GBrain
af57d757eb
新增 / New - PoseReference 节点：姿势素材库（线稿 / 深度两套）+ 节点内缩略图，选中素材以 data URI 内嵌进工作流 - 采样器「⚡预设」下拉：一键填 采样器/调度器/步数/CFG/Sigma曲线；预设表只有一份在 sampler.py， 前端走 /toggle_pass/sampler_presets 拉取；手改任一参数自动退回「手动（不套用）」 - ToggleEmptyLatent 新增 latent类型（4 档：SD 4ch/8x、SD3-Flux-Krea2 16ch/8x、Qwen2.1 64ch/16x、Flux2 128ch/16x） - DynamicRefImageEncode 新增专门的 negative 输出（内部编空文本），替代 ConditioningZeroOut 变更 / Changed - 模型类型：SD / SDXL → SD / SDXL / Z-Image（前后端双层别名迁移，旧工作流自动改写） - DynamicRefImageEncode 输出口顺序 → conditioning / negative / latent，老工作流按 link 类型自动迁移连线 - clip_type 从 14 项补到 40 项；auto 重做为「精确表 → MRO 继承链 → 前缀规则」三层 - CLIP 校验加固：手工表未登记的模型改问 model_config.clip_target()，覆盖 106/108 个模型类 - 统一加载器支持 FIBO 2.0 / MOSS-Image；采样器预设新增 FIBO 2.0 turbo（4步） 修复 / Fixed - Krea2 的 reference_latents_method: index_timestep_zero → index（前者会把参考线稿「描」进画面） - Krea2 参考图 latent fit 到出图网格 - 修正三处 Flux2 空 latent 的通道 / 尺寸（Ref Image Encode / OneImageMultiPrompt / Ref Independent） - 广播：槽位小点点亮、端点圆点改到节点下层、修「紫线伸向空白处断掉」（零尺寸 boundingRect 误判） - 输出口顺序迁移只在 onConfigure 跑一次太早（graph.links 要几百毫秒才建好）⇒ 改为阶梯重试， 否则会出现「槽位改名了、连线没搬」 其他 / Misc - 新增诊断命令 __tpAudit()：打印每条广播线两端坐标 / 归属节点 / 是否悬空 - 画布菜单新增「🔄 刷新姿势素材库」 - 移除示例 Krea2 Pose Editing (Ostris).json（硬依赖外部节点包且会换人） 版本 / Version - 新增 pyproject.toml（ComfyUI-Manager / Registry 元数据入口），版本号 v1.0.0 - __init__.py 增加 __version__，启动打印 [Toggle-Pass] v1.0.0 loaded — 13 nodes - README 顶部加版本标识
638 lines
24 KiB
Python
638 lines
24 KiB
Python
"""
|
||
ComfyUI Toggle Pass - Toggle nodes + Image IO + Multi-Ref Encode
|
||
Author: Clever
|
||
"""
|
||
# 版本号：与 pyproject.toml 的 [project].version 保持一致。
|
||
# ComfyUI-Manager / Comfy Registry 读 pyproject.toml；这里供运行时打印、方便看日志排查。
|
||
__version__ = "1.0.0"
|
||
|
||
from .multi_ref_encode import NODE_CLASS_MAPPINGS as MULTI_REF_NODES, NODE_DISPLAY_NAME_MAPPINGS as MULTI_REF_NAMES
|
||
from .unified_loader import NODE_CLASS_MAPPINGS as UNIFIED_NODES, NODE_DISPLAY_NAME_MAPPINGS as UNIFIED_NAMES
|
||
from .sampler import NODE_CLASS_MAPPINGS as SAMPLER_NODES, NODE_DISPLAY_NAME_MAPPINGS as SAMPLER_NAMES
|
||
from .pose_reference import NODE_CLASS_MAPPINGS as POSE_NODES, NODE_DISPLAY_NAME_MAPPINGS as POSE_NAMES
|
||
|
||
import os
|
||
import json
|
||
import math
|
||
import time
|
||
import random
|
||
import hashlib
|
||
import torch
|
||
import numpy as np
|
||
from PIL import Image, ImageDraw, ImageOps, ImageSequence, ImageFile, ImageFont
|
||
from PIL.PngImagePlugin import PngInfo
|
||
import comfy.model_management
|
||
import comfy.utils
|
||
import folder_paths
|
||
import node_helpers
|
||
|
||
|
||
WEB_DIRECTORY = "./js"
|
||
|
||
_last_saved_path = {}
|
||
|
||
|
||
# ==================== Toggle Load Image ====================
|
||
|
||
class ToggleLoadImage:
|
||
"""Load image from input folder with privacy toggle support."""
