main
38
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 malkuthroandCursor
|
cf32beb8b0 |
fix(install-guard): skip duplicate scan in linked git worktrees (#281) (#282)
* fix(install-guard): skip duplicate scan in linked git worktrees Linked worktrees outside custom_nodes/ are not ComfyUI installs; stop treating sibling worktrees as #162 duplicates so pytest can register nodes. Closes #281. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(install-guard): keep #162 active for custom_nodes→worktree symlinks Use the unresolved load path for the worktree skip so Path.resolve() cannot drop a custom_nodes parent and disable the duplicate scan. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
|
040b30733d |
feat(matte): first-party mask-guided one-step video matting nodes (#269)
* feat(matte): first-party mask-guided one-step video matting nodes
Adds matte/ — an independent implementation of the published VideoMaMa
one-step SVD method (matte/pipeline.py) plus Koolook's temporal windowing,
subject-focused crop/stitch, resolution control (/64, shortest-side) and
auto OOM resolution-backoff. Nodes: KoolookMatteLoader, KoolookMatteSampler,
KoolookMatteFocusCrop, KoolookMatteFocusStitch.
Registered via the existing guarded _merge_node_group("Koolook Matte",
".matte"), so a missing diffusers disables only these nodes. diffusers is
already pulled in by the radiance/whatdreamscost forks, so no new dependency.
Loads (never bundles) the VideoMaMa UNet (CC-BY-NC) + SVD VAE at runtime;
matte/NOTICE records the model licenses and user responsibility.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(dev-sync): include matte/ in RUNTIME_PATHS
So `sync_to_dev` copies the new matte node package into the live ComfyUI
install alongside the rest of the koolook pack.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(matte): raw cond latents; fp16 default; preserve checkpoints on dev-sync
pipeline.py: _encode_latents must NOT divide by vae scaling_factor. The
VideoMaMa UNet is trained on raw VAE latents (matching okdalto's
VideoInferencePipeline, which multiplies by scaling_factor then divides it
back out -> net x1). Dividing over-scaled the conditioning ~5.5x and produced
a systematically blurry matte on every frame. Verified 5/5 vs the desktop /
vanilla reference on identical single-frame inputs.
nodes.py: harden the loader precision fallback so fp16 is the true default -
only an explicit "bf16" opts out; a stale "auto" widget now resolves to fp16
instead of bf16.
sync_to_dev.py: the rmtree+copytree mirror deleted matte/checkpoints/ (the
gitignored ~10GB SVD+VideoMaMa weights) on every matte sync, forcing a
re-download. Add PRESERVE_IN_DEST + selective _clean_dest so runtime-
downloaded model data survives a sync.
docs: add docs/maintainers/matte-pipeline.md (latent-scaling invariant,
black-output causes, fidelity levers, single-frame + vanilla-oracle debug
loop, checkpoint preservation).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(matte): address PR #269 review
- license: align matte to the repo's GPL-3.0 (was self-contradictory: claimed
"MIT licensed - see LICENSE" while LICENSE is GPL-3.0). pipeline.py header,
NOTICE, and the maintainer doc updated. NOTICE still credits VideoMaMa
(CC BY-NC) + SVD (Stability Community License).
- nodes.py _ensure_2d_mask: squeeze first, then handle [H,W,C]/[C,H,W]
explicitly and RAISE on ambiguous shapes -- previously [H,W,3] silently
became [W,3] (wrong geometry).
- nodes.py OOM backoff: catch RuntimeError (OutOfMemoryError subclasses it, but
cuBLAS/cuDNN/older paths raise plain "RuntimeError: CUDA out of memory"); the
existing message guard still filters non-OOM errors.
- nodes.py: warn when the guide mask is empty instead of silently falling back
to full-frame.
- sync_to_dev.py: PRESERVE_IN_DEST is now an exact relative path
("matte/checkpoints"), not a bare name matched at any depth; handle symlinks
before is_dir() in _clean_dest so a symlinked dir isn't recursed/rmtree'd.
- CHANGELOG: add the Koolook Matte nodes under [Unreleased] with the
non-commercial model-license caveat.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(matte): address PR #269 re-review
- sync_to_dev.py [BLOCKING]: a top-level symlinked dest (e.g. target/matte -> X)
was passed straight into _clean_dest()/copytree(), which follow the link and
can mutate files outside the target. Now unlink a symlinked dest and own the
path as a real dir (checked before exists(), which follows symlinks). Verified
with a repro: the symlink target's file survives.
- nodes.py: force float32 on a CPU fallback (CUDA unavailable). Half precision
is unreliable on CPU; fp16 kernels would error. Warn that CPU is very slow.
- docs: correct the preserved-path guidance to the target-relative
"matte/checkpoints" (was a bare "checkpoints" name); reword "near-verbatim
reimplementation" -> "independent implementation targeting behavioural parity"
to match the NOTICE/pipeline framing.
- nodes.py: also warn on a degenerate (<8px) guide-mask region before the
full-frame fallback (was silent).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(matte): address PR #269 re-review (CI red + FocusCrop frame check)
- nodes.py: KoolookMatteFocusCrop now validates images vs coarse_mask frame
counts (raises like KoolookMatteSampler.run_inference). A mismatch previously
made cropped_images/cropped_mask diverge and silently drop/misalign frames at
FocusStitch.
- pipeline.py: split "vae.to('cpu'); _clear_cache(...)" onto two lines (Ruff
E702 was failing Lint & Smoke).
- constraints-test.txt: bump pillow 12.2.0 -> 12.3.0 (pip-audit CVE, fixed in
12.3.0). Pillow is a leaf dep so the closure is unaffected; CI re-audits.
- README: note the Koolook Matte nodes need diffusers (already via the LTX /
Radiance forks) and huggingface_hub for runtime weight downloads.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: AI Machine <ai.machine@x.baconcph.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
|
||
|
|
961a95b9bd |
feat(ltx-director): reference-image identity — Ghost Mask + schedulable guide attention (#264)
* feat(ltx-director): Ghost Mask reference image (REF / character sheet) Add a no-LoRA reference-image identity anchor to LTXDirector__koolook, the way a WAN VACE reference works: attach reference image(s) and the model keeps the character's face / mouth consistent even when keyframes show a closed (or different) mouth than the audio-driven motion needs. Targets the mouth / identity drift seen in the talking-goat lip-sync workflow. Mechanism (Ghost Mask): the reference image(s) are appended as masked guide frames *after* the clean video region, so the transformer attends to them for identity but they never appear in the output; the trailing reference frames are then sliced off the result. - LTXDirector__koolook: new optional `reference_images` (IMAGE) + `reference_strength` inputs, and appended `clean_latent_frames` / `clean_pixel_frames` outputs. All additive/optional -> saved workflows unaffected. Refs are appended AFTER snap/ease so they are never grid-snapped or ease-expanded; the auto latent is grown to hold them. ltx_director.py is under ruff's forks/ exclude. - reference_ghost.py (new, Koolook-original, pure stdlib): trailing-reference geometry (insert frames, grown latent length, slice bounds). Unit-tested. - k_clean_latent_slice.py (new root node `CleanLatentSlice`): torch.narrow slice to drop the trailing reference frames (start=0, length=clean_latent_frames). torch imported lazily; clamp math unit-tested, slice path tested when real torch is present (skips under the CI stub). - Registered the node; added to sync_to_dev RUNTIME_PATHS. - Attribution: the Ghost Mask APPROACH is adapted (idea only, reimplemented) from CGlide/WhatDreamsCost-CSGlide (main_cs @ 21dd076, GPL-3.0, license verified 2026-06-24, GPL->GPL OK). THIRD_PARTY.md, forks_manifest.yaml, UPSTREAM_PIN.yaml, CHANGELOG updated. Phase 1: Ghost Mask only. The LoRA-backed Licon MSR prefix mode and the editor's character-slot UI are intentionally NOT ported. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(sidebar): action-led star footer ("Give a ★ on GitHub") Reword the GitHub discovery footer from the passive "Help creators find us on GitHub" to an action-led "Help creators find us! Give a ★ on GitHub", with the star as the explicit ask. Render the sentence as one inline text flow (not separate flex children) so the footer's `gap` doesn't space out the words; the ★ glyph keeps its gold styling inline and now carries role="img"/aria-label="star" so it is announced. Verified rendering in an isolated harness. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ltx-director): Ghost Mask broke motion in the append-and-crop pipeline The upstream guide pipeline (LTXDirectorGuide + LTXDirectorCropGuides) appends every guide_data image as an extra latent frame and crops them back off via `exact_crop_frames = latent.shape[2] - initial_latent_length`. The Ghost Mask code (modeled on CGlide's CleanLatentSlice flow) ALSO pre-grew the video latent by n_refs, so initial_latent_length became clean+n_refs: CropGuides then left n_refs frames uncropped AND the audio latent stayed at the clean length, desyncing the A/V joint transformer and breaking motion. Fix: don't pre-grow the latent. References are added to guide_data exactly like timeline keyframes and ride the same append-and-crop path — the guide node appends them and CropGuides removes them, audio untouched. Auto latent is sized to the clean length again. CleanLatentSlice/clean_latent_frames remain available for bare pipelines that don't use LTXDirectorCropGuides. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(ltx): LTX Guide Reference Strength — per-stage reference boost Adds a small Koolook node that rewrites the strength of ONLY the Ghost Mask reference frames in guide_data (the trailing entries the Director now tags with `reference_count`), leaving the keyframe pins untouched. Insert it before a stage-2 refinement guide and set ~0.9 to lock identity during the low-denoise pass without disturbing the motion set in stage 1 — the deterministic form of "more reference in stage 2". Chosen over vendoring the 600-line upstream LTXDirectorGuide (which would force a node-ID rename + graph rewire): this is one small node inserted on the stage-2 guide_data wire. The sigma-ramp alternative (ramp reference up at low sigma) is deferred — the reference hook is noise_mask + a core-model attention term, so it needs a live-model spike, not a clean static patch. - Director tags guide_data["reference_count"] = n_refs. - k_ltx_guide_reference_strength.py: pure strength rewrite (absolute / multiply, clamped), unit-tested; registered; added to sync paths. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(ltx): LTX Reference Bind Schedule — sigma-ramp the reference attention Confirmed against core (comfy/ldm/lightricks/model.py): the reference's guide attention strength is turned into a log-space additive self-attention bias in `_build_guide_self_attention_mask`, which is rebuilt EVERY denoise step and explicitly supports strength > 1.0 to amplify. That's a clean per-step hook — so the "inverse sigma curve" idea is buildable, not just a stage knob. New node mirrors LTXAVBindSchedule: a model_function_wrapper captures the per-step sigma and stashes a gain in transformer_options; an object-patch on diffusion_model._build_guide_self_attention_mask multiplies the strength of ONLY the trailing `num_references` guide entries (references are appended after keyframes) by a smoothstep ramp — neutral early (motion forms), rising to peak_strength late (identity locks during refinement). Keyframe pins and noise-mask freezing untouched; orthogonal to reference_strength. One pass, no re-noise, no LoRA. - ref_gain ramp + _boosted_entries unit-tested (11 tests); model wiring validated by live render. - Registered; added to sync paths. