(height))
|
||||
across, _, down = str(ratio).partition(":")
|
||||
return max(1, int(across)), max(1, int(down))
|
||||
|
||||
|
||||
def plan_load_size(
|
||||
source_width: int,
|
||||
source_height: int,
|
||||
size: str = "off",
|
||||
megapixels: float = 1.0,
|
||||
longest_side: int = 1024,
|
||||
scale: float = 1.0,
|
||||
width: int = 1024,
|
||||
height: int = 1024,
|
||||
multiple: int = 1,
|
||||
enlarge: bool = True,
|
||||
crop_from: str = "center",
|
||||
ratio: str = "1:1",
|
||||
) -> dict[str, int]:
|
||||
"""What one picture is resized to (``resize_width``/``resize_height``),
|
||||
where a crop starts in the resized picture (``crop_x``/``crop_y``) and
|
||||
the size that comes out (``width``/``height``).
|
||||
|
||||
``off`` gives the picture back as it is; ``multiple`` rounds only a size
|
||||
that is being changed. With ``enlarge`` off nothing comes out larger than
|
||||
the picture: a target above it leaves the picture at its own size, and
|
||||
``crop to fill`` cuts the box's shape out of the picture without growing
|
||||
it. ``fit inside`` rounds down, so the result never leaves its box.
|
||||
|
||||
``crop`` cuts a width x height piece out of the picture and ``crop to
|
||||
ratio`` cuts the largest piece of a shape; neither resizes. ``crop_from``
|
||||
names the part every cut keeps.
|
||||
"""
|
||||
source_w, source_h = int(source_width), int(source_height)
|
||||
if source_w <= 0 or source_h <= 0:
|
||||
raise ValueError("Load Image needs a picture larger than 0x0.")
|
||||
if size not in SIZE_MODES:
|
||||
raise ValueError(f"size must be one of {SIZE_MODES}, got '{size}'.")
|
||||
if crop_from not in CROP_SIDES:
|
||||
raise ValueError(f"crop_from must be one of {CROP_SIDES}, got '{crop_from}'.")
|
||||
if ratio not in RATIOS:
|
||||
raise ValueError(f"ratio must be one of {RATIOS}, got '{ratio}'.")
|
||||
if size == "off":
|
||||
return {
|
||||
"resize_width": source_w,
|
||||
"resize_height": source_h,
|
||||
"crop_x": 0,
|
||||
"crop_y": 0,
|
||||
"width": source_w,
|
||||
"height": source_h,
|
||||
}
|
||||
step = max(1, int(multiple))
|
||||
|
||||
def cut(resize_w: int, resize_h: int, out_w: int, out_h: int) -> dict[str, int]:
|
||||
return {
|
||||
"resize_width": resize_w,
|
||||
"resize_height": resize_h,
|
||||
"crop_x": _crop_start(resize_w - out_w, crop_from, "left", "right"),
|
||||
"crop_y": _crop_start(resize_h - out_h, crop_from, "top", "bottom"),
|
||||
"width": out_w,
|
||||
"height": out_h,
|
||||
}
|
||||
|
||||
if size == "crop":
|
||||
out_w = _multiple_inside(min(max(1, int(width)), source_w), step)
|
||||
out_h = _multiple_inside(min(max(1, int(height)), source_h), step)
|
||||
return cut(source_w, source_h, out_w, out_h)
|
||||
|
||||
if size == "crop to ratio":
|
||||
across, down = _ratio_terms(ratio, width, height)
|
||||
if source_w * down >= source_h * across:
|
||||
out_h = source_h
|
||||
out_w = min(source_w, max(1, _round_half_up(source_h * across / down)))
|
||||
else:
|
||||
out_w = source_w
|
||||
out_h = min(source_h, max(1, _round_half_up(source_w * down / across)))
|
||||
return cut(source_w, source_h, _multiple_inside(out_w, step), _multiple_inside(out_h, step))
|
||||
|
||||
if size == "crop to fill":
|
||||
box_w, box_h = max(1, int(width)), max(1, int(height))
|
||||
if not enlarge:
|
||||
grow = max(box_w / source_w, box_h / source_h)
|
||||
if grow > 1.0:
|
||||
# The box keeps its shape and shrinks to what the picture holds.
|
||||
box_w = max(1, int(math.floor(box_w / grow + 1e-6)))
|
||||
box_h = max(1, int(math.floor(box_h / grow + 1e-6)))
|
||||
canvas_w = min(_multiple_below(box_w, step), max(step, source_w))
|
||||
canvas_h = min(_multiple_below(box_h, step), max(step, source_h))
|
||||
else:
|
||||
canvas_w = snap_to_multiple(box_w, step)
|
||||
canvas_h = snap_to_multiple(box_h, step)
|
||||
plan = plan_resize(source_w, source_h, canvas_w, canvas_h, "cover_crop", 1)
|
||||
return cut(plan["resize_width"], plan["resize_height"], canvas_w, canvas_h)
|
||||
|
||||
if size == "megapixels":
|
||||
target_w, target_h = resolve_target(source_w, source_h, "megapixels", megapixels=megapixels)
|
||||
elif size == "longest side":
|
||||
target_w, target_h = resolve_target(source_w, source_h, "longest_edge", edge_length=longest_side)
|
||||
elif size == "scale":
|
||||
target_w, target_h = resolve_target(source_w, source_h, "scale_factor", scale_factor=scale)
|
||||
else: # fit inside
|
||||
fit = min(max(1, int(width)) / source_w, max(1, int(height)) / source_h)
|
||||
target_w = max(1, _round_half_up(source_w * fit))
|
||||
target_h = max(1, _round_half_up(source_h * fit))
|
||||
if not enlarge and (target_w > source_w or target_h > source_h):
|
||||
target_w, target_h = source_w, source_h
|
||||
|
||||
def rounded(target: int, source: int, box: int) -> int:
|
||||
value = snap_to_multiple(target, step)
|
||||
if size == "fit inside" and value > box:
|
||||
value = _multiple_below(box, step)
|
||||
if not enlarge and value > source:
|
||||
value = _multiple_below(source, step)
|
||||
return value
|
||||
|
||||
out_w = rounded(target_w, source_w, max(1, int(width)))
|
||||
out_h = rounded(target_h, source_h, max(1, int(height)))
|
||||
return {
|
||||
"resize_width": out_w,
|
||||
"resize_height": out_h,
|
||||
"crop_x": 0,
|
||||
"crop_y": 0,
|
||||
"width": out_w,
|
||||
"height": out_h,
|
||||
}
|
||||
|
||||
|
||||
def frames_to_image_and_mask(frames) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""RGBA PIL frames to a BHWC RGB batch and a BHW mask.