|
||
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
input_dir = folder_paths.get_input_directory()
|
||
files = [f for f in os.listdir(input_dir) if os.path.isfile(os.path.join(input_dir, f))]
|
||
files = folder_paths.filter_files_content_types(files, ["image"])
|
||
return {
|
||
"required": {
|
||
"image": (sorted(files), {"image_upload": True}),
|
||
},
|
||
}
|
||
|
||
RETURN_TYPES = ("IMAGE", "MASK")
|
||
RETURN_NAMES = ("IMAGE", "MASK")
|
||
FUNCTION = "load_image"
|
||
CATEGORY = "Toggle-Pass"
|
||
DESCRIPTION = "Load image with privacy toggle."
|
||
|
||
def load_image(self, image):
|
||
image_path = folder_paths.get_annotated_filepath(image)
|
||
|
||
# 尝试视频帧提取（如果安装了 comfy-extras 的 InputImpl）
|
||
try:
|
||
from comfy_extras.nodes_input import InputImpl
|
||
components = InputImpl.VideoFromFile(image_path).get_components()
|
||
if components.images.shape[0] > 0:
|
||
return (components.images, 1.0 - components.alpha[..., -1] if components.alpha is not None else torch.zeros((components.images.shape[0], 64, 64), dtype=torch.float32, device="cpu"))
|
||
except Exception:
|
||
pass
|
||
|
||
img = node_helpers.pillow(Image.open, image_path)
|
||
|
||
output_images = []
|
||
output_masks = []
|
||
w, h = None, None
|
||
|
||
dtype = comfy.model_management.intermediate_dtype()
|
||
|
||
for i in ImageSequence.Iterator(img):
|
||
i = node_helpers.pillow(ImageOps.exif_transpose, i)
|
||
|
||
if i.mode == 'I':
|
||
i = i.point(lambda i: i * (1 / 255))
|
||
image = i.convert("RGB")
|
||
|
||
if len(output_images) == 0:
|
||
w = image.size[0]
|
||
h = image.size[1]
|
||
|
||
if image.size[0] != w or image.size[1] != h:
|
||
continue
|
||
|
||
image = np.array(image).astype(np.float32) / 255.0
|
||
image = torch.from_numpy(image)[None,]
|
||
if 'A' in i.getbands():
|
||
mask = np.array(i.getchannel('A')).astype(np.float32) / 255.0
|
||
mask = 1. - torch.from_numpy(mask)
|
||
elif i.mode == 'P' and 'transparency' in i.info:
|
||
mask = np.array(i.convert('RGBA').getchannel('A')).astype(np.float32) / 255.0
|
||
mask = 1. - torch.from_numpy(mask)
|
||
else:
|
||
mask = torch.zeros((64, 64), dtype=torch.float32, device="cpu")
|
||
output_images.append(image.to(dtype=dtype))
|
||
output_masks.append(mask.unsqueeze(0).to(dtype=dtype))
|
||
|
||
if img.format == "MPO":
|
||
break
|
||
|
||
if len(output_images) > 1:
|
||
output_image = torch.cat(output_images, dim=0)
|
||
output_mask = torch.cat(output_masks, dim=0)
|
||
else:
|
||
output_image = output_images[0]
|
||
output_mask = output_masks[0]
|
||
|
||
return (output_image, output_mask)
|
||
|
||
@classmethod
|
||
def IS_CHANGED(cls, image):
|
||
image_path = folder_paths.get_annotated_filepath(image)
|
||
m = hashlib.sha256()
|
||
with open(image_path, 'rb') as f:
|
||
m.update(f.read())
|
||
return m.digest().hex()
|
||
|
||
@classmethod
|
||
def VALIDATE_INPUTS(cls, image):
|
||
if not folder_paths.exists_annotated_filepath(image):
|
||
return "Invalid image file: {}".format(image)
|
||
return True
|
||
|
||
|
||
# ==================== Toggle Preview Image ====================
|
||
|
||
class TogglePreviewImage:
|
||
"""Preview image with privacy toggle support. Saves to temp like built-in PreviewImage."""