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ltx): mark Reference Bind Schedule inert without an IC-LoRA Live diagnostic on the AV model showed the builder is reached and the sigma gain flows in, but resolved_guide_entries=None despite 3276 guide tokens. Root cause: LTXDirectorGuide only records guide_attention_entries when an IC-LoRA is active (`if is_lora_active:` gates _append_guide_attention_entry). With no LoRA there is no per-guide attention bias to scale, so the attention-ramp lever does not exist in the Ghost Mask path — it can only attenuate/amplify entries that an IC-LoRA created. Rather than silently no-op, the node now logs a clear INERT warning when the entry list is empty, and the docstring/DESCRIPTION/CHANGELOG state the IC-LoRA requirement. The node remains correct for the IC-LoRA path (its partner). No-LoRA reference strengthening stays on the noise-mask lever (reference_strength / LTX Guide Reference Strength). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(ltx): Reference Bind Schedule works without an IC-LoRA (fabricates entries) The attention-bias identity channel (_build_guide_self_attention_mask) is base-model functionality, but LTXDirectorGuide only populates the per-guide entries when an IC-LoRA is active, so the node was inert in the no-LoRA Ghost Mask path. Now, when entries are missing, the node FABRICATES them for the trailing reference tokens: it splits num_guide_tokens across (num_keyframes + num_references) guide frames and assigns the sigma-ramped strength to the reference share, then hands off to the core builder. This gives the reference the same amplified-attention channel an IC-LoRA would use — inside the Director, no guide-node vendor, no graph rewire (the node is already in the user's setup). When an IC-LoRA IS active it scales the real entries as before. - New `fabricate_reference_entries(num_guide_tokens, num_references, num_keyframes, strength)` (pure, unit-tested) + `num_keyframes` input. - Wrapper now does scale-or-fabricate; clearer one-shot diagnostics. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ltx): keep num_keyframes last so pre-existing RefBind widgets stay aligned num_keyframes was inserted mid widget-order, which shifted ramp_start/ramp_end for workflows saved before it existed. Move it to the end of the optional inputs so saved [num_references, peak_strength, ramp_start, ramp_end] values keep mapping correctly; num_keyframes just defaults to 0 on reload. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(test): drop unused pytest import (ruff F401) The guide-reference-strength test uses plain asserts; the stray `import pytest` tripped ruff F401 and reddened the Lint & Smoke CI job. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ltx-director): default keyframe guidance 0.8 (smooth), not 1.0 The timeline editor auto-built the guide_strength string from each clip's per-segment guideStrength, filling 1.0 for any clip left unset (and the Python fallback matched). 1.0 is a hard pin, so a clip you forgot to lower snapped instead of transitioning smoothly. Unset clips now default to 0.8 in the editor (display + serialization) and the node; explicit per-clip values still override. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(ltx): Keyframe Soften Schedule — smooth motion early, sharp poses late New model-patcher node LTXKeyframeSoftenSchedule: sigma-schedules the LTX keyframe denoise-mask pins so the model forms smooth motion between poses early (pins loosened at high sigma) and resolves sharp poses late (pins restored by a crossover). Uses the sampler's per-step denoise_mask_function hook — no IC-LoRA required — and adds no guide frames (so it can't fabricate extra poses). Audio-safe: only the video half of the NestedTensor A/V mask is softened; the audio mask is untouched, so lip-sync is unaffected. Two normalized 0-1 widgets map onto the sensitive useful range: max_soften (ease, ->0-0.20) and crossover (invention window, ->0-0.05). Pure soften_gain ramp unit-tested (11 tests); model wiring render-validated. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(review): chain denoise hook, optional_latent clean-length, audio test, CGlide manifest Addresses PR #264 review findings: - MEDIUM (Soften node): chain a pre-existing denoise_mask_function instead of clobbering it (mirrors the sibling LTX patchers' wrapper-chaining), so the node composes with anything else that sets that hook. - HIGH #2 (Director): clean_pixel_frames/clean_latent_frames now derive from the ACTUAL optional_latent shape when one is connected, so Ghost-Mask reference geometry and the reported counts follow the real latent instead of the timeline duration (with a safe fallback + warning). - LOW (Soften node): reconcile docstring (crossover is sigma-progress, maps to internal 0-0.05), one-time warn on the sigma-read fallback, and extract a testable soften_denoise_mask() helper. - LOW: lock the audio-safety guarantee with NestedTensor-stub tests — video mask softened, audio mask untouched, part-ordering asserted. - #6: add a formal forks_manifest.yaml entry for the CSGlide Ghost Mask idea-only adaptation (was prose-only in another entry's notes; THIRD_PARTY.md already carries the full attribution). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> --------- Co-authored-by: AI Machine <ai.machine@x.baconcph.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> |
||
|
|
ba37ec3d41 |
feat: LTX Director 2.0.2 fork, keyframe grid-snap, and audio→video bind schedule (#259)
* feat(forks): LTX Director v2_0_2 fork + keyframe latent-grid snap (#258) Bump the WhatDreamsCost Koolook fork to upstream 2.0.2 (pinned fe09f73) in a new versions/v2_0_2 namespace, and add `snap_keyframes_to_grid` (default on) to fix hard-pinned image keyframes "jumping" at latent-bucket edges. Why: a Director image keyframe is injected as a single latent frame at the segment's left edge, quantized to the model temporal stride s (L(f)=0 if f==0 else 1+(f-1)//s). Pins near a bucket boundary round ambiguously and two pins in one bucket collide. Snapping each pin to its bucket center removes the ambiguity without changing which latent frame it lands on; colliding pins are bumped to the next free bucket with a warning. - forks/whatdreamscost_koolook/versions/v2_0_2/: vendored 2.0.2 sources with the v1.3.9 Koolook deltas carried forward (relay_overrides, audio_transcript_json, per-segment sigma — build_segments is byte-identical 1.3.9->2.0.2 so the sigma patch drops in verbatim), plus new pure keyframe_grid.py snapper wired into the insert_frames build. - Canonical LTXDirector__koolook rolls to 2.0.2 (entrypoint loads v2_0_2 last); LTXDirector__koolook_v1_3_2 alias preserved; v1.3.9 kept on disk. - LICENSE re-verified GPL-3.0 (2026-06-22). Manifest + THIRD_PARTY + CHANGELOG + audio-lipsync findings/CHEATSHEET updated. - 37 new unit tests for the snapper; 107 fork tests pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(forks): append Koolook LTX Director widgets last for clean load Move the v2_0_2 Koolook widgets (relay_overrides, audio_transcript_json, snap_keyframes_to_grid) to the end of the schema, after override_audio. They were inserted after img_compression, which shifted the last upstream widget when an upstream-2.0.x-saved workflow is loaded onto LTXDirector__koolook. With them last, an upstream graph's positional widgets_values map onto the upstream widgets untouched and only the Koolook widgets append with defaults (snap_keyframes_to_grid=on). Execution is unaffected (kwargs by name). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * refactor(forks): make v2_0_2 a clean upstream-2.0.2 replica + grid-snap only Per maintainer direction: the v2_0_2 Koolook LTX Director should faithfully replicate upstream 2.0.2 (so all latest features work) rather than carry the older customizations, which are unused. - Drop relay_overrides, audio_transcript_json, and the per-segment sigma patch. prompt_relay.py and patches.py are now vendored VERBATIM from upstream 2.0.2; _relay_overrides.py removed. - ltx_director.py is upstream 2.0.2 verbatim except two additive changes: the namespaced node ID LTXDirector__koolook, and the snap_keyframes_to_grid latent-bucket keyframe snap (issue #258), appended last in the schema. - v2_0_2 registers only LTXDirector__koolook; the legacy LTXDirector__koolook_v1_3_2 ID stays backed by the v1_3_9 package. v1.3.9 (with the older customizations) remains on disk for rollback. - Manifest / THIRD_PARTY / UPSTREAM_PIN / CHANGELOG updated to the clean scope. grid-snap unit tests unchanged (107 fork tests pass). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(forks): vendor upstream 2.0.2 timeline editor JS for LTXDirector__koolook The v2_0_2 Python was upstream-exact but the node still rendered with the old 1.3.9-era timeline editor (fewer toolbar actions, no IC-LoRA track, leaked *_second widgets, frame_rate NaN). Replace web/whatdreamscost_koolook/ ltx_director.js with upstream 2.0.2's js/ltx_director.js vendored verbatim, retargeted with only two edits: unique extension name 'LTXDirector_Koolook' and node guard nodeData.name === 'LTXDirector__koolook'. UX now matches upstream 2.0.2 exactly; only the Python differs (grid-snap). The old koolook base64-scrub deltas are dropped as obsolete — upstream 2.0.2 persists timeline media via imageFile/audioFile + /api/view URLs (saved timelines are already lean), and the localStorage quota guard stays in the global web/koolook_draft_guard.js (upstream does not override localStorage). Rewrote tests/sidebar/test_ltx_director_draft_guard.py to the new contract (no embedded guard, retargeted to the koolook node, pinned-upstream provenance). 110 fork+contract tests pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(scripts): resolve .env from main repo when dev-sync runs in a worktree Unrelated to the LTX work: plain `dev-sync` failed with "KOLOOK_COMFYUI_DEV_PATH not set" from any worktree because scripts/sync_to_dev.py only loaded `.env` from the worktree root, where the committed `.env` does not exist (it lives in the main checkout). sync_to_dev_audio.py already had the worktree->main fallback, which is why dev-sync-audio worked. Move that fallback into a shared sync_to_dev.find_dotenv() and use it in both scripts so they can't drift. Adds regression tests for the worktree fallback, the worktree-own-.env preference, and the not-a-worktree case. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * chore: gitignore root-level generated workflow JSONs Node-swapped/converted workflow copies (e.g. *.koolook.json) are dropped at the repo root during dev iteration; they're scratch deliverables, not package content. Ignore root-level *.json only (leading `/`), with a `!config.json` exception, so the 30+ tracked JSON in subdirs (test fixtures, web/, .claude/, docs/runs/*/workflow.json) stay tracked. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat: LTX A/V Bind Schedule node — gate audio→video binding by sigma New Koolook node LTXAVBindSchedule decouples big motion from lip-sync in LTX 2.3. The AV transformer (comfy/ldm/lightricks/av_model.py, BasicAVTransformerBlock) adds audio into video via a gated `audio_to_video_attn` per block, run only when transformer_options["a2v_cross_attn"] is set. This node patches each block's audio_to_video_attn.forward to scale its output by a per-step gain, and a unet-function-wrapper computes that gain from the current sigma: early/high-sigma steps run near `early_gain` (coarse motion settles audio-blind), ramping (smoothstep) to full at low sigma (lips bind onto settled motion). One pass, no re-noise — the fix for keyframe transitions jumping on audio peaks. early_gain=1.0 reproduces stock behavior; a graph without the node is unchanged. Registered as a root node + added to RUNTIME_PATHS so dev-sync ships it. Pure ramp (a2v_gain) unit-tested (9 cases); model wiring validated by live render. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat: LTX A/V Bind Schedule — add transition-window mode (lips synced + smooth) The global sigma ramp decoupled motion from lips but desynced the mouth (lip timing is coarse motion too, so holding audio out globally moved the mouth audio-blind). Add a WINDOWED mode: audio binds at full strength everywhere (lips sync across the whole clip) EXCEPT a +/-transition_window band around each keyframe transition, where the sigma ramp applies — so only those moments are motion-protected early and rejoined late. Implemented as a per-latent-frame gain on a2v_out: read grid_sizes from transformer_options to map frame-major video tokens -> latent frames, build a gain vector (1.0 outside bands, smoothstep-tapered to sigma_gain at each transition center), expand per-frame -> per-token, scale. New optional inputs: transition_frames, total_frames, transition_window. Empty transition_frames = prior global-ramp behavior. windowed_frame_gains unit-tested (7 cases). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat: LTXAVBindSchedule — one-shot diagnostic log (shapes/grid/mode) Logs out.shape, grid_sizes, transition_fracs, and whether the per-frame mask fired vs fell back to scalar — so a single render reveals whether the a2v patch hooks the model and whether the windowed token->frame mapping aligns, instead of guessing blind. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs(audio-lipsync): record validated a2v bind-schedule knee (0.35/0.25/0.55) Promote the finding that LTX 2.3's audio->video coupling must be sigma-ramped (not full-on) to keep motion smooth AND lips synced, with the validated knee early_gain=0.35 / bind_start=0.25 / bind_end=0.55 on the 8-step base pass, plus the trade-axis table and the per-frame-mode caveat (grid_sizes None + appended guide frames -> scalar fallback). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * feat(forks): keyframe ease — strength-ramped neighbor pins to de-robotic transitions Adds `keyframe_ease` / `ease_falloff` to the Koolook LTX Director (default off). Each hard image pin keeps its exact center pose, but emits same-pose neighbor pins one latent bucket apart at decreasing strength (center*falloff**k), so the model glides into and out of the locked pose instead of snapping to a dead stop — softening the "stop/dissolve/stop" robotic feel without softening the pose. Implemented as a pure list expansion (`expand_keyframe_ease` in keyframe_grid.py, 6 unit tests) applied to guide_data after grid-snap, so the upstream guide node is untouched (no fork). Center stays exact; neighbors that fall off-clip or onto another center are skipped. Costs extra guide frames (more compute), hence opt-in. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs(audio-lipsync): downgrade a2v knee to 'promising' — note lows-placement confound The 0.35/0.25/0.55 win was measured on a clip whose transitions sit on audio lows, so 'no jumps' may be the placement, not the node. Mark the controlled on-peak test (early_gain 1.0 vs 0.35 on the same on-peak layout) as still pending. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * chore: mark vendored WhatDreamsCost fork files linguist-vendored The koolook Director ships upstream 2.0.2 code (the 8.5k-line timeline-editor JS + verbatim prompt_relay/patches) so the pack runs and distributes, but it's upstream, not authored here. Mark those linguist-vendored (drop from language stats) and the JS -diff (collapse the large blob in PR diffs). Our edited/new files stay attributed to us. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ltx-av): drop unused `import math` (ruff F401) a2v_gain uses an explicit smoothstep polynomial, not math.*, so the import was dead. Unblocks the Lint & Smoke CI gate. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * ci(codeql): advanced setup that excludes vendored upstream code CodeQL default setup cannot read a config file, so it scanned the verbatim-vendored WhatDreamsCost editor (web/whatdreamscost_koolook/ ltx_director.js) and raised 4 high-sev DOM-XSS alerts in upstream code we do not own. Switch to advanced setup with a committed config whose paths-ignore mirrors the .gitattributes linguist-vendored block, so the scanner treats vendored upstream as out-of-scope while still fully scanning our own nodes and sidebar JS. Languages match the prior default-setup matrix (actions, javascript-typescript, python). NOTE: requires disabling "default setup" once in repo Settings before this workflow can run (GitHub rejects advanced runs while default setup is enabled). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * fix(ci): scope vendored fork exclusions to v2_0_2 * fix(forks): preserve final LTX keyframe snap bucket --------- Co-authored-by: AI Machine <ai.machine@x.baconcph.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> |
||
|
malkuthro
|
3e2b541a61 | fix(registry): expose nodes to scanners (#251) | ||
|
malkuthro
|
83b170bdae |
feat(setups): add publish contract nodes (#228)
* feat(setups): infer app contract from named fields * feat(setups): add publish contract nodes * feat(setups): split publish result contract * fix(setups): resolve publish output reroutes in preview * fix(setups): validate app surface on read * fix(setups): harden app contract validation |
||
|
malkuthro
|
8c8546c17c |
Add queue-driven loop status controller (#205)
* Add queue-driven loop status node * Document ComfyUI server API testing * Harden loop queue controller * Guard loop controller HTTP requests |
||
|
malkuthroandClaude Opus 4.8
|
447465ecd8 |
refactor(nodes): remove superseded LTX Director extra nodes + harden registry (#198) (#199)
* refactor(nodes): remove superseded LTX Director extra nodes + harden registry (#198) Remove the two experimental nodes layered on the WhatDreamsCost fork — KoolookAudioTranscriptTimeline (k_audio_timeline.py) and KoolookTimelineEditor (k_timeline_editor.py) — plus the transcribe_audio_timeline.py CLI and its test. The audio-transcript to Director-timing job now lives inside the fork via the Director's audio_transcript_json input, so the standalone nodes are redundant. Also drops the orphaned [audio]/faster-whisper extra and updates packaging, dev-sync, web, and docs references. Harden the node registry against the #183 failure: __init__.py now imports each node group independently and installs the /koolook/presets/* routes regardless of node-import results, so one missing or broken module degrades to 'that node absent' instead of zeroing every node and 404'ing the sidebar disk routes. A new test keeps every root k_*.py listed in dev-sync's RUNTIME_PATHS. Closes #198. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(nodes): ship loader gates via dev-sync-audio; generalize RUNTIME_PATHS guard (#198) Address the K-Flame review on #199: MEDIUM — the #198 refactor made koolook_versioning.py a single registration gate, but dev-sync-audio shipped the new __init__.py without it (and the k_*.py-glob guard test couldn't see it). Add koolook_versioning.py AND koolook_install_guard.py (whose absolute-import fallback raises uncaught) to AUDIO_PATHS so a scoped sync can never leave the loader pointing at a module the target lacks. Generalize the guard test: parse __init__.py's actual load-time import targets (relative imports + _merge_node_group calls) and assert each is in RUNTIME_PATHS, instead of globbing k_*.py which missed the non-k_ gates. Add an AUDIO_PATHS gate-coverage test. LOW — CHANGELOG now records the [audio]/faster-whisper + transcribe_audio_timeline.py removal as an external-contract break. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
|
8058e17456 |
fix(snapshot): paired drift-corruption fix — duplicate-install guard + boot-time mismatch (#161, #162) (#186)
* fix(snapshot): prevent drift corruption from duplicate installs + boot-time mismatch (#161, #162) Closes #162 and #161 as the paired fix the issues describe — both touch the same data-loss surface and verify against the same dev-sync rig. installs at import time, log critical, register nothing in the alphabetical loser; client-side `window` sentinel as belt-and-suspenders; align .env.example / dev-iteration-loop / sync_to_dev / README at the Registry-derived `koolook/` path. compare against live /userdata + picks, set a drift flag on mismatch; gate periodic + pre-load autosave routing to `_unsaved_autosave/` while drifted; promote status pill to "drifted" precedence over "saved" with a recovery tooltip; auto-clear on the next markStateSaved (Save / Load). 26 new tests (14 install guard, 4 sidebar sentinel, 8 boot drift) — 316 total passing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(snapshot): apply review fixes — canonical fingerprint + early loser exit + drift timestamp Review-driven follow-ups on PR #186: 1. **Symmetrize file-side fingerprint (MEDIUM).** `detectBootDrift` now uses a new `_canonicalStateFingerprint()` that normalizes the workflows tree through `normalizeWorkflowsStore` AND deep-sorts keys via `sortJsonValue`. Without this, hand-edited / cross-machine / pre-v0.3 files would false-positive as drifted on first boot (e.g. `archived: "false"` string vs `false` bool, missing `directories: {}`, key insertion-order asymmetry from the `{...wf, archived, module, tags}` spread). `_computeFingerprint` stays unchanged so the localStorage save baseline contract holds for existing users. New regression test verifies a shape-only diff does NOT trigger drift. 2. **Gate fork imports on `_is_winning_install` (LOW).** Non-winning duplicate installs now short-circuit before the heavy torch / PIL / diffusers / aiohttp imports, exiting in ~5 LoC instead of paying full import cost just to discard the mappings. 3. **Drift timestamp uses `driftDetectedAt`, not stale `lastNamedSaveAt` (LOW).** The localStorage save baseline could be from a prior session whose live state was already corrupt — surfacing it as "Last named save" in the drift tooltip would mislead. Replaced with the session-noticed-at timestamp from `_bootDriftDiagnostics`. 352 tests passing (316 prior + 26 from PR #186 + 1 new normalize- symmetry case). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(install-guard): fail safe on unreadable / non-UTF-8 siblings (#162) Re-review follow-up. The import-time duplicate-install scan could crash the entire plugin load (fail-silent — the alphabetical winner vanished too, not just the loser) on two unguarded paths: - detect_duplicate_koolook_installs: the per-entry is_dir()/is_file() over custom_nodes/ siblings was unguarded, so a non-traversable sibling (restrictive perms, broken mount, a special dir) raised PermissionError up into __init__.py. No duplicate even required. Now caught per-entry and skipped. - read_pyproject_version: read_text(encoding="utf-8") raises UnicodeDecodeError (a ValueError, NOT OSError) on a non-UTF-8 / binary pyproject.toml, escaping the `except OSError`. Now degrades to "?" via `except (OSError, ValueError)`. - __init__.py: wrapped the guard call site as a final backstop — any unforeseen guard failure now degrades to register-normally rather than aborting the import. Both paths reproduced before the fix; the first surfaced while running the suite from a /tmp checkout whose parent held a root-owned dir. 3 regression tests added (non-traversable sibling skipped; non-UTF-8 pyproject -> "?"; report survives a non-UTF-8 sibling). Full suite: 452 passed (449 + 3). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-authored-by: kflame99 <261674183+kflame99@users.noreply.github.com> |