|
||||
|
||||
The mask is what ComfyUI's own Load Image gives: white where the picture
|
||||
is see-through, black where it is solid, and all black for a picture
|
||||
with no transparency (here at the picture's own size).
|
||||
"""
|
||||
images, masks = [], []
|
||||
for frame in frames:
|
||||
array = np.asarray(frame.convert("RGBA"), dtype=np.float32) / 255.0
|
||||
images.append(torch.from_numpy(array[..., :3].copy()))
|
||||
masks.append(torch.from_numpy(1.0 - array[..., 3]))
|
||||
return torch.stack(images, dim=0), torch.stack(masks, dim=0)
|
||||
|
||||
|
||||
def apply_load_size(
|
||||
image: torch.Tensor, mask: torch.Tensor, plan: dict[str, int], method: str
|
||||
) -> tuple[torch.Tensor, torch.Tensor]:
|
||||
"""Run a plan on a BHWC picture and its BHW mask. A plan that changes
|
||||
nothing hands both back untouched."""
|
||||
resize_w, resize_h = int(plan["resize_width"]), int(plan["resize_height"])
|
||||
out_w, out_h = int(plan["width"]), int(plan["height"])
|
||||
if (int(image.shape[2]), int(image.shape[1])) != (resize_w, resize_h):
|
||||
image = resample_batch(image, resize_w, resize_h, method)
|
||||
mask = resample_batch(mask.unsqueeze(-1), resize_w, resize_h, method).squeeze(-1).clamp(0.0, 1.0)
|
||||
if (resize_w, resize_h) != (out_w, out_h):
|
||||
x0, y0 = int(plan["crop_x"]), int(plan["crop_y"])
|
||||
image = image[:, y0 : y0 + out_h, x0 : x0 + out_w, :].contiguous()
|
||||
mask = mask[:, y0 : y0 + out_h, x0 : x0 + out_w].contiguous()
|
||||
return image, mask
|
||||
|
||||
|
||||
__all__ = ["CROP_SIDES", "RATIOS", "SIZE_MODES", "apply_load_size", "frames_to_image_and_mask", "plan_load_size"]
|
||||
+23
-19
@@ -13,6 +13,8 @@ from collections import OrderedDict
|
||||
from pathlib import Path
|
||||
from typing import Any
|
||||
|
||||
from ._route_errors import error_detail
|
||||
|
||||
try:
|
||||
import folder_paths
|
||||
except ImportError: # pragma: no cover - only outside ComfyUI
|
||||
@@ -136,27 +138,29 @@ _remap_noted: set[str] = set()
|
||||
def resolve_lora_path(name: str) -> Path:
|
||||
if folder_paths is None:
|
||||
raise ValueError("folder_paths is unavailable outside ComfyUI.")
|
||||
full = folder_paths.get_full_path("loras", name)
|
||||
if not full:
|
||||
# Only a name on ComfyUI's own LoRA list reaches the disk. A route's
|
||||
# name or a widget value is never joined to a folder as given: it is
|
||||
# matched against the list first, and anything else is refused here.
|
||||
candidate, status = match_lora_name(str(name), list_loras())
|
||||
if status == "remapped":
|
||||
# The saved path is stale; a moved file should just keep working.
|
||||
candidate, status = match_lora_name(str(name), list_loras())
|
||||
if status == "remapped":
|
||||
if name not in _remap_noted:
|
||||
_remap_noted.add(name)
|
||||
print(
|
||||
f"[AusBoss] LoRA Loader: '{name}' is not at its saved "
|
||||
f"path; using '{candidate}'."
|
||||
)
|
||||
full = folder_paths.get_full_path("loras", candidate)
|
||||
elif status == "ambiguous":
|
||||
raise ValueError(
|
||||
f"LoRA name matches several files in models/loras: {name}"
|
||||
if name not in _remap_noted:
|
||||
_remap_noted.add(name)
|
||||
print(
|
||||
f"[AusBoss] LoRA Loader: '{name}' is not at its saved "
|
||||
f"path; using '{candidate}'."
|
||||
)
|
||||
elif status == "ambiguous":
|
||||
raise ValueError(
|
||||
f"LoRA name matches several files in models/loras: {name}"
|
||||
)
|
||||
elif status != "exact":
|
||||
raise ValueError(f"LoRA file not found in models/loras: {name}")
|
||||
full = folder_paths.get_full_path("loras", candidate)
|
||||
if not full:
|
||||
raise ValueError(f"LoRA file not found in models/loras: {name}")
|
||||
# Containment is checked on the path as written, with ".." folded but
|
||||
# links not followed: a LoRA kept on another drive and linked into
|
||||
# models/loras is the user's own choice, and ComfyUI loads it too.
|
||||
# The listed file must also sit below one of the loras folders, read
|
||||
# from the path as ComfyUI built it, with ".." folded.
|
||||
path = Path(os.path.abspath(full))
|
||||
roots = [Path(os.path.abspath(root)) for root in folder_paths.get_folder_paths("loras")]
|
||||
if not any(root in path.parents for root in roots):
|
||||
@@ -388,8 +392,8 @@ def apply_lora_stack(
|
||||
except MemoryError:
|
||||
raise
|
||||
except Exception as exc:
|
||||
# ComfyUI's own message here is PyTorch's and never names the file.
|
||||
detail = f"{type(exc).__name__}: {' '.join(str(exc).split())}"
|
||||
# ComfyUI can add the file's path; error_detail drops words that name one.
|
||||
detail = error_detail(exc)
|
||||
raise ValueError(
|
||||
unreadable_lora_message(row["name"], unreadable_lora_reason(path), detail)
|
||||
) from exc
|
||||
|
||||
@@ -41,6 +41,16 @@ RADIUS_MAX = 4096.0
|
||||
# The face's own picture: the long side is at most this.
|
||||
SMALL_SIDE = 1024
|
||||
|
||||
# SAM 3 is asked about a copy of a big picture: its long side is at most
|
||||
# this. ComfyUI's SAM 3 reads every picture at 1008 pixels and returns each
|
||||
# answer at the size of the picture it was given, on the graphics card. At
|
||||
# 8192 pixels one word with sixteen matches needed over 27 GB there, and a
|
||||
# look at 3840 took twice as long as at 2048, all for outlines the model
|
||||
# never saw that finely. Twice what the model reads keeps its second pass
|
||||
# (a close-up of each thing it found) as sharp as before on pictures up to
|
||||
# this size, which come out exactly as they did.
|
||||
WORK_SIDE = 2048
|
||||
|
||||
# How a find stroke is asked (find_points, choose_found). Measured on 27
|
||||
# strokes over three pictures, each drawn 13 slightly different ways: four
|
||||
# points alone returned the stroked thing in 87 % of 351 tries, two points
|
||||
@@ -386,6 +396,73 @@ def read_ask(ask, height: int, width: int) -> tuple:
|
||||
return "look", None
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# The copy of a big picture that SAM 3 is asked about
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
|
||||
class WorkView:
|
||||
"""One picture as SAM 3 is asked about it.
|
||||
|
||||
A picture whose long side is over WORK_SIDE is asked about on a smaller
|
||||
copy (`frame`, averaged down, so nothing shimmers), and what SAM 3
|
||||
answers there is brought back to the picture's own size with `full`.
|
||||
A smaller picture is its own copy: `same` is true and nothing changes.
|
||||
|
||||
Touches and pieces keep their places in pixels of the picture. `point`
|
||||
gives a place on the copy, `full_point` a place on the picture that
|
||||
`point` leads back to the same pixel of the copy.
|
||||
"""
|
||||
|
||||
def __init__(self, frame: torch.Tensor):
|
||||
self.height, self.width = int(frame.shape[1]), int(frame.shape[2])
|
||||
long_side = max(self.height, self.width)
|
||||
self.same = long_side <= WORK_SIDE
|
||||
if self.same:
|
||||
self.frame, self.work_height, self.work_width = frame, self.height, self.width
|
||||
return
|
||||
scale = WORK_SIDE / long_side
|
||||
self.work_height = max(1, int(math.floor(self.height * scale + 0.5)))
|
||||
self.work_width = max(1, int(math.floor(self.width * scale + 0.5)))
|
||||
small = F.interpolate(
|
||||
frame.detach().float().cpu().movedim(-1, 1), size=(self.work_height, self.work_width), mode="area",
|
||||
)
|
||||
self.frame = small.movedim(1, -1).clamp(0.0, 1.0).contiguous()
|
||||
|
||||
def point(self, point) -> tuple:
|
||||
x = min(self.work_width - 1, max(0, int(point[0]) * self.work_width // self.width))
|
||||
y = min(self.work_height - 1, max(0, int(point[1]) * self.work_height // self.height))
|
||||
return (x, y)
|
||||
|
||||
def full_point(self, point) -> tuple:
|
||||
x = min(self.width - 1, -(-int(point[0]) * self.width // self.work_width))
|
||||
y = min(self.height - 1, -(-int(point[1]) * self.height // self.work_height))
|
||||
return (x, y)
|
||||
|
||||
def points(self, points) -> tuple:
|
||||
return tuple(self.point(point) for point in points)
|
||||
|
||||
def length(self, value: float) -> float:
|
||||
"""A length of the picture (a brush's radius) on the copy."""
|
||||
return float(value) * self.work_width / self.width
|
||||
|
||||
def full_box(self, box) -> list:
|
||||
if self.same:
|
||||
return list(box)
|
||||
x0, y0, x1, y1 = (int(value) for value in box)
|
||||
return [
|
||||
x0 * self.width // self.work_width, y0 * self.height // self.work_height,
|
||||
min(self.width, -(-x1 * self.width // self.work_width)), min(self.height, -(-y1 * self.height // self.work_height)),
|
||||
]
|
||||
|
||||
def full(self, mask: torch.Tensor) -> torch.Tensor:
|
||||
"""An answer from the copy at the picture's own size (bool)."""
|
||||
if self.same or tuple(mask.shape) == (self.height, self.width):
|
||||
return mask
|
||||
soft = F.interpolate(mask[None, None].float(), size=(self.height, self.width), mode="bilinear", align_corners=False)
|
||||
return soft[0, 0] >= 0.5
|
||||
|
||||
|
||||
# ---------------------------------------------------------------------------
|
||||
# Strokes
|
||||
# ---------------------------------------------------------------------------
|
||||
|
||||
@@ -242,13 +242,20 @@ def _locate_comfy_file(text: str, rule: str) -> tuple[str, str, Path]:
|
||||
if (
|
||||
parts[0] == "~"
|
||||
or any(part in ("", ".", "..") for part in parts)
|
||||
or any(re.match(r"[A-Za-z]:", part) for part in parts)
|
||||
or (os.name == "nt" and ":" in inside)
|
||||
or re.search(r"[\x00-\x1f\x7f]", inside)
|
||||
):
|
||||
raise ValueError(rule)
|
||||
root = _server_folders().get(folder)
|
||||
if root is None:
|
||||
raise ValueError(rule)
|
||||
path = root.joinpath(*parts).resolve()
|
||||
# Checked as text before the disk resolves it, and again after, so a
|
||||
# link that leads out of the folder is refused too.
|
||||
lexical = root.joinpath(*parts)
|
||||
if root not in lexical.parents:
|
||||
raise ValueError(rule)
|
||||
path = lexical.resolve()
|
||||
if root not in path.parents:
|
||||
raise ValueError(rule)
|
||||
return folder, inside, path
|
||||
|
||||
@@ -0,0 +1,35 @@
|
||||
"""Pick one choice from each {a|b|c} set in a prompt, decided by a seed.