|
||
|
||
def __init__(self):
|
||
self.output_dir = folder_paths.get_temp_directory()
|
||
self.type = "temp"
|
||
self.compress_level = 1
|
||
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
return {
|
||
"required": {
|
||
"images": ("IMAGE",),
|
||
},
|
||
"hidden": {
|
||
"prompt": "PROMPT",
|
||
"extra_pnginfo": "EXTRA_PNGINFO",
|
||
},
|
||
}
|
||
|
||
RETURN_TYPES = ("IMAGE",)
|
||
RETURN_NAMES = ("IMAGE",)
|
||
OUTPUT_NODE = True
|
||
FUNCTION = "preview"
|
||
CATEGORY = "Toggle-Pass"
|
||
DESCRIPTION = "Preview image with privacy toggle. Saves to temp like PreviewImage."
|
||
|
||
def preview(self, images, prompt=None, extra_pnginfo=None):
|
||
results = []
|
||
for idx, img in enumerate(images):
|
||
img_np = 255. * img.cpu().numpy()
|
||
pil_img = Image.fromarray(np.clip(img_np, 0, 255).astype(np.uint8))
|
||
filename = f"toggle_preview_{idx:05}.png"
|
||
filepath = os.path.join(self.output_dir, filename)
|
||
pil_img.save(filepath, format="PNG", compress_level=self.compress_level)
|
||
results.append({
|
||
"filename": filename,
|
||
"subfolder": "",
|
||
"type": self.type,
|
||
})
|
||
|
||
return {"result": (images,), "ui": {"images": results}}
|
||
|
||
|
||
class SaveLoadImage:
|
||
"""Save image AND output it. Toggle OFF to auto-load last saved image instead."""
|
||
|
||
def __init__(self):
|
||
self.output_dir = folder_paths.get_output_directory()
|
||
self.type = "output"
|
||
self.prefix_append = ""
|
||
self.compress_level = 4
|
||
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
return {
|
||
"required": {
|
||
"images": ("IMAGE",),
|
||
"use_saved": (["Save", "Load"],),
|
||
},
|
||
"optional": {
|
||
"filename_prefix": ("STRING", {"default": "ComfyUI"}),
|
||
},
|
||
"hidden": {
|
||
"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO"
|
||
},
|
||
}
|
||
|
||
RETURN_TYPES = ("IMAGE", "MASK")
|
||
RETURN_NAMES = ("IMAGE", "MASK")
|
||
OUTPUT_NODE = True
|
||
FUNCTION = "execute"
|
||
CATEGORY = "Toggle-Pass"
|
||
DESCRIPTION = "Save image and output IMAGE. Toggle use_saved=ON to reuse last saved image."
|
||
|
||
def execute(self, images, use_saved="Save", filename_prefix="ComfyUI", prompt=None, extra_pnginfo=None):
|
||
if use_saved == "Load":
|
||
return self._load_last()
|
||
else:
|
||
return self._save(images, filename_prefix, prompt, extra_pnginfo)
|
||
|
||
def _save(self, images, filename_prefix, prompt, extra_pnginfo):
|
||
if images is None:
|
||
images = torch.zeros((1, 64, 64, 3), dtype=torch.float32)
|
||
|
||
filename_prefix += self.prefix_append
|
||
full_output_folder, filename, counter, subfolder, filename_prefix = folder_paths.get_save_image_path(
|
||
filename_prefix, self.output_dir, images[0].shape[1], images[0].shape[0]
|
||
)
|
||
results = list()
|
||
for (batch_number, image) in enumerate(images):
|
||
i = 255. * image.cpu().numpy()
|
||
img = Image.fromarray(np.clip(i, 0, 255).astype(np.uint8))
|
||
metadata = None
|
||
if not comfy.utils.args.disable_metadata:
|
||
metadata = PngInfo()
|
||
if prompt is not None:
|
||
metadata.add_text("prompt", json.dumps(prompt))
|
||
if extra_pnginfo is not None:
|
||
for x in extra_pnginfo:
|
||
metadata.add_text(x, json.dumps(extra_pnginfo[x]))