||
|
malkuthro
|
1d9d40ab70 |
Add audio transcript timeline node (#189)
* Add audio transcript timeline node * Hook audio transcripts into Koolook Director * Add modular timeline audio timing workflow * fix(frontend): guard Comfy workflow draft cache * fix(audio): compose transcript timeline prompts * fix(audio): address transcript review findings * fix(ci): keep transcript tests lightweight * fix(timeline): restore image previews from saved paths * fix(timeline): scrub restored preview blobs |
||
|
malkuthroandClaude Opus 4.7
|
4739dd4203 |
feat: Easy_Utility, Easy_LoadVideo, and preview fixes (#184)
* feat(nodes): add Easy_Utility — INT → padded STRING dispatcher First mode `int_to_padded_string`: takes an INT widget (with up/down arrows) and emits a zero-padded STRING (default `001`, no `v` prefix). The downstream `koolook_versioning.resolve_version_token` prepends `v` and re-pads, so wiring this into `EasyAIPipeline.version` / `Easy_VideoCombine.version` yields the standard `vNNN` token at execution. Designed for extension: the `mode` widget shipped from day one so adding further conversions (string→int, formatting variants, etc.) doesn't rewire saved workflows. Note: the previously-attempted JS preview-resolver refactor that aimed to fix the "preview shows stale version" bug was reverted in this worktree before commit — it didn't actually fix the bug (the real cause is the JS resolver not descending into ComfyUI subgraphs) and it broke shot_name resolution in the process. The preview-resolver work is NOT in this commit; only the Python node + tests. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(nodes): guard version widget against pysssss .replace crash PR #180 added a STRING `version` widget to EasyAIPipeline and Easy_VideoCombine. Three workflow states leave widget.value undefined or non-string when graphToPrompt iterates widgets: - Pre-#180 workflows replay an INT (e.g. `1`) into the new STRING slot. - Widget-to-input conversion voids the value in newer LiteGraph. - Easy_VideoCombine's dict-based widgets_values omits the new key entirely on workflows saved before #180. With pysssss/ComfyUI-Custom-Scripts installed, presetText.js patches every STRING TextWidget.serializeValue to run callbacks that do `value.replace(...)` for `{variable}` substitution. Any of the states above crashes graphToPrompt and turns Run into a no-op for the whole workflow ("Cannot read properties of undefined (reading 'replace')"). Fix: trap `widget.value` reads/writes on the version widget via `Object.defineProperty` so the property is always coerced to a string regardless of caller. `configurable: true` so downstream code can still override the descriptor if it needs to. Tests: 231 passing; both JS files parse OK. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(ai-pipeline): resolve subgraph version preview * feat(video): add path-aware load video node * fix(ai-pipeline): preview connected text names * fix(pr184): address review cleanup --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
malkuthroandClaude Opus 4.7
|
9f063345f7 |
feat(investigations): LTX23 audio-file lip-sync scaffold (#177) (#178)
* feat(investigations): scaffold #177 LTX23 audio-file lip-sync investigation Reorganize docs/investigations/ into numbered topic folders, each self-contained with narrative + upstream pin + patches + per-run snapshots. Add a generic sync helper that mirrors the dev-sync pattern for any investigation's third-party node modifications. Structure (new): docs/investigations/ 00_LTX23-base-1-step/README.md (rename of ltx-director-4k-transitions.md) 01_LTX23-audio-file-lipsync/ README.md -- problem, mechanism, patches, reversion, promotion upstream.json -- repo + version + pinned commit + sync target env + files patches/ -- source of truth for modified upstream files ltx_director.py -- adds `relay_overrides` multiline JSON input prompt_relay.py -- per-segment sigma per Prompt-Relay paper formula runs/ LOOP.md -- iteration protocol log.md -- rolling table (empty header for now) Sync helper: scripts/sync_investigation_patches.py <NN>_<topic> [--dry-run] - Reads upstream.json for target env var + file list - Backs up the target's current state once per day (<file>.bak.<YYYYMMDD>) - Worktree-aware: falls back to MAIN's .env via git rev-parse --git-common-dir - Prints a two-line report matching scripts/sync_to_dev.py shape - Per-investigation trigger phrase via `trigger:` key (sync-audio for #177) Env: .env.example -- documents new KOLOOK_WHATDREAMSCOST_PATH .env -- local path set Glossary updates (docs/reference/glossary.md): + Investigation folder + JSON file folder + Loop + Run + Investigation patches + sync_investigation_patches.py Closes the workflow gap from #177: MAIN is now the source of truth for investigation patches; the live install is a sync destination only. No manual file copying between iterations — `sync-audio` does it. * chore(ruff): exclude docs/investigations from lint The investigation patches under docs/investigations/*/patches/ are vendored upstream source (WhatDreamsCost-ComfyUI v1.3.2) carried so sync_investigation_patches.py can mirror them into the live install and so iterations can be diff'd against the pinned upstream commit. Linting them would force divergence from upstream and defeat that purpose. Mirrors the existing forks/ exclude precedent. Fixes the CI Lint & Smoke failure on PR #178. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor: per-task automation modules + WhatDreamsCost Koolook fork Pivot docs/automations/ from "one folder per model" to "one folder per automation module" so the same diffusion model can host several independent tasks side by side. Replace the parallel docs/investigations/ tree with a fork of WhatDreamsCost-ComfyUI under forks/. Automation modules (new layout) - docs/automations/LTX-2.3/base-1step/ the existing single-stage automation, moved from docs/automations/LTX-2.3/ with its README, findings, and handoff-checklist rescoped to the task. The 4k-transitions backstory that produced the module moves here as backstory/4k-transitions.md. The earlier docs/investigations/experiments/ card PNGs/HTML land in example-cards/. - docs/automations/LTX-2.3/audio-lipsync/ the audio-file lip-sync module previously living under docs/investigations/01_…/. README, handoff, findings stub, backstory/, and runs/{LOOP,log}.md, all referencing the new fork. - docs/automations/README.md + CHEATSHEET.md updated to describe the module-per-task pivot and the new tree shape. - docs/automations/CONVENTIONS.md unchanged (cross-cutting loop contract still holds). WhatDreamsCost Koolook fork (new) - forks/whatdreamscost_koolook/versions/v1_3_2/ UPSTREAM_PIN.yaml pinned to upstream e81223a (v1.3.2), GPL-3.0 verified 2026-05-28. ltx_director.py MODIFIED — adds the relay_overrides multiline JSON widget on LTXDirector. Moved from docs/investigations/01_…/patches/. prompt_relay.py MODIFIED — per-segment σ per the Prompt-Relay paper formula σ = (L − w_eff) / (2 · √ln(1/ε)). Moved from docs/investigations/01_…/patches/. patches.py UNMODIFIED — vendored verbatim from upstream because ltx_director.py imports detect_model_type and apply_patches from it (same precedent as radiance_koolook's color_helpers.py). - Registered as LTXDirector__koolook_v1_3_2 ("LTX Director (Koolook v1.3.2)") so it coexists with an installed copy of upstream WhatDreamsCost-ComfyUI. Wired into the root __init__.py NODE_CLASS_MAPPINGS merge. - forks/forks_manifest.yaml + forks/THIRD_PARTY.md updated with the new entry (attribution, change log, license verified date, exposed node IDs). Removed - docs/investigations/ entire tree retired; its contents either moved to automation modules or to forks/. - scripts/sync_investigation_patches.py no longer needed — fork code lives in the Koolook tree and is synced via the standard dev-sync flow. - KOLOOK_WHATDREAMSCOST_PATH in .env.example same reason. - docs/investigations in pyproject.toml's ruff extend-exclude added in |
||
|
malkuthroandClaude Opus 4.7
|
ff79b0f418 |
feat(video): add Easy_VideoCombine (#154)
* feat(video): add Easy_VideoCombine with absolute-path support Adds a Koolook subclass of VHS VideoCombine that bypasses ComfyUI's output-directory sandbox when filename_prefix is an absolute path — the VFX-pipeline missing piece for writing renders into a project tree (e.g. E:/projects/shot01/v003) instead of ComfyUI/output/. Implementation: - k_video_combine.py: ~60 LoC subclass + a pure _resolve_abs_target helper. Scoped-patches folder_paths.get_save_image_path for the duration of one combine_video() call so upstream's encoder, audio mux, batch manager, and progress bar run unchanged. No VHS code copied. Self-skips registration when VHS isn't importable (clean one-line stderr notice). - Discriminator: os.path.isabs(filename_prefix). No new input pin; pattern borrowed from spacepxl/ComfyUI-HQ-Image-Save's SaveEXR. - Optional create_path_if_missing BOOLEAN (default False) so typos surface as FileNotFoundError instead of spawning stray dirs. License audit (2026-05-16): - VHS upstream: GPL-3.0 (verified via raw LICENSE fetch). - HQ-Image-Save pattern source: MIT, copyright 2023 spacepxl. - Koolook: already GPL-3.0; both upstreams are compatible. - No source code from either upstream is incorporated. Verification: 18 pytest cases green (12 existing + 6 new covering relative passthrough, absolute resolution, create-on/off branches, trailing-separator normalization, and empty-basename rejection). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(starter-preset): pre-favourite Easy_VideoCombine Adds the new node to the shipped starter_preset.json picks so fresh installs surface it in the Kforge Labs sidebar's Nodes section without the user having to discover it via Comfy's stock search first. Pairs naturally with the existing LoadEXR / SaveEXRFrames picks from HQ-Image-Save. Existing installs with non-empty picks are unaffected (the seeder skips them by design); they can either Load -> "starter" in the Snapshot dialog or hit "+" once in the Nodes section to pull it in. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(dev-sync): include k_video_combine.py in RUNTIME_PATHS Without this entry the new Easy_VideoCombine module is never copied into the dev ComfyUI install, so the node fails to register on restart with a silent ImportError. Caught when the maintainer restarted ComfyUI and couldn't find the node in the search palette. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(starter-preset): include all Koolook nodes in default picks Fresh installs should land on a populated Nodes (favorites) panel that immediately surfaces what Koolook provides — otherwise new users see only third-party picks and have no entry point into the Koolook-native workflow. Adds the 8 currently-registered Koolook node IDs to the shipped starter preset alongside the existing third-party essentials: EasyAIPipeline, EasyResize_Koolook, EasyWan22Prompt, Easy_Pattern, Easy_VideoCombine, Easy_hdr_VAE_decode, Easy_hdr_VAE_encode, KoolookLoadCameraPosesAbsolute, easy_ImageBatch Also adds Easy_VideoCombine to the README's Included Nodes table. Existing installs with non-empty picks are unaffected by the seeder (by design). To pull the Koolook defaults into an existing install, Quick Save first, then Snapshot -> Load -> "starter". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(video): reach VHS_VideoCombine via nodes.NODE_CLASS_MAPPINGS The previous `from videohelpersuite.nodes import VideoCombine` fails inside ComfyUI's loader context because VHS uses relative imports throughout (`from .logger ...`, `from .utils ...`) that only resolve when VHS is loaded through ComfyUI's importlib spec — not through a plain absolute import path. The error surfaced as No module named 'utils.install_util'; 'utils' is not a package at Koolook load time, which caused the node to self-skip with an empty NODE_CLASS_MAPPINGS — invisible in ComfyUI's search palette. Fix: look up VHS_VideoCombine through `nodes.NODE_CLASS_MAPPINGS`, which is the documented cross-pack interface. ComfyUI's loader (custom_nodes/.../__init__.py exec_module → merge into global) runs to completion for each pack before moving to the next, and VHS sorts before Koolook alphabetically in every layout we've seen, so by the time Koolook's __init__.py runs the global registry already has VHS_VideoCombine in it. Verified against ComfyUI's nodes.py load_custom_node() at load_custom_node() lines 2196 (exec_module) and 2225-2228 (merge), which guarantee this ordering. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(video): add output_directory input for split path/name editing Overloading filename_prefix for absolute paths means losing the "just type a name" affordance — once you put the path there, there's nowhere to specify just the file root. Across many renders into the same project tree this becomes tedious: every iteration retypes the whole path. Adds an optional `output_directory` STRING input. When set: - filename_prefix is treated as just the filename root; any path components are stripped (so accidental typos like "sub/clip" don't double-up under the directory). - Absolute output_directory writes directly to that location. - Relative output_directory joins under ComfyUI's output/. When output_directory is empty, the original overloaded filename_prefix behavior is preserved unchanged (back-compat for the one-field flow). Implementation: - New pure helper `_compose_prefix(filename_prefix, output_directory)` combines the two fields into one effective prefix. The existing isabs discrimination then handles sandboxed vs absolute writes uniformly downstream. - 6 new tests cover all combinations: empty/abs/relative directory, path stripping from filename_prefix, trailing-separator trim, empty-name fallback. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(video): coerce frontend 'undefined' STRING values to empty The ComfyUI frontend serializes a STRING widget with empty default as the literal string "undefined" when left untouched (a known long-running quirk). Without defensive normalization, an untouched output_directory was treated as a real directory name and prepended to the user's path: filename_prefix = "E:/renders/clip_v003" (typed by user) output_directory = "undefined" (frontend default, not typed) _compose_prefix(...) -> "undefined/clip_v003" (relative) _resolve_abs_target -> None (not absolute) super().combine_video -> writes to <output>/undefined/clip_v003_* The QuickTime ended up at <ComfyUI>/output/undefined/<name>.mov with a spurious `undefined/` subfolder, instead of at the absolute path the user actually specified. Fix: a `_normalize_text_input()` helper coerces the literal strings "undefined" / "null" / "None" (case-insensitive, post-strip) to "". Applied to both filename_prefix and output_directory at the top of combine_video() before any path composition. 4 new tests cover the sentinels and a regression test for the exact path-composition behavior that produced the bug. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(video-fe): wire format-specific widgets for Easy_VideoCombine VHS's frontend extension (web/js/VHS.core.js:2030) gates its format-widget machinery on `nodeData?.name == "VHS_VideoCombine"` — strict-equality on the registered node ID. Our subclass uses a different ID (`Easy_VideoCombine`), so none of VHS's per-format widget rendering ran. The user saw: [VideoHelperSuite] - WARNING - Missing input for p has been set to hq on every execution (profile defaulting to "hq" for ProRes), and had no way to pick anything other than the default — because the profile dropdown was never rendered. Fix: a minimal frontend extension at web/easy_video_combine.js (~80 LoC) that replicates `addFormatWidgets` for our node name. Hooks the format dropdown's callback to add/remove per-format widgets the same way upstream does. Self-contained — does not import from VHS's module (their helpers aren't exported as ES imports). Verifies via the actual ComfyUI install after dev-sync; static preview servers can't exercise it (depends on LiteGraph's registered_node_types at runtime). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(video): strip 'undefined' from anywhere + one-shot debug log The previous normalization only caught the EXACT literal "undefined" / "null" / "None" at field-level. Adds two stronger defenses for the case where a sentinel still slips through: 1. _normalize_text_input now also strips surrounding quotes (paste- from-Explorer artifact) and handles embedded newlines by taking the first non-empty non-sentinel line (defends against multi-line Text Multiline inputs with trailing blank or junk lines). 2. New _strip_sentinel_components helper drops any path component that is exactly a sentinel string from the composed effective_ prefix — belt-and-suspenders for any path that somehow carries `\undefined\` as a literal segment by the time it reaches us. Also adds a one-shot print() of fp_raw, od_raw, normalized values, and effective_prefix at the top of combine_video. The user is seeing an undefined/ folder under their custom path despite the field-level normalization, and pure code tracing doesn't reproduce it — the log will surface what kwargs actually contains at run time so we can fix the real cause if defensive stripping doesn't already handle it. To be removed in a follow-up once root cause is known. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(video): trim node docstring + tooltips, remove debug print - Class docstring collapsed from ~20 lines to a single sentence (the frontend was showing it as the node description on right-click, cluttering the UI). - Widget tooltips trimmed to one short sentence each — same intent, no walls of text. - Debug print() added in |
||
|
malkuthroandClaude Opus 4.7
|
8e489ad37a |
feat(snapshot): redesigned Save/Load dialogs + folder picker (#137) (#144)
* feat(snapshot): row-augment max(periodic, pre_load_*) + pytest infra (#137) Server-side preparation for the snapshot dialogs redesign. The Load dialog's "newer auto-save available" row flag now reflects the freshest recovery file regardless of which mechanism wrote it — before this we only stat'd ``periodic.json``, missing cases where the user's most recent Load wrote a fresher ``pre_load_*.json`` snapshot inside the same autosave subfolder. Extracts ``_latest_autosave_mtime(autosave_dir, named_mtime)`` from the inline block in ``list_presets`` so the scan is unit-testable. The helper walks ``periodic.json`` and every ``pre_load_*.json`` once, returns the max mtime if strictly newer than the named file, else ``None``. Same shape as before, broader source set. Introduces pytest as a first-class test runner: * ``tests/server/test_listing_autosave_mtime.py`` — 5 row-augment test cases (periodic-only, pre_load-only, max-of-both, no-newer, missing-dir). * ``[tool.pytest.ini_options]`` in ``pyproject.toml`` scopes discovery to ``tests/`` and uses ``importmode=importlib``. * ``[project.optional-dependencies] test = [...]`` defines the test extra (pytest + aiohttp, which is a runtime dep of ``koolook_routes`` that must be importable even when only a pure helper is under test). * ``.github/workflows/ci.yml`` gains a ``pytest`` job that installs via ``pip install .[test]`` and runs ``pytest -q``. Guards ``__init__.py``'s relative imports in a ``try / except ImportError / else`` block. Pytest's package collector walks up from test files looking for ``__init__.py`` and tries to import the one at the repo root — but the relative imports inside it only resolve under ComfyUI's custom-node loader, not under a standalone Python import. Without the guard, every pytest run aborts during collection with ``attempted relative import with no known parent package``. The fallback emits a one-line diagnostic and registers empty mappings so the package directory can still be imported by tooling. ComfyUI's loader still hits the ``else`` branch and behaviour is unchanged for real users. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(snapshot): folder picker + agent-driven visual harness (#137) Commit 2 of the snapshot dialogs redesign. Introduces the navigate- into folder picker from mockup section 6 — the picker that opens when the user clicks Save to… / Load from… in the redesigned dialogs (and that the Settings cog will be removed in favour of, once commit 6 lands). Path input is the source of truth: type/paste + Enter navigates, clicking a folder row drills in, ↑ Up climbs one level. Files appear greyed below the directories as a "yes, this is the folder I expected" affordance. End-visible path overflow: the path input uses ``direction: rtl`` + ``unicode-bidi: plaintext`` so when the path exceeds the input width the leading ``/Users/…`` clips on the left, never the trailing folder name. Per mockup section 6. Visual harness — the path the issue called out as "visual verification is a hard gate": * ``.claude/launch.json`` defines ``design-server`` (``python3 -m http.server 8765``). Agent / maintainer can spin this up via ``preview_start`` or directly from a terminal. * ``docs/designs/_harness/folder-picker.html`` mounts the real ``showFolderPicker`` from ``web/sidebar/modals.js`` against a canned filesystem, with one trigger button per mockup-illustrated state (default / new-folder-input / empty / error). Cache-busts module imports so editing modals.js while the harness is open picks up fresh code on reload. * ``docs/maintainers/visual-harness.md`` documents the workflow, when to use the harness vs. dev-sync, and the backtick-in-CSS- comments pitfall that broke this commit twice during development. * ``CLAUDE.md`` now lists the harness as the preferred agent-driven path; dev-sync stays the integration-level gate. Server: ``/koolook/presets/browse`` gains an opt-in ``?files=1`` parameter that returns child JSON files alongside the directories. Default response shape unchanged so the legacy Settings dialog keeps working. ``_list_child_files`` mirrors ``_list_child_dirs``'s swallow-OSError-per-entry policy so a broken symlink mid-listing doesn't sink the whole response. Tests: six new pytest cases in ``tests/server/test_browse_list_children.py`` covering directory sorting, autosave-subfolder hiding, JSON-only filter, case-insensitive extension match, and the empty-result paths. ``pytest`` runs all 11 tests cleanly under the existing CI job from commit 1. What's still verified via dev-sync (not via the harness): * Real Comfy theme tokens (the harness fakes ``--comfy-menu-bg`` etc. with dark-theme defaults). * End-to-end Save → Load round-trip across the new library path. * Picker interaction with ComfyUI's keyboard manager. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(snapshot): modal headers stay title-only; route gloss to tooltip (#137) Follow-up to the folder picker commit. Maintainer flagged that the "Where snapshots are saved and loaded from." subtitle below the title was extra information the user doesn't need at first glance — if the title isn't self-explanatory, fix the title, don't pad it with a sentence. The mockup at docs/designs/snapshot-dialogs.html section 6 shows a title-only header for exactly this reason. Locks the rule in as a project convention so every dialog box behaves the same way going forward: * docs/maintainers/conventions.md — new "Modal headers" section spells out the rule, what counts as a "descriptive subtitle" (remove) vs. a "functional info row" (keep), and how to plumb explanatory copy via a hover tooltip on the title instead. * web/sidebar/modals.js — makeModalShell now accepts ``titleTooltip``, applied as the ``title`` attribute on the modal title element so the gloss is mouseover-discoverable but never visible by default. showFolderPicker drops its ``subtitle`` param (it was added one commit ago, no other callers) in favour of ``titleTooltip``. * docs/designs/_harness/folder-picker.html — harness re-points its explanatory text to ``titleTooltip``. Verified in the harness: the visible header is title-only, the tooltip surfaces on hover. Applies prospectively to the Save / Load dialogs (commits 3 and 4) — they will also be title-only. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(snapshot): redesigned Save dialog with inline library row + Save to... + 4-state button (#137) Commit 3 of the snapshot dialogs redesign. The Save dialog is now a single modal (was three: three-way overwrite / input prompt / confirm) that matches mockup section 2: * Inline library-path info row at the top of the body — "Saved to" label on the left, "Open folder ↗" link on the right, full path below. Pure information; every action lives in the bottom command bar. * Bottom command bar with split layout: Save to... on the left, Cancel + Save as new... + Save on the right (spacer in between). * Save to... opens the folder picker from commit 2. On Use this folder, the dialog calls saveSettings() to persist the new library path, refreshes its in-place library row, and flips the primary button into Dirty state. * The primary Save button cycles through the four states from mockup section 5: - Default ("Save") — clean state, primary blue. - Dirty ("Save to new folder") — folder differs from the one the dialog opened with; the label spells out the consequence. - In progress ("Saving...") — disabled, dimmed primary. - Done ("Saved") — disabled, subtle green, brief pause before the dialog closes. * Save as new... still routes through the existing showInputModal plus presetExists guard, but the prompt no longer carries a library-path subtitle — the parent dialog already shows the target. Convention from the previous commit applied. tree.js now passes the new capability functions (saveSettings, browseDirectories, createBrowseDirectory, revealPresetFolder) into showSaveSnapshotDialog so the dialog can invoke the folder picker end-to-end inside the Save flow. Visual harness: docs/designs/_harness/save-dialog.html exercises all five mockup states (default tracked, untracked Save-as-new, dirty, in-progress, done) so the maintainer (or future agent) can re-verify without dev-syncing. Verified all states render correctly in the preview server. Static guard against the recurring backtick-in-CSS-comments trap: tests/sidebar/test_css_template_literal.py asserts that the CSS template literal in web/sidebar/constants.js contains exactly two backticks (open + close) and nothing in between. The trap bit this work three times (commits 2 and 3) — a literal backtick inside a comment closes the template early and silently bricks the entire sidebar. The test runs under the same pytest job introduced in commit 1. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(snapshot): redesigned Load dialog with scoped recovery + inline delete (#137) Commit 4 of the snapshot dialogs redesign. The Load dialog is now a single modal matching mockup section 3: * Inline library row at the top, same shape as the Save dialog — "Loaded from" label + Open folder ↗ link + leaf folder name + full path. No 📂 emoji button (the link supersedes it). * Preset rows: click to load. The "Replace current state?" confirm modal is gone — clicking Load is the intent, the dialog title already says Load snapshot, no double-confirm. * Scoped recovery: when a preset has a newer autosave (per the server-side row-augment from commit 1), the first click DOES NOT load. It expands exactly ONE recovery row directly beneath the parent, showing the freshest entry (newest of periodic.json or pre_load_*.json) with its kind badge + meta line. The user then chooses: re-click the named row to load the named version, OR click the recovery row to restore the auto-save. Both options are surfaced inline; the user picks by clicking the one they want. * Inline delete confirm: × outlines the target row red, the bottom Close button transforms to 'Yes — delete "<name>"' (danger styling). A second click commits; Escape cancels and reverts. No second modal. * Bottom command bar matches mockup section 3 split layout: [Load from…] on the left, [Close] on the right. Load from… opens the folder picker from commit 2; on Use this folder it persists via saveSettings() and refreshes the listing. The pre-#137 YES/NO choice modal (``promptApplyChoice``) and the collapsible Recovery auto-saves <details> section are both removed — the scoped-recovery inline row + the click flow above subsume their job. (Commit 5 in the issue plan was specifically for dropping the YES/NO modal, which happens here as a natural consequence of the routing change; commit 5 will be a small follow-up if needed.) Behavior preserved from the pre-#137 implementation: * doNamedLoad — pre-load auto-save backup chain, partial-failure tracker handling, identical toast messages. * doAutosaveRestore — same backup chain, same markStateAutosaved() semantics, same restored-name derivation from subdir. Tree.js plumbs the three new caps (saveSettings, browseDirectories, createBrowseDirectory) into showLoadSnapshotDialog so the dialog can invoke the folder picker end-to-end. Visual harness: docs/designs/_harness/load-dialog.html exercises default, empty, scoped-recovery, and pending-delete states. Verified all four render correctly in the preview server. Net diff: -56 lines in the sidebar source (the redesign deletes more than it adds because the YES/NO modal + collapsible recovery disclosure + confirm modal for plain loads are all gone). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(designs): reconcile snapshot redesign sketch with shipped behaviour (#137) Commit 5 of the snapshot dialogs redesign. The code-level removal of the autosave-newer YES/NO modal happened in commit 4 as a natural consequence of the scoped-recovery inline row taking over its job; this commit reconciles the design docs so they describe what was actually built. docs/designs/snapshot-dialogs-redesign.md: * "What changes" item 4 reworded — the YES/NO modal is dropped; the affordance is now an inline scoped recovery row that opens directly beneath the clicked preset. * "What changes" gains items 2, 3, 5 covering the Save/Load dialog redesigns and the inline delete confirm that landed in commits 3 and 4. * Mermaid flow rewritten to reflect the new routing (preset click → scoped recovery expand → user picks named or recovery row), matching commit 4's onPresetClick implementation. docs/designs/snapshot-dialogs.html "Locked-in decisions": * Updated the autosave-modal item from "modal head reads X" to "modal — dropped" and explains the inline scoped recovery row replacement. * Added the title-only modal header convention (commit 2 follow-up) pointing to docs/maintainers/conventions.md. * Added a pointer to the visual harness workflow added in commit 2. No code changes — the YES/NO modal was already gone after commit 4. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(snapshot): remove Settings cog + showSnapshotSettingsDialog + CHANGELOG (#137) Final commit of the snapshot dialogs redesign. The Settings cog disappears from the Snapshot toolbar — the snapshot row is now status indicator | Load | Quick Save | Save, matching mockup section 1. Library-path management now lives inside the Save dialog ("Save to…") and Load dialog ("Load from…"), both of which open the folder picker introduced in commit 2 and persist via the existing `saveSettings` route. Removals: * `showSnapshotSettingsDialog` (~325 lines) — the dialog body, its inline directory browser, and the "save folder location" state machine. The same browse functionality lives in `showFolderPicker` now; the same persistence happens via `saveSettings()` called from inside Save / Load. * `showThreeWayDialog` (~45 lines) — only consumer was the pre-#137 Save flow (overwrite / Save as new / Cancel), which the redesigned Save dialog absorbed into its primary action bar. * `renderLibraryLocation` (~17 lines) — replaced by inline rendering in the Save and Load dialogs' library rows. * Settings cog toolbar button + `openSettingsDialog` wiring in tree.js (~12 lines). * Orphan `.koolook-browse-*`, `.koolook-path-input-row`, `.koolook-load-library-row*`, `.koolook-settings-save-note` CSS rules in constants.js (~16 lines) — only consumed by the removed dialog. `.koolook-settings-folder-name` and `.koolook-settings- folder-path` stay; the new Save / Load library rows reuse them verbatim (git-blame continuity for the visual style). Adds: * CHANGELOG.md [Unreleased] entry summarising the whole #137 feature — picker, redesigned Save/Load dialogs, scoped recovery, inline delete, removed Settings cog, pytest infrastructure. Net diff for this commit: -377 lines. Cumulative PR diff (commits 1–6 + harness/convention follow-ups): all six issue ACs satisfied. The maintainer should dev-sync once before merging to confirm real-Comfy theme tokens render correctly (harness fakes the dark-theme defaults) and end-to-end Save → Load round-trips against the live preset folder behave. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(ci): install pytest + aiohttp directly, not via pip install . (#137) The Pytest job introduced in commit 1 of this PR was failing on CI because ``pip install .[test]`` tries to build a wheel of the entire project. setuptools's flat-layout discovery then trips on the non-Python ``web/`` and ``forks/`` directories at the repo root and errors out with "Multiple top-level packages discovered in a flat- layout: ['web', 'forks']". The project is a ComfyUI custom node loaded directly from its source directory, not a pip-installable package. There's no wheel to build. Install the test deps explicitly instead — the ``[project.optional- dependencies] test`` block in pyproject.toml is the spec, CI mirrors it by hand. Worked locally because the venv install path went straight to ``pip install pytest aiohttp`` and never touched the project. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(snapshot): untracked dirty state + Load dialog Escape listener leak (#137) Two findings from the external review on PR #144: **Blocker — Save dialog's primary action with no tracked preset + folder changed via Save to...** ``updatePrimaryLabel`` was entering the "dirty" state whenever the library path diverged from the original, regardless of whether a tracked preset existed. The primary click handler then routed "dirty" to ``doOverwrite(current)`` — and with ``current === null``, ``sanitizeName(null)`` would throw / silently coerce to an empty string instead of asking the user for a name. Fix: ``isDirty()`` only matters when there is something to overwrite. With no tracked preset, the primary stays in ``default-new`` regardless of folder divergence. The Save to... button still updates the library row, so the user has feedback the destination changed; the eventual primary click then routes to the Save-as-new name prompt and writes into the new folder cleanly. Verified in the visual harness: opening Save as untracked, switching the library to ``/Volumes/Shared/different-library`` via the picker, then inspecting the primary button: * Label stays "Save" (was "Save to new folder" pre-fix) * Tooltip "Choose a name and write the snapshot." (was "Write into /Volumes/Shared/different-library.") * Library row still reflects the new path (saveSettings DID run) **Major — Load dialog leaked a document keydown listener per session** The inline-delete Escape-to-cancel handler was registered with ``document.addEventListener`` but never removed. Each open/close cycle of the Load dialog accumulated another stale handler; long sessions ended up firing the cancelDelete check on every Escape press across every dialog (no-op because ``pendingDelete`` was null in closed dialogs, but the listener was still alive and called). Fix: capture the handler by name, attach on open, tear down on close by wrapping ``overlay.remove`` (mirrors the pattern the modal shell already uses at ``modals.js:52-66`` for its own Escape handler). Verified by instrumenting ``document.{add,remove}EventListener`` and opening + closing the Load dialog five times in a row: every keydown listener added during the cycle was removed before the next cycle opened. No accumulation. CI gates: ruff, compileall, bandit (via preflight equivalent), pytest 12/12, preflight 4/4 — all green. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
malkuthroandClaude Opus 4.7
|
9dab85c695 |
feat(easy_pattern): add test-pattern node + auto-restart dev-sync (#135)
* feat(easy_pattern): add test-pattern node + auto-restart dev-sync Adds an `Easy_Pattern` node (Koolook/Testing) for solid-color frame batches with optional per-frame index overlays — verifies sequence merging, frame insertion, batch indexing, and video-pipeline frame order at a glance. Color inputs use the same `White/Black/Gray/Custom` dropdown idiom `EasyResize_Koolook` already exposes for pad/panel colors, so "Gray" means the same 50% sRGB across nodes. `scripts/sync_to_dev.py` now restarts the live ComfyUI server after a successful sync (via ComfyUI-Manager's `/manager/reboot`). Default-on because dev-sync is only ever invoked to see new code, and Python custom-node files load once at server start — `--no-restart` is the escape hatch. The `RUNTIME_PATHS` manifest also picks up `k_easy_pattern.py` so adding a new top-level node doesn't silently break the pack at import time on the next sync. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(sync-dev): restrict restart URL to http(s) for Bandit B310 `urllib.request.urlopen` accepts arbitrary schemes including `file://`, which Bandit B310 flags as a path-traversal vector. The script's typical caller is the maintainer's own shell, but a typo in `--restart-url` could surprise — and the static gate is right to flag it. Validate the scheme is http/https before the urlopen call and annotate it so Bandit accepts the now-truthful narrowing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
malkuthroandClaude Opus 4.7
|
93363cc625 |
feat(sidebar): snapshot library v2 — configurable filesystem path via KFORGELABS_PRESETS (#68)
* feat(sidebar): snapshot library — named presets for picks + workflows Implements #46 item 1 (custom save location + cross-machine portability) as a preset library: each preset is one named JSON file under /userdata/koolook-presets/, capturing the full Kforge Labs state. Use cases addressed ------------------- - **Personal portability:** copy `koolook-presets/` between machines, or sync ComfyUI's userdata folder via Dropbox/iCloud/Drive — your kit follows. - **Team / facility shared library:** multiple workstations on the same ComfyUI server share `/userdata/koolook-presets/` natively. One person publishes "facility kit", everyone else loads it. - **One-off transfer (different servers):** Manage tab's Download button saves a snapshot to local disk; Upload imports it into another machine's library. UI -- New top-level **Snapshot** action row above the search field, mirroring the Nodes/Workflows action-row pattern. Three icon buttons (Load / Save / Manage), each opens the same modal at a different starting tab. - Load tab: lists every preset with metadata (workflow count · pick count · export date). Click a row → "Replace current state?" confirm → restores both picks and workflows. - Save tab: input field defaults to `preset YYYY-MM-DD`, fully editable. Existing-name save prompts to overwrite. - Manage tab: per-row Download / Delete buttons + a top-of-tab Upload action (file picker imports a JSON snapshot from disk into the library). Schema ------ ```json { "kind": "koolook-snapshot", "version": 1, "name": "Wan video kit", "exportedAt": "2026-05-05T08:42:13.000Z", "picks": [...], "workflows": { "directories": {...} } } ``` `kind` + `version` for future schema migrations. Anything without `kind === "koolook-snapshot"` is rejected on import. Module shape ------------ - `web/sidebar/snapshot.js` (new) — server I/O on the preset directory, schema validation, gather/apply, download/upload helpers. UI-free. - `web/sidebar/modals.js` — new `showSnapshotModal({initialTab, ...})` that takes capability functions instead of importing snapshot.js directly, so the UI module stays UI-only. - `web/sidebar/workflows_store.js` — new `replaceAllWorkflows(rawStore)` for bulk-replace (bypasses persistMutation since the entire cache is being swapped) + `getAllWorkflowsForExport()` for the gather flow. - `web/sidebar/picks_store.js` — new `setAllPicks(picks)` exported helper. - `web/sidebar/tree.js` — new top action row + 3 toolbar buttons + one wiring helper. - `web/sidebar/constants.js` — CSS for tab strip, list rows, manage buttons. Deliberate scope cuts (per the design discussion) ------------------------------------------------- - No continuous sync. Each save is a deliberate act. - No conflict resolution. Same preset name written from two machines on the same server = last-write-wins. Documented as a known limit; revisit when it bites. - No selective import (preset = full state). Merge can be a v2. - No custom userdata path config. The `koolook-presets/` location is fixed; portability comes from the snapshot files themselves, not from rerouting storage. Verification ------------ - `node --input-type=module --check` on all modified + new JS files — clean. - `python tools/preflight_release.py --skip manager-meta -v` PASS. - Synced to live install on Mac for browser smoke-test. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(sidebar): snapshot library now uses configurable filesystem path Reworks PR #67 (which served snapshots from /userdata/koolook-presets/) to expose the library through Koolook's own server-side endpoints, making the storage location configurable via an env var. The userdata namespace is per-install — fine for personal use, useless for the "facility shared library" use case the user explicitly called out. Architectural change -------------------- - The client no longer talks to ComfyUI's `/userdata/*` API for snapshot files. New JS-side endpoints: `/koolook/presets/{info,list,file}`. - Server side: new `koolook_routes.py` registers aiohttp routes on ComfyUI's PromptServer at custom-node load time. Path-traversal protection at the route boundary via a strict filename whitelist regex (`^[A-Za-z0-9 _.()\-]+\.json$`). - Storage location is read from the `KFORGELABS_PRESETS` env var on every request (so re-launching ComfyUI with a new value picks it up without any cache fuss). Default fallback: `<comfyui-userdata>/koolook-presets/`. - Auto-creates the configured directory on first save (parent must already exist — for NFS/SMB mounts the mount must be present). Use cases this enables vs. PR #67 --------------------------------- | Use case | PR #67 | This PR | |-------------------------|--------|---------| | Personal, single-machine | ✓ | ✓ | | Personal cross-machine | symlink-or-sync userdata | point env var at Dropbox folder | | NFS/SMB facility share | ✗ (locked to userdata) | ✓ point env var at mount | | Multiple parallel libs | ✗ | per-instance env var | | Read-only distribution | ✗ | ✓ (server detects, save fails cleanly) | UI difference vs. PR #67 ------------------------ - The Manage tab gains a small `i` icon at the right edge of the upload row. Hover tooltip shows the current library path, source (default vs `KFORGELABS_PRESETS` override), and writability status. No header line, no extra UI weight. Files added/changed ------------------- - `koolook_routes.py` (new, ~180 lines) — aiohttp routes + env-var config + filename validation. Loaded by `__init__.py` at startup; failure is non-fatal (the rest of Koolook still works, just the snapshot feature is unavailable). - `__init__.py` — registers `koolook_routes` import-time side-effect. - `scripts/sync_to_dev.py` — `koolook_routes.py` added to `RUNTIME_PATHS` so dev-sync includes it. - `web/sidebar/snapshot.js` — server I/O section rewritten to call `/koolook/presets/*` instead of `/userdata/koolook-presets/*`. Drops the URL-encoding workaround (we own the route now, so the `name=<file>` query param is clean and the server validates). - `web/sidebar/modals.js` — `showSnapshotModal` accepts a `getLibraryInfo` capability function; the Manage tab calls it and populates the info-icon tooltip. - `web/sidebar/tree.js` — passes `getLibraryInfo` from `snapshot.js` to the modal. - `.env.example` — new `KFORGELABS_PRESETS` block with macOS / Linux / Windows examples for personal-sync, NFS facility share, and SMB facility share patterns. - `CHANGELOG.md` — Added entry rewritten for the env-var design; Internal entry on the new server-side module. Why a separate v2 PR (instead of patching #67) ---------------------------------------------- The userdata-namespace design in #67 was the wrong architecture for what the user actually wanted. Patching it to the env-var design would have meant a near-total rewrite of #67's server I/O paths mid-review. Cleaner: branch fresh from main, cherry-pick #67's UI + stores work as a base, layer the new endpoint design on top. PR #67 will be closed as superseded once this lands. Verification ------------ - `node --input-type=module --check` on every modified JS file — clean. - `python -c "import ast; ast.parse(...)"` on `koolook_routes.py` + `__init__.py` — clean. - `python tools/preflight_release.py --skip manager-meta -v` PASS. - Synced to live install on Mac. User to smoke-test against a real ComfyUI session (the previous URL-encoding 405 should be gone since the new route handlers own the URL space). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(sidebar): redesign snapshot UX — single-click Save, real Settings panel User feedback on #68's first cut: tabbed modal is too complicated, can't see existing presets in the Save flow, the env var alone isn't a real settings UI, the default-path text was stale (still claimed `/userdata/koolook-presets/` after the configurable-path landed). UX changes ---------- - **Save button → one click.** If a preset is currently loaded (tracked in localStorage as `koolook.snapshot.currentPresetName.v1`, survives reloads), prompts "Save over '<name>'?" with three options: Save (overwrite) / Save as new… (rename) / Cancel. If nothing is loaded yet, prompts for a name with `preset YYYY-MM-DD` default, fully editable. - **Load → single dialog.** Preset list with per-row × Delete button. Header line shows the current library path so the user can see where they're loading from. No tabs. - **Cog → real Settings panel.** Single field for the library path (text input, absolute path). Save / Reset to default / Close. Read- only line shows the currently-resolved path + source label (settings panel / env var / built-in default). The "Manage" tab and its Download / Upload sub-actions are gone — the user can already reach the file on disk after a save, no in-app file ops needed. Server side ----------- - New `GET /koolook/presets/settings` returns the saved library path + resolved path + source. - New `POST /koolook/presets/settings` writes the saved library path to `<comfyui-userdata>/koolook-settings.json`. Empty value clears the override. - Resolution chain in `_configured_dir()`: settings file > KFORGELABS_PRESETS env var > built-in default Reads each on every call; no module-load caching. - `info` endpoint gains a `source` field ("settings" | "env" | "default") the UI uses for the readout line. Client side ----------- - `showSnapshotModal` (tabbed) replaced by three exports: showSaveSnapshotDialog, showLoadSnapshotDialog, showSnapshotSettingsDialog. Each is self-contained, no shared state. - `snapshot.js` adds `getSettings`/`saveSettings` for the Settings dialog and `getCurrentPresetName`/`setCurrentPresetName` for the one-click Save tracking. Settings ride on the new server endpoints; current-preset tracking is browser-local. - `applySnapshot` now also updates the current-preset tracker (via the dialog wrapper, not the function itself) on successful load. - `tree.js` toolbar wires three direct callbacks instead of one `openSnapshotModal(initialTab)` helper. Stale text fix -------------- - The Save dialog's hint text in #68's first cut hardcoded "/userdata/koolook-presets/<name>.json" — stale from the userdata- era PR #67. The redesigned Save flow doesn't show that text at all (the path is no longer relevant to the user-typed name); the Load dialog's header line and the Settings dialog's read-only field are the canonical surfaces for "where my presets actually live." Deliberately out of scope ------------------------- - No per-save path picker. The library is one location at a time, set in advance via Settings or env var. Saving to ad-hoc locations would fragment the library and break the "facility shared library" use case. - No in-app Download / Upload (was in #68's first cut). Files are on disk at the configured path; the OS file manager handles transfer. Verification ------------ - `node --input-type=module --check` on every modified JS file — clean. - `python -c "import ast; ast.parse(...)"` on `koolook_routes.py` — clean. - `python tools/preflight_release.py --skip manager-meta -v` PASS. - Synced to live install on Mac. User to smoke-test against a real ComfyUI session. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(sidebar): address PR #68 review — atomicity, security, error UX Multi-agent review surfaced 2 BLOCKING + 5 HIGH + 4 MEDIUM + 3 LOW. Validated the BLOCKINGs against the actual code (both real) and applied the bundle of fixes that hold up. Modal-size HIGH (#7 file size) deferred to a follow-up — that's a wider refactor. BLOCKING fixes -------------- - **`applySnapshot` partial-failure tracker gating** (modals.js Load dialog). `setCurrentPresetName(...)` was called unconditionally before checking `{picksOk, workflowsOk}` from the apply. Partial apply (one of picks/workflows persisted, the other didn't) left the tracker pointing at the preset, so the next Save would offer "Save over '<name>'?" and silently overwrite the on-disk preset with the corrupted half-state. Now: tracker is set only on `picksOk && workflowsOk`; partial failure clears it (and the toast tells the user "Reload to recover prior state. Tracker cleared."). - **`replaceAllWorkflows` non-atomic rebind** (workflows_store.js). Rebinds `workflowsCache` before awaiting persist, with no rollback. Now: snapshots the cache via the existing `snapshotCache()` primitive that `persistMutation` uses; on persist failure, restore the prior cache and return false. Load is fully atomic. HIGH fixes ---------- - **Workflows-store invariant block updated.** Lists `replaceAllWorkflows` as the fourth legitimate rebind site (with rollback semantics). "Fifth rebind" warning bumped to "sixth." - **Explicit HEAD route on `/koolook/presets/file`** that just does `is_file()` + status, no body work. The auto-synthesis path was reading the full file from disk just to discard it — wasteful on multi-MB snapshots over Dropbox/iCloud/NFS. - **Client `sanitizeName` mirrors server whitelist regex** (`^[A-Za-z0-9 _.()\-]+$`). Characters outside the whitelist (`+`, `&`, `'`, accented letters, emoji, etc.) collapse to `_` rather than triggering an opaque HTTP 400 at save time. - **Error-reason surfacing via `await resp.text()`.** RFC 7540 strips HTTP/2 reason phrases; behind any HTTP/2-terminating proxy `resp.statusText` is empty. Reading the body always carries aiohttp's `reason=` text. Used by writePreset, deletePreset, saveSettings, readPreset. MEDIUM fixes ------------ - **Symlink follow-through hardening.** `_resolve_within_base` helper computes `(base / name).resolve()` and asserts it's still under `base.resolve()`. Defends against a symlink planted in a facility- shared library dir redirecting reads/writes off-share. - **Atomic settings write.** `_write_settings` writes to `<tmp>` then `os.replace`, so an interrupted process can't truncate the file and silently revert the user's saved library path on next load. - **`__init__.py` import wrapped in try/except.** The "non-fatal" comment promised that an import failure wouldn't break node registration; previously only `install()` returning False was guarded. Now any module-load exception is caught and logged with the `[Koolook]` prefix, and the rest of the plugin loads. LOW housekeeping ---------------- - Dropped dead exports `downloadSnapshotAsFile` + `readSnapshotFile` (left over from an earlier iteration that had Download/Upload UI). - Dropped 5 dead CSS rules (`.koolook-snapshot-tabs*`, `.koolook-snapshot-tab-pane`) from when the modal was tabbed. - `currentPresetName` cleared when the user deletes the currently- loaded preset (Load dialog's promptDelete now checks via the new `getCurrentPresetName` capability). - `loadPreview` warns to console on `kind` mismatch — non-snapshot JSON files in the library directory are now visible in DevTools rather than silently filtered. - `presetExists` is tri-state (`true` / `false` / `null`) — the Save flow refuses to write when the library is unreachable rather than silently skipping the overwrite-confirm prompt. Deferred -------- - modals.js size (1080 lines) — extracting the snapshot dialogs into their own file is a follow-up refactor. Not blocking merge. - Doc drift on maintainer docs / README env-var convention — separate documentation sweep. Verification ------------ - `python -c "import ast; ast.parse(...)"` on all .py files — clean. - `node --input-type=module --check` on every modified JS file — clean. - `python tools/preflight_release.py --skip manager-meta -v` PASS. - `grep -rn "downloadSnapshotAsFile\\|readSnapshotFile\\|koolook-snapshot-tab" web/` — zero hits (dead code gone). - Synced to live install for browser smoke-test. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * ci: nudge — re-trigger CI after webhook miss on d6d2bf8 The CI workflow didn't fire on the prior push (Socket Security ran but GitHub Actions Lint/Smoke/Pre-flight/Security Gate didn't). Empty commit to refire the pull_request webhook. * fix(routes): drop unnecessary f-string prefixes flagged by ruff (F541) Two adjacent string literals in the symlink-escape error message had an `f` prefix despite containing no placeholders. CI's ruff check fails on F541 — strip the prefixes so the lint gate passes. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
malkuthroandClaude Opus 4.7
|
428d4f538b |
chore(nodes): drop Easy_Version + bump 0.1.9 (#41)
Per maintainer direction: Easy_Version was a one-line node that turned
an INT N into the string `vNNN` (zero-padded). The first trivial node
ever written for the pack, never used in production workflows. Removed
to keep the pack's surface focused on the higher-value nodes.
Anyone needing the exact behavior can either inline `f"v{N:03d}"`
downstream of an INT primitive in their workflow, or pull a more
general string-format node from KJ Nodes / ComfyUI-Custom-Scripts.
BREAKING for any saved workflow that references the `Easy_Version` ID
(node will fail to load on import). Same migration shape as the v0.1.5
__koolook_v1_0_1 cleanup.
Also rolls the [Unreleased] PR #39 LATENT-dict-wrap fix into [0.1.9]
so it ships in the same release.
Touched:
- Deleted: k_easy_version.py
- __init__.py: removed version_mappings / version_display imports + merges
- README.md: removed node-table row, Pipeline category list update,
Features bullet, Usage section, Node Code Overview line, install
category list — Easy Version no longer mentioned anywhere
- pyproject.toml: 0.1.8 -> 0.1.9
- CHANGELOG.md: [0.1.9] entry covering the removal + the carried-over
PR #39 fix
Net effect: pack drops from 8 nodes to 7. Koolook/Pipeline subfolder
now has only EasyAIPipeline.
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
|
||
|
malkuthro
|
7f8eda4809 |
Restructure fork tracking and versioned Radiance layout.
Consolidate fork metadata into forks/, move Radiance custom code into versioned folders, and align external reference naming so upgrades and cross-machine parity are easier to manage. Made-with: Cursor |
||
|
malkuthro
|
13d78b776e | Make a Koolook cam-track-loader | ||
|
malkuthro
|
b41f6e17ca | renames | ||
|
malkuthro
|
8960af66c3 | add Easy batch | ||
|
malkuthro
|
0303343d06 | WORKING button | ||
|
malkuthro
|
92cfcaa7e9 | working - version - PIPELINE | ||
|
malkuthro
|
ccd091be82 | adding buttons - A | ||
|
malkuthro
|
949e0c52c2 | AI pipeline update - A | ||
|
malkuthro
|
d2301ee0a6 | init again | ||
|
malkuthro
|
e272606fa8 | add AI_pipline node | ||
|
malkuthro
|
94886fdf49 | Multiple files - new names | ||
|
malkuthro
|
789ebd35b1 | All working - A | ||
|
malkuthro
|
7a13bc1d03 | Size working -A | ||
|
malkuthro
|
a6a800a2a6 | Color RGB | ||
|
malkuthro
|
f3bd016ba4 | system-A | ||
|
malkuthro
|
f2c40b5a94 | Update __init__.py | ||
|
malkuthro
|
2dfd19deb0 | add in-out = positive Prompt tool | ||
|
malkuthro
|
6882ebc46b | add multiLine prompt box | ||
|
malkuthro
|
74713d0b24 | fixed init 2 comfyui | ||
|
malkuthro
|
86637d07da | init_ONLY | ||
|
malkuthro
|
bb114e48ac | add init |