|
||||
|
||||
Only a set that holds a ``|`` is a choice. Braces without one are left as
|
||||
they are, so a JSON prompt passes through unchanged. Sets can sit inside
|
||||
other sets; the innermost ones are picked first.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import random
|
||||
import re
|
||||
|
||||
# An innermost {...} that holds at least one "|", with the one space in
|
||||
# front of it, so an empty pick does not leave two spaces behind.
|
||||
_CHOICE = re.compile(r"( ?)\{([^{}]*\|[^{}]*)\}")
|
||||
|
||||
|
||||
def pick_variations(text: str, seed: int) -> tuple[str, list[str]]:
|
||||
"""Return the prompt with every choice set replaced, and the picks in order."""
|
||||
rng = random.Random(int(seed))
|
||||
picks: list[str] = []
|
||||
|
||||
def choose(match: re.Match) -> str:
|
||||
options = [option.strip() for option in match.group(2).split("|")]
|
||||
pick = rng.choice(options)
|
||||
picks.append(pick)
|
||||
return match.group(1) + pick if pick else ""
|
||||
|
||||
while True:
|
||||
text, count = _CHOICE.subn(choose, text)
|
||||
if not count:
|
||||
return text, picks
|
||||
|
||||
|
||||
__all__ = ["pick_variations"]
|
||||
+15
-3
@@ -1,11 +1,12 @@
|
||||
"""What a route says back when a request fails."""
|
||||
"""What a route or a node says back when something fails."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import re
|
||||
|
||||
# A whole path: a word that opens with "/", "\" or a drive letter.
|
||||
_SERVER_PATH = re.compile(r"""(?:^|[\s'"(\[=:,;])(?:[A-Za-z]:[\\/]|[\\/])[^\s'"]""")
|
||||
# A whole path: a word that opens with "/", "\", "~/" or a drive letter,
|
||||
# after anything but a letter, digit, "_" or "." ("and/or" and "3/4" stay).
|
||||
_SERVER_PATH = re.compile(r"""(?:^|[^\w.])(?:[A-Za-z]:[\\/]|~[\\/]|[\\/])[^\s'"]""")
|
||||
|
||||
|
||||
def safe_route_error(exc: Exception, fallback: str = "The request failed.") -> str:
|
||||
@@ -16,3 +17,14 @@ def safe_route_error(exc: Exception, fallback: str = "The request failed.") -> s
|
||||
if not text or _SERVER_PATH.search(text):
|
||||
return fallback
|
||||
return text
|
||||
|
||||
|
||||
def error_detail(exc: BaseException) -> str:
|
||||
"""A library's error for a node's message: its kind and its words, or
|
||||
its kind alone when the words name a place on the server's disk. Node
|
||||
errors reach the browser that queued the run and stay in the history."""
|
||||
kind = type(exc).__name__
|
||||
text = " ".join(str(exc).split())
|
||||
if not text or _SERVER_PATH.search(text):
|
||||
return kind
|
||||
return f"{kind}: {text}"
|
||||
|
||||
@@ -0,0 +1,97 @@
|
||||
"""Only one copy of the pack loads.
|
||||
|
||||
A Manager install lands in ``custom_nodes/ausboss-nodes`` and a git clone in
|
||||
``custom_nodes/ComfyUI-AusBoss``. With both in place every route is added
|
||||
twice and ComfyUI stops while starting. So each copy looks for the others
|
||||
and all of them agree on the one that loads; the rest step aside and say
|
||||
which folder to delete.
|
||||
|
||||
A copy from before this file existed cannot step aside, so it always loads
|
||||
and a newer copy yields to it. Among copies that can, the newest version
|
||||
loads.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import os
|
||||
import tomllib
|
||||
|
||||
PACK_NAME = "ausboss-nodes"
|
||||
# A copy that has this file knows how to step aside.
|
||||
_GUARD_FILE = os.path.join("nodes", "_single_copy.py")
|
||||
|
||||
|
||||
def _read_pack(folder: str) -> dict | None:
|
||||
"""Name the pack copy in ``folder``, or None when it is another pack."""
|
||||
try:
|
||||
with open(os.path.join(folder, "pyproject.toml"), "rb") as f:
|
||||
project = tomllib.load(f).get("project", {})
|
||||
except Exception:
|
||||
return None
|
||||
if project.get("name") != PACK_NAME:
|
||||
return None
|
||||
if not os.path.isfile(os.path.join(folder, "__init__.py")):
|
||||
return None
|
||||
version = project.get("version")
|
||||
return {
|
||||
"folder": folder,
|
||||
"version": version if isinstance(version, str) else "unknown",
|
||||
"guarded": os.path.isfile(os.path.join(folder, _GUARD_FILE)),
|
||||
}
|
||||
|
||||
|
||||
def _version_key(version: str) -> tuple:
|
||||
parts = []
|
||||
for part in version.split("."):
|
||||
digits = "".join(ch for ch in part if ch.isdigit())
|
||||
parts.append(int(digits) if digits else -1)
|
||||
return tuple(parts)
|
||||
|
||||
|
||||
def find_copies(own_folder: str, roots: list[str]) -> list[dict]:
|
||||
"""Every copy of the pack below ``roots``, this one included."""