|
||
|
||
filename_with_batch_num = filename.replace("%batch_num%", str(batch_number))
|
||
file = f"{filename_with_batch_num}_{counter:05}_.png"
|
||
filepath = os.path.join(full_output_folder, file)
|
||
img.save(filepath, pnginfo=metadata, compress_level=self.compress_level)
|
||
results.append({
|
||
"filename": file,
|
||
"subfolder": subfolder,
|
||
"type": self.type
|
||
})
|
||
|
||
if id(self) not in _last_saved_path:
|
||
_last_saved_path[id(self)] = filepath
|
||
|
||
counter += 1
|
||
|
||
last_file = os.path.join(full_output_folder, results[0]["filename"])
|
||
_last_saved_path[id(self)] = last_file
|
||
|
||
mask = torch.zeros((images.shape[0], images.shape[1], images.shape[2]), dtype=torch.float32)
|
||
return {"result": (images, mask), "ui": {"images": results}}
|
||
|
||
def _load_last(self):
|
||
node_id = id(self)
|
||
if node_id in _last_saved_path and os.path.isfile(_last_saved_path[node_id]):
|
||
filepath = _last_saved_path[node_id]
|
||
img = node_helpers.pillow(Image.open, filepath)
|
||
output_images = []
|
||
output_masks = []
|
||
w, h = None, None
|
||
dtype = comfy.model_management.intermediate_dtype()
|
||
|
||
for i in ImageSequence.Iterator(img):
|
||
i = node_helpers.pillow(ImageOps.exif_transpose, i)
|
||
if i.mode == 'I':
|
||
i = i.point(lambda x: x * (1 / 255))
|
||
image = i.convert("RGB")
|
||
if len(output_images) == 0:
|
||
w = image.size[0]
|
||
h = image.size[1]
|
||
if image.size[0] != w or image.size[1] != h:
|
||
continue
|
||
image = np.array(image).astype(np.float32) / 255.0
|
||
image = torch.from_numpy(image)[None,]
|
||
if 'A' in i.getbands():
|
||
mask = np.array(i.getchannel('A')).astype(np.float32) / 255.0
|
||
mask = 1. - torch.from_numpy(mask)
|
||
elif i.mode == 'P' and 'transparency' in i.info:
|
||
mask = np.array(i.convert('RGBA').getchannel('A')).astype(np.float32) / 255.0
|
||
mask = 1. - torch.from_numpy(mask)
|
||
else:
|
||
mask = torch.zeros((h, w), dtype=torch.float32, device="cpu")
|
||
output_images.append(image.to(dtype=dtype))
|
||
output_masks.append(mask.unsqueeze(0).to(dtype=dtype))
|
||
if img.format == "MPO":
|
||
break
|
||
|
||
if len(output_images) > 1:
|
||
output_image = torch.cat(output_images, dim=0)
|
||
output_mask = torch.cat(output_masks, dim=0)
|
||
else:
|
||
output_image = output_images[0]
|
||
output_mask = output_masks[0]
|
||
|
||
return {"result": (output_image, output_mask), "ui": {"images": []}}
|
||
|
||
dummy = torch.zeros((1, 64, 64, 3), dtype=torch.float32)
|
||
dummy_mask = torch.zeros((1, 64, 64), dtype=torch.float32)
|
||
return {"result": (dummy, dummy_mask), "ui": {"images": []}}
|
||
|
||
|
||
class ToggleEmptyLatent:
|
||
"""开关空 Latent。
|
||
|
||
`latent类型` 决定通道数与空间压缩比，必须和底模的 latent_format 对齐：
|
||
· SD / SDXL 4 通道 8×
|
||
· SD3 / Flux / Krea 2 16 通道 8× （Krea2 走 Wan21 格式，latent_dimensions=3）
|
||
· Qwen-Image 2.1 64 通道 16×
|
||
· Flux2 128 通道 16×
|
||
通道给错不会立刻报错 —— K采样器里的 fix_empty_latent_channels 只会在
|
||
「latent 全零」时按模型补齐通道，属于隐式兜底；直接接别的 latent 处理节点就不会修。
|
||
"""
|
||
|
||
LATENT_PRESETS = {
|
||
"SD / SDXL（4 通道 · 8×）": (4, 8),
|
||
"SD3 / Flux / Krea 2（16 通道 · 8×）": (16, 8),
|
||
"Qwen-Image 2.1（64 通道 · 16×）": (64, 16),
|
||
"Flux2（128 通道 · 16×）": (128, 16),
|
||
}
|
||
DEFAULT_LATENT = "SD / SDXL（4 通道 · 8×）"
|
||
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
return {
|
||
"required": {
|
||
"width": ("INT", {"default": 1024, "min": 16, "max": 16384, "step": 8}),
|
||
"height": ("INT", {"default": 1024, "min": 16, "max": 16384, "step": 8}),
|
||
"batch_size": ("INT", {"default": 1, "min": 1, "max": 4096}),
|
||
"enabled": (["ON", "OFF"],),
|
||
},
|
||
"optional": {
|
||
"fallback_width": ("INT", {"default": 1024, "min": 16, "max": 16384, "step": 8}),
|
||
"fallback_height": ("INT", {"default": 1024, "min": 16, "max": 16384, "step": 8}),
|
||
# ⚠️ 必须放在 fallback_height 之后 —— 前端控件顺序是 required 全部 + optional，
|
||
# 老存档的 widgets_values 是 [width,height,batch_size,enabled,fb_w,fb_h] 六项，
|
||
# 早先把它插在 enabled 后面（fb_w 之前），老图加载时 fb_w 的 1024 会灌进来，
|
||
# 点运行直接 value_not_in_list: latent类型 1024 ⇒ 400，看着就是「跑不了」。
|
||
# 放最后 ⇒ 老图前 6 位对得上，这一项拿 default（4 通道，即老行为）。
|
||
"latent类型": (list(cls.LATENT_PRESETS.keys()), {"default": cls.DEFAULT_LATENT}),
|
||
},
|
||
}
|
||
|
||
RETURN_TYPES = ("LATENT",)
|
||
RETURN_NAMES = ("LATENT",)
|
||
FUNCTION = "generate"
|
||
CATEGORY = "Toggle-Pass"
|
||
|
||
def generate(self, width, height, batch_size=1, enabled="ON", latent类型=None,
|
||
fallback_width=None, fallback_height=None):
|
||
if fallback_width is None:
|
||
fallback_width = 1024
|
||
if fallback_height is None:
|
||
fallback_height = 1024
|
||
if enabled != "ON":
|
||
width = fallback_width
|
||
height = fallback_height
|
||
channels, scale = self.LATENT_PRESETS.get(latent类型, self.LATENT_PRESETS[self.DEFAULT_LATENT])
|
||
# 宽高先向下对齐到压缩比的倍数，避免出现非整数网格
|
||
h = max(1, height // scale)
|
||
w = max(1, width // scale)
|
||
latent = torch.zeros([batch_size, channels, h, w],
|
||
device=comfy.model_management.intermediate_device(),
|
||
dtype=comfy.model_management.intermediate_dtype())
|
||
return ({"samples": latent, "downscale_ratio_spacial": scale},)
|
||
|
||
|
||
# ==================== Image Compare ====================
|
||
# 把「N 张原图 + N 张结果」拼成一张对比图
|
||
# 后端合成 + 节点内预览（走标准 ui.images，和 SaveLoadImage 同一机制）
|
||
# =======================================================
|
||
|
||
_compare_counter = 0
|
||
|
||
|
||
class ToggleImageCompare:
|
||
"""图片对比：多张原图与多张结果拼成一张对比图，节点内直接预览。
|
||
|
||
- 两组动态输入：source1..source10（原图）、result1..result10（结果）
|
||
- 排版：自动 / 成对并排 / 上下两排 / 自动网格
|
||
- 排版按**实际接上的图**走（不看你设了几个槽），所以「N 原图 + 1 结果」
|
||
和「N 原图 + N 结果」都不用改设置
|
||
"""
|
||
|
||
_font_cache = {}
|
||
|
||
_CJK_FONTS = (
|
||
"msyh.ttc", "msyhl.ttc", "simhei.ttf", "simsun.ttc",
|
||
"/usr/share/fonts/opentype/noto/NotoSansCJK-Regular.ttc",
|
||
"/usr/share/fonts/truetype/wqy/wqy-microhei.ttc",
|
||
"/System/Library/Fonts/PingFang.ttc",
|
||
)
|
||
_LATIN_FONTS = ("arial.ttf", "segoeui.ttf", "DejaVuSans.ttf")
|
||
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
optional = {}
|
||
for i in range(1, 11):
|
||
optional[f"source{i}"] = ("IMAGE",)
|
||
optional[f"result{i}"] = ("IMAGE",)
|
||
return {
|
||
"required": {
|
||
"num_sources": ("INT", {"default": 2, "min": 1, "max": 10, "step": 1}),
|
||
"num_results": ("INT", {"default": 2, "min": 1, "max": 10, "step": 1}),