|
||||
own = os.path.realpath(own_folder)
|
||||
seen = {own}
|
||||
copies = [_read_pack(own) or {"folder": own, "version": "unknown", "guarded": True}]
|
||||
for root in roots:
|
||||
try:
|
||||
names = sorted(os.listdir(root))
|
||||
except OSError:
|
||||
continue
|
||||
for name in names:
|
||||
# ComfyUI skips these too.
|
||||
if name.endswith(".disabled") or name.startswith("."):
|
||||
continue
|
||||
folder = os.path.realpath(os.path.join(root, name))
|
||||
if folder in seen or not os.path.isdir(folder):
|
||||
continue
|
||||
seen.add(folder)
|
||||
pack = _read_pack(folder)
|
||||
if pack:
|
||||
copies.append(pack)
|
||||
return copies
|
||||
|
||||
|
||||
def pick_copy(copies: list[dict]) -> dict:
|
||||
"""The one copy that loads. Every copy reaches the same answer."""
|
||||
unguarded = [c for c in copies if not c["guarded"]]
|
||||
pool = unguarded or copies
|
||||
return max(pool, key=lambda c: (_version_key(c["version"]), c["folder"]))
|
||||
|
||||
|
||||
def other_copy_that_loads(own_folder: str, roots: list[str] | None = None) -> dict | None:
|
||||
"""The copy this one steps aside for, or None when this one loads."""
|
||||
if roots is None:
|
||||
# Outside ComfyUI nothing is served, so there is nothing to clash.
|
||||
try:
|
||||
import folder_paths
|
||||
|
||||
roots = list(folder_paths.get_folder_paths("custom_nodes"))
|
||||
except Exception:
|
||||
return None
|
||||
roots = list(dict.fromkeys(os.path.realpath(r) for r in roots))
|
||||
copies = find_copies(own_folder, roots)
|
||||
if len(copies) < 2:
|
||||
return None
|
||||
chosen = pick_copy(copies)
|
||||
return None if chosen["folder"] == os.path.realpath(own_folder) else chosen
|
||||
@@ -0,0 +1,67 @@
|
||||
"""Cut a block of text into the items Text Lines runs one at a time.
|
||||
|
||||
Pure text handling with no ComfyUI imports, so it runs under the
|
||||
standard-library tests.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import re
|
||||
|
||||
SPLIT_EACH_LINE = "each line"
|
||||
SPLIT_BLANK_LINE = "blank line"
|
||||
SPLIT_MODES = (SPLIT_EACH_LINE, SPLIT_BLANK_LINE)
|
||||
|
||||
# The words a list writer puts before an item's number: "Prompt 3:", "Shot 2 -".
|
||||
# Not "image" or "picture": an edit prompt may open with "Image 1: ...".
|
||||
_LABEL = r"(?:prompt|shot|scene|line|step|item)"
|
||||
|
||||
# A marker at the start of an item that only counts or bullets it. A bare
|
||||
# number needs punctuation after it, and that punctuation must not be followed
|
||||
# by a digit, so "2 women on a beach", "1.5x zoom" and "10:30 at night" stay
|
||||
# whole. A bullet needs a space after it, so "*bold* text" stays whole.
|
||||
_MARKER = re.compile(
|
||||
r"""^\s*(?:
|
||||
[\[【(]\s*(?:%(label)s[\s_#-]*)?\d+\s*[\]】)]\s*[:.\-–—]? # [prompt 1] (2)
|
||||
| %(label)s[\s_#-]*\d+\s*[:.)\-–—] # Prompt 1: Shot 2 -
|
||||
| \#?\d+\s*[.):](?!\d) # 1. 2) #3:
|
||||
| [-*•·–—](?=\s|$) # - bullet
|
||||
)\s*"""
|
||||
% {"label": _LABEL},
|
||||
re.IGNORECASE | re.VERBOSE,
|
||||
)
|
||||
|
||||
_BLANK_LINE = re.compile(r"\n[ \t]*(?:\n[ \t]*)+")
|
||||
|
||||
|
||||
def strip_marker(item: str) -> str:
|
||||
"""Take one leading number or bullet off an item."""
|
||||
return _MARKER.sub("", item, count=1).strip()
|
||||
|
||||
|
||||
def split_items(text, split=SPLIT_EACH_LINE, remove_numbering=True, limit=0) -> list[str]:
|
||||
"""The items of `text`, in order, with empty ones left out.
|
||||
|
||||
`split` picks where one item ends: at every line break, or only at an
|
||||
empty line (so an item can hold several lines). `remove_numbering` takes
|
||||
a leading "1." or "-" off each item. `limit` keeps the first so many;
|
||||
0 keeps them all.
|
||||
"""
|
||||
text = str(text or "").replace("\r\n", "\n").replace("\r", "\n")
|
||||
if split == SPLIT_BLANK_LINE:
|
||||
parts = _BLANK_LINE.split(text)
|
||||
else:
|
||||
parts = text.split("\n")
|
||||
|
||||
items = []
|
||||
for part in parts:
|
||||
item = part.strip()
|
||||
if remove_numbering:
|
||||
item = strip_marker(item)
|
||||
if item:
|
||||
items.append(item)
|
||||
|
||||
limit = int(limit or 0)
|
||||
if limit > 0:
|
||||
items = items[:limit]
|
||||
return items
|
||||
+19
-11
@@ -480,17 +480,25 @@ def transform_tensor_batch_chunked(
|
||||
return output, mask, geometry
|
||||
|
||||
|
||||
def stable_file_fingerprint(path: str | os.PathLike[str], inputs: dict[str, object]) -> str:
|
||||
normalized_path = str(Path(path).resolve()).casefold()
|
||||
try:
|
||||
stat = os.stat(path)
|
||||
file_state: dict[str, object] = {
|
||||
"path": normalized_path,
|
||||
"mtime_ns": stat.st_mtime_ns,
|
||||
"size": stat.st_size,
|
||||
}
|
||||
except OSError:
|
||||
file_state = {"path": normalized_path, "missing": True}
|
||||
def stable_file_fingerprint(path: Path | None, inputs: dict[str, object]) -> str:
|
||||
"""A key that changes when the inputs or the file do.
|
||||
|
||||
`path` is what resolve_input_path or resolve_video_path returned, or
|
||||
None when they refused the value. Only such a path is ever looked at on
|
||||
disk: a refused value is a string, and a string is only hashed, so a
|
||||
whole path or a network share typed into a widget is never opened."""