|
||
"排版": (["自动", "成对并排", "上下两排", "自动网格"],),
|
||
"标签": (["显示", "隐藏"],),
|
||
},
|
||
"optional": optional,
|
||
}
|
||
|
||
RETURN_TYPES = ("IMAGE",)
|
||
RETURN_NAMES = ("对比图",)
|
||
OUTPUT_NODE = True
|
||
FUNCTION = "compare"
|
||
CATEGORY = "Toggle-Pass"
|
||
DESCRIPTION = "把多张原图与多张结果拼成一张对比图（before / after）。排版自动适配 N:1 或 N:N。"
|
||
|
||
# ---- 字体 ----
|
||
|
||
@classmethod
|
||
def _get_font(cls, size):
|
||
size = int(size)
|
||
if size in cls._font_cache:
|
||
return cls._font_cache[size]
|
||
font, cjk = None, False
|
||
for name in cls._CJK_FONTS:
|
||
try:
|
||
font = ImageFont.truetype(name, size)
|
||
cjk = True
|
||
break
|
||
except Exception:
|
||
font = None
|
||
if font is None:
|
||
for name in cls._LATIN_FONTS:
|
||
try:
|
||
font = ImageFont.truetype(name, size)
|
||
break
|
||
except Exception:
|
||
font = None
|
||
if font is None:
|
||
try:
|
||
font = ImageFont.load_default(size=size)
|
||
except Exception:
|
||
try:
|
||
font = ImageFont.load_default()
|
||
except Exception:
|
||
font = None
|
||
cls._font_cache[size] = (font, cjk)
|
||
return cls._font_cache[size]
|
||
|
||
@classmethod
|
||
def _draw_label(cls, pil, text, size):
|
||
font, _ = cls._get_font(size)
|
||
if font is None or not text:
|
||
return pil
|
||
draw = ImageDraw.Draw(pil, "RGBA")
|
||
pad = max(3, size // 4)
|
||
x0, y0 = max(4, size // 3), max(4, size // 3)
|
||
try:
|
||
box = draw.textbbox((x0, y0), text, font=font)
|
||
except Exception:
|
||
try:
|
||
tw, th = draw.textsize(text, font=font)
|
||
except Exception:
|
||
tw, th = size * len(text), size
|
||
box = (x0, y0, x0 + tw, y0 + th)
|
||
draw.rectangle([box[0] - pad, box[1] - pad, box[2] + pad, box[3] + pad],
|
||
fill=(0, 0, 0, 150))
|
||
draw.text((x0, y0), text, font=font, fill=(255, 255, 255, 255))
|
||
return pil
|
||
|
||
# ---- 张量 -> PIL ----
|
||
|
||
@staticmethod
|
||
def _to_pils(img):
|
||
if img is None:
|
||
return []
|
||
try:
|
||
arr = img.detach().cpu().numpy()
|
||
except Exception:
|
||
arr = np.asarray(img)
|
||
if arr.ndim == 3:
|
||
arr = arr[None, ...]
|
||
out = []
|
||
for i in range(arr.shape[0]):
|
||
a = np.clip(arr[i].astype(np.float32) * 255.0, 0, 255).astype(np.uint8)
|
||
if a.ndim == 2:
|
||
out.append(Image.fromarray(a, "L").convert("RGB"))
|
||
elif a.shape[-1] == 1:
|
||
out.append(Image.fromarray(a[..., 0], "L").convert("RGB"))
|
||
elif a.shape[-1] == 4:
|
||
out.append(Image.fromarray(a[..., :3], "RGB"))
|
||
else:
|
||
out.append(Image.fromarray(a[..., :3], "RGB"))
|
||
return out
|
||
|
||
@staticmethod
|
||
def _scale_to_height(pil, h):
|
||
w0, h0 = pil.size
|
||
if h0 <= 0 or h0 == h:
|
||
return pil
|
||
w = max(1, int(round(w0 * h / h0)))
|
||
return pil.resize((w, h), Image.LANCZOS)
|
||
|
||
# ---- 合成 ----
|
||
|
||
@classmethod
|
||
def _compose(cls, rows, labels_on=True, gap=8, pad=8, bg=(28, 28, 28),
|
||
blank=(52, 52, 52)):
|
||
norm = []
|
||
for row in rows:
|
||
items = [(p, lab) for (p, lab) in row if p is not None]
|
||
if not items:
|
||
continue
|
||
target_h = max(p.size[1] for p, _ in items)
|
||
fixed = []
|
||
for (p, lab) in row:
|
||
if p is None:
|
||