|
||||
if isinstance(path, Path):
|
||||
try:
|
||||
stat = os.stat(path)
|
||||
file_state: dict[str, object] = {
|
||||
"path": str(path).casefold(),
|
||||
"mtime_ns": stat.st_mtime_ns,
|
||||
"size": stat.st_size,
|
||||
}
|
||||
except OSError:
|
||||
file_state = {"path": str(path).casefold(), "missing": True}
|
||||
else:
|
||||
file_state = {"unresolved": True}
|
||||
payload = {"file": file_state, "inputs": inputs}
|
||||
encoded = json.dumps(payload, sort_keys=True, separators=(",", ":"), ensure_ascii=True)
|
||||
return hashlib.sha256(encoded.encode("utf-8")).hexdigest()
|
||||
|
||||
@@ -13,6 +13,7 @@ import numpy as np
|
||||
import torch
|
||||
from av.video.reformatter import ColorRange, Colorspace
|
||||
|
||||
from ._route_errors import error_detail
|
||||
from ._execution_helpers import advance_progress, frame_progress, raise_if_interrupted
|
||||
|
||||
AAC_FRAME_SIZE = 1024
|
||||
@@ -363,9 +364,9 @@ def _encoder_start_error(codec: str, exc: Exception) -> RuntimeError:
|
||||
return RuntimeError(
|
||||
"Save Video: the nvenc formats need an NVIDIA GPU with a working "
|
||||
"driver. Pick a software format such as 'mp4 h264' instead. "
|
||||
f"(The '{codec}' encoder said: {exc})"
|
||||
f"(The '{codec}' encoder said: {error_detail(exc)})"
|
||||
)
|
||||
return RuntimeError(f"Save Video could not start the '{codec}' encoder: {exc}.")
|
||||
return RuntimeError(f"Save Video could not start the '{codec}' encoder: {error_detail(exc)}.")
|
||||
|
||||
|
||||
def _add_video_stream(container, spec: VideoFormat, rate, width: int, height: int, crf: int):
|
||||
|
||||
@@ -0,0 +1,242 @@
|
||||
"""Edit Lock 🆎 and Edit Lock Restore 🆎: an edit leaves the rest of the picture alone.
|
||||
|
||||
An edit model redraws the whole picture, so the parts nobody asked to change
|
||||
come back a little moved and a little tinted. Edit Lock watches the model's
|
||||
own prediction at every step, reads where it differs from the original, and
|
||||
puts the original back everywhere else, so the model finishes the edit on
|
||||
top of the real picture. Edit Lock Restore then swaps the original's own
|
||||
pixels in for those parts after the decode. The deciding and blending live
|
||||
in _edit_lock_helpers.py.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import torch
|
||||
|
||||
from ._edit_lock_helpers import (
|
||||
LOCK_MODES,
|
||||
EditLock,
|
||||
describe,
|
||||
restore_pixels,
|
||||
)
|
||||
from ._edit_tint_helpers import take_out_tint
|
||||
|
||||
|
||||
class AusBossEditLock:
|
||||
CATEGORY = "🆎 AusBoss/Image"
|
||||
DESCRIPTION = (
|
||||
"Keeps every part of your picture that an edit did not change "
|
||||
"exactly where it was, with the colors it had. Edit models redraw "
|
||||
"the whole picture, so untouched parts often come back slightly "
|
||||
"zoomed, shifted or tinted. Put this between the model and the "
|
||||
"sampler and give it the picture you are editing. It needs no "
|
||||
"mask: it sees what the edit changes while the picture is made. "
|
||||
"When the edit changes the whole picture (a new style, new light, "
|
||||
"new weather) it steps aside. Add Edit Lock Restore 🆎 after the "
|
||||
"decode to get your own pixels back in the untouched parts."
|
||||
)
|
||||
SEARCH_ALIASES = [
|
||||
"edit lock",
|
||||
"pixel drift",
|
||||
"pixel shift",
|
||||
"color shift",
|
||||
"consistency",
|
||||
"keep unchanged",
|
||||
"preserve original",
|
||||
"tail steer",
|
||||
"no drift edit",
|
||||
"ausboss",
|
||||
]
|
||||
|
||||
@classmethod
|
||||
def INPUT_TYPES(cls):
|
||||
return {
|
||||
"required": {
|
||||
"model": (
|
||||
"MODEL",
|
||||
{"tooltip": "The edit model. Patch last, after any LoRAs."},
|
||||
),
|
||||
"vae": (
|
||||
"VAE",
|
||||
{"tooltip": "The same VAE the workflow decodes with."},
|
||||
),
|
||||
"image": (
|
||||
"IMAGE",
|
||||
{
|
||||
"tooltip": (
|
||||
"The picture you are editing, at the size of the "
|
||||
"result. Use the same picture the edit model gets."
|
||||
),
|
||||
},
|
||||
),
|
||||
"lock": (
|
||||
list(LOCK_MODES),
|
||||
{
|
||||
"default": LOCK_MODES[0],
|
||||
"tooltip": (
|
||||
"auto locks the untouched parts and steps aside "
|
||||
"when the whole picture changes. always locks "
|
||||
"even then. off does nothing."
|
||||
),
|
||||
},
|
||||
),
|
||||
"tolerance": (
|
||||
"FLOAT",
|
||||
{
|
||||
"default": 1.0,
|
||||
"min": 0.25,
|
||||
"max": 4.0,
|
||||
"step": 0.05,
|
||||
"tooltip": (
|
||||
"How big a change must be to count as part of the "
|
||||
"edit. Raise it if the lock lets small unwanted "
|
||||
"changes through; lower it if it holds back part "
|
||||
"of your edit."
|
||||
),
|
||||
},
|
||||
),
|
||||
},
|
||||
"optional": {
|
||||
"mask": (
|
||||
"MASK",
|
||||
{
|
||||
"tooltip": (
|
||||
"Optional. White marks where the edit may happen; "
|
||||
"everything else is locked. With a mask the node "
|
||||
"does not look for the change itself."
|
||||
),
|
||||
},
|
||||
),
|
||||
},
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("MODEL", "AUSBOSS_EDIT_LOCK")
|
||||
RETURN_NAMES = ("model", "edit_lock")
|
||||
OUTPUT_TOOLTIPS = (
|
||||
"The model with the lock on. Plug it into the sampler.",
|
||||
"What the lock kept. Plug it into Edit Lock Restore 🆎 after the decode.",
|
||||
)
|
||||
FUNCTION = "patch"
|
||||
|
||||
def patch(self, model, vae, image, lock, tolerance, mask=None):
|
||||
if image.ndim != 4:
|
||||
raise ValueError("Edit Lock: image must be an IMAGE batch.")
|
||||
state = EditLock(mode=str(lock), tolerance=float(tolerance))
|
||||
source = image[:1, ..., :3]
|
||||
state.source = source.detach().float().cpu()
|
||||
if mask is not None:
|
||||
state.user_mask = mask.detach().float().cpu()
|
||||
if state.mode == "off":
|
||||
state.status = "off"
|
||||
return (model, state)
|
||||
latent = vae.encode(source)
|
||||
patched = model.clone()
|
||||
state.latent = patched.model.process_latent_in(latent).detach().float().cpu()
|
||||
patched.set_model_sampler_post_cfg_function(state.step)
|
||||
return (patched, state)
|
||||
|
||||
|
||||
class AusBossEditLockRestore:
|
||||
CATEGORY = "🆎 AusBoss/Image"
|
||||
DESCRIPTION = (
|
||||
"Puts your picture's own pixels back in every part the edit did not "
|
||||
"change. Use it right after the VAE decode, with the edit_lock "
|
||||
"output of Edit Lock 🆎. The untouched parts then match your "
|
||||
"original pixel for pixel; only the edit is new. It also takes out "
|
||||
"the faint pink tint some edit models add to the whole picture, "
|
||||
"which matters most when the lock stepped aside (a new hair color, "
|
||||
"background or weather)."
|
||||
)
|
||||
SEARCH_ALIASES = [
|
||||
"edit lock restore",
|
||||
"restore original pixels",
|
||||
"paste back",
|
||||
"keep unchanged",
|
||||
"ausboss",
|
||||
]
|
||||
|
||||
@classmethod
|
||||
def INPUT_TYPES(cls):
|
||||
return {
|
||||
"required": {
|
||||
"image": (
|
||||
"IMAGE",
|
||||
{"tooltip": "The edited picture, straight from the VAE decode."},
|
||||
),
|
||||
"edit_lock": (
|
||||
"AUSBOSS_EDIT_LOCK",
|
||||
{"tooltip": "From Edit Lock 🆎."},
|
||||
),
|
||||
"feather": (
|
||||
"INT",
|
||||
{
|
||||
"default": 16,
|
||||
"min": 0,
|
||||
"max": 256,
|
||||
"step": 1,
|
||||
"tooltip": (
|
||||
"How softly your original fades into the edit, in "
|
||||
"pixels. The fade stays outside the edit."
|
||||
),
|
||||
},
|
||||
),
|
||||
},
|
||||
"optional": {
|
||||
"remove_tint": (
|
||||
"BOOLEAN",
|
||||
{
|
||||
"default": True,
|
||||
"tooltip": (
|
||||
"Takes out the faint pink tint the edit model adds "
|
||||
"to the whole picture. Edits meant to recolor "
|
||||
"everything, like golden hour, are left alone."
|
||||
),
|
||||
},
|
||||
),
|
||||
},
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("IMAGE", "MASK", "STRING")
|
||||
RETURN_NAMES = ("image", "edit_mask", "report")
|
||||
OUTPUT_TOOLTIPS = (
|
||||
"The edit with your original pixels in the untouched parts.",
|
||||
"White where the edit changed the picture.",
|
||||
"How much of the picture was kept. Plug it into Show Text 🆎 to read it.",
|
||||
)
|
||||
FUNCTION = "restore"
|
||||
|
||||
def restore(self, image, edit_lock, feather, remove_tint=True):
|
||||
if image.ndim != 4:
|
||||
raise ValueError("Edit Lock Restore: image must be an IMAGE batch.")
|
||||
if not isinstance(edit_lock, EditLock):
|
||||
raise ValueError("Edit Lock Restore: edit_lock needs the output of Edit Lock.")
|
||||
height, width = int(image.shape[1]), int(image.shape[2])
|
||||
report = describe(edit_lock)
|
||||
if remove_tint and edit_lock.source is not None:
|
||||
# before the paste, so the edit itself loses the tint too and meets the original evenly
|
||||
image, taken = take_out_tint(image, edit_lock.source)
|
||||
if taken is not None:
|
||||
report += " " + tint_report(taken)
|
||||
mask = edit_lock.pixel_mask(height, width)
|
||||
if mask is None or edit_lock.source is None:
|
||||
return (image, torch.ones((int(image.shape[0]), height, width)), report)
|
||||
result, edit_mask = restore_pixels(image, edit_lock.source, mask, int(feather))
|
||||
return (result, edit_mask, report)
|
||||
|
||||
|
||||
def tint_report(shift):
|
||||
"""One plain sentence; the numbers are Lab units, for anyone who wants them."""
|
||||
light, a, b = shift
|
||||
return (f"It also took out the faint pink tint the model added to the whole picture "
|
||||
f"(L {light:+.1f}, a {a:+.1f}, b {b:+.1f}).")
|
||||
|
||||
|
||||
NODE_CLASS_MAPPINGS = {
|
||||
"AUSBOSS_NODES_EditLock": AusBossEditLock,
|
||||
"AUSBOSS_NODES_EditLockRestore": AusBossEditLockRestore,
|
||||
}
|
||||
|
||||
NODE_DISPLAY_NAME_MAPPINGS = {
|
||||
"AUSBOSS_NODES_EditLock": "Edit Lock 🆎",
|
||||
"AUSBOSS_NODES_EditLockRestore": "Edit Lock Restore 🆎",
|
||||
}
|
||||
@@ -0,0 +1,87 @@
|
||||
"""FlashVSR Upscale 🆎."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
from ._flashvsr_helpers import EXPECTED_MODEL_NAME, list_flashvsr_models, resolve_flashvsr_model, run_flashvsr
|
||||
|
||||
|
||||
class AusBossFlashVSRUpscale:
|
||||
CATEGORY = "🆎 AusBoss/Video"
|
||||
DESCRIPTION = (
|
||||
"Makes a video larger and sharper with the FlashVSR model. Pick the "
|
||||
"model file, set how much larger, and run. It reads the clip a few "
|
||||
"frames at a time, so a long clip needs no more graphics memory than "
|
||||
"a short one, and every frame comes back. Connect the Wan 2.1 VAE "
|
||||
"for a slower, finer finish."
|
||||
)
|
||||
SEARCH_ALIASES = [
|
||||
"flashvsr",
|
||||
"video upscale",
|
||||
"upscale video",
|
||||
"video super resolution",
|
||||
"vsr",
|
||||
"sharpen video",
|
||||
"ausboss",
|
||||
]
|
||||
|
||||
@classmethod
|
||||
def INPUT_TYPES(cls):
|
||||
models = list_flashvsr_models()
|
||||
return {
|
||||
"required": {
|
||||
"frames": (
|
||||
"IMAGE",
|
||||
{"tooltip": "The video frames to make larger."},
|
||||
),
|
||||
"model": (
|
||||
models,
|
||||
{
|
||||
"default": EXPECTED_MODEL_NAME if EXPECTED_MODEL_NAME in models else models[0],
|
||||
"tooltip": "The FlashVSR model file, from ComfyUI/models/diffusion_models.",
|
||||
},
|
||||
),
|
||||
"scale": (
|
||||
"FLOAT",
|
||||
{
|
||||
"default": 2.0,
|
||||
"min": 1.0,
|
||||
"max": 4.0,
|
||||
"step": 0.25,
|
||||
"tooltip": "How many times larger each side gets. 1 keeps the size and only sharpens.",
|
||||
},
|
||||
),
|
||||
"color_match": (
|
||||
"BOOLEAN",
|
||||
{
|
||||
"default": True,
|
||||
"tooltip": "Keeps each frame's colours the same as your video. Leave it on.",
|
||||
},
|
||||
),
|
||||
},
|
||||
"optional": {
|
||||
"vae": (
|
||||
"VAE",
|
||||
{
|
||||
"tooltip": (
|
||||
"Optional. Connect the Wan 2.1 VAE for a slower, finer finish. "
|
||||
"Left empty, the fast decoder inside the model file is used."
|
||||
)
|
||||
},
|
||||
),
|
||||
},
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("IMAGE",)
|
||||
RETURN_NAMES = ("frames",)
|
||||
OUTPUT_TOOLTIPS = ("The larger, sharper frames: as many as went in.",)
|
||||
FUNCTION = "upscale"
|
||||
|
||||
def upscale(self, frames, model, scale, color_match, vae=None):
|
||||
output = run_flashvsr(resolve_flashvsr_model(model), frames, float(scale), bool(color_match), vae)
|
||||
return (output,)
|
||||
|
||||
|
||||
NODE_CLASS_MAPPINGS = {"AUSBOSS_NODES_FlashVSRUpscale": AusBossFlashVSRUpscale}
|
||||
NODE_DISPLAY_NAME_MAPPINGS = {"AUSBOSS_NODES_FlashVSRUpscale": "FlashVSR Upscale 🆎"}
|
||||
|
||||
__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS"]
|
||||
@@ -164,7 +164,7 @@ class AusBossH3PromptWriter:
|
||||
"INT",
|
||||
{
|
||||
"default": 8,
|
||||
"min": 5,
|
||||
"min": 4,
|
||||
"max": 15,
|
||||
"step": 1,
|
||||
"tooltip": "How long the clip is. Plan on about one moment a second at most.",
|
||||
@@ -254,8 +254,8 @@ class AusBossH3PromptWriter:
|
||||
"default": VIDEO_JOBS[0],
|
||||
"tooltip": (
|
||||
"What the video is for: copy its moves, swap the person in it, keep it going, give it "
|
||||
"a new voice, or change something in it and keep the person (a hat, a shirt). Swap, "
|
||||
"new voice and change something ignore Seconds and keep the video's length."
|
||||
"a new voice, or change something in it and keep the person (a hat, a shirt). Copy its "
|
||||
"moves, swap, new voice and change something ignore Seconds and keep the video's length."
|
||||
),
|
||||
},
|
||||
),
|
||||
@@ -393,7 +393,7 @@ def _register_routes():
|
||||
async def read(request):
|
||||
try:
|
||||
body = await request.json()
|
||||
seconds = min(max(int(body.get("seconds", 8)), 5), LONGEST_SECONDS)
|
||||
seconds = min(max(int(body.get("seconds", 8)), 4), LONGEST_SECONDS)
|
||||
jobs = {int(slot): job for slot, job in dict(body.get("jobs") or {}).items() if job in JOBS and int(slot) in PICTURES}
|
||||
video = body.get("video") if body.get("video") in VIDEO_JOBS else ""
|
||||
return web.json_response(script_report(str(body.get("script", ""))[:LONGEST_SCRIPT], seconds, jobs, video))
|
||||
|
||||
@@ -168,7 +168,7 @@ class AusBossImageCropRotatePad:
|
||||
try:
|
||||
path = resolve_input_path(image)
|
||||
except Exception:
|
||||
path = image or ""
|
||||
path = None # refused: hashed by its text only, never opened
|
||||
spec = spec_from_values(**values)
|
||||
# Every widget value is already part of ComfyUI's cache key, so a
|
||||
# changed transform or budget re-runs the node without this; the
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user