fixed.append((Image.new("RGB", (target_h, target_h), blank), None))
|
||
else:
|
||
fixed.append((cls._scale_to_height(p, target_h), lab))
|
||
norm.append(fixed)
|
||
if not norm:
|
||
return Image.new("RGB", (64, 64), bg)
|
||
|
||
row_w = [sum(p.size[0] for p, _ in r) + gap * (len(r) - 1) for r in norm]
|
||
row_h = [max(p.size[1] for p, _ in r) for r in norm]
|
||
W = max(row_w) + pad * 2
|
||
H = sum(row_h) + gap * (len(norm) - 1) + pad * 2
|
||
canvas = Image.new("RGB", (W, H), bg)
|
||
y = pad
|
||
for r, rh, rw in zip(norm, row_h, row_w):
|
||
x = pad + (W - pad * 2 - rw) // 2
|
||
for p, lab in r:
|
||
if labels_on and lab:
|
||
size = int(max(12, min(38, p.size[1] * 0.055)))
|
||
cls._draw_label(p, lab, size)
|
||
canvas.paste(p, (x, y))
|
||
x += p.size[0] + gap
|
||
y += rh + gap
|
||
return canvas
|
||
|
||
@classmethod
|
||
def _build_rows(cls, sources, results):
|
||
ns, nr = len(sources), len(results)
|
||
s_lab = [f"原{i + 1}" for i in range(ns)]
|
||
r_lab = [f"结{i + 1}" for i in range(nr)]
|
||
n = max(ns, nr, 1)
|
||
rows = []
|
||
for i in range(n):
|
||
left = (sources[i], s_lab[i]) if i < ns else (None, None)
|
||
right = (results[i], r_lab[i]) if i < nr else (None, None)
|
||
rows.append([left, right])
|
||
return rows
|
||
|
||
@classmethod
|
||
def _build_stacked(cls, sources, results):
|
||
rows = []
|
||
if sources:
|
||
rows.append([(p, f"原{i + 1}") for i, p in enumerate(sources)])
|
||
if results:
|
||
rows.append([(p, f"结{i + 1}") for i, p in enumerate(results)])
|
||
return rows
|
||
|
||
@classmethod
|
||
def _build_grid(cls, sources, results):
|
||
items = [(p, f"原{i + 1}") for i, p in enumerate(sources)]
|
||
items += [(p, f"结{i + 1}") for i, p in enumerate(results)]
|
||
if not items:
|
||
return []
|
||
cols = max(1, int(math.ceil(math.sqrt(len(items)))))
|
||
return [items[i:i + cols] for i in range(0, len(items), cols)]
|
||
|
||
def compare(self, num_sources, num_results, **kwargs):
|
||
sources, results = [], []
|
||
for i in range(1, 11):
|
||
sources.extend(self._to_pils(kwargs.get(f"source{i}")))
|
||
results.extend(self._to_pils(kwargs.get(f"result{i}")))
|
||
|
||
layout = kwargs.get("排版", "自动")
|
||
labels_on = kwargs.get("标签", "显示") != "隐藏"
|
||
|
||
if layout == "自动":
|
||
layout = "成对并排" if (len(sources) == len(results) and sources) else "上下两排"
|
||
if layout == "成对并排":
|
||
rows = self._build_rows(sources, results)
|
||
elif layout == "上下两排":
|
||
rows = self._build_stacked(sources, results)
|
||
else:
|
||
rows = self._build_grid(sources, results)
|
||
|
||
canvas = self._compose(rows, labels_on=labels_on)
|
||
|
||
arr = np.asarray(canvas).astype(np.float32) / 255.0
|
||
out = torch.from_numpy(arr)[None,]
|
||
|
||
global _compare_counter
|
||
_compare_counter += 1
|
||
ui_images = []
|
||
try:
|
||
temp_dir = folder_paths.get_temp_directory()
|
||
fn = f"image_compare_{int(time.time() * 1000)}_{_compare_counter:04d}.png"
|
||
canvas.save(os.path.join(temp_dir, fn), format="PNG", compress_level=1)
|
||
ui_images = [{"filename": fn, "subfolder": "", "type": "temp"}]
|
||
except Exception as e:
|
||
print(f"[ToggleImageCompare] 预览图保存失败: {e}")
|
||
|
||
return {"ui": {"images": ui_images}, "result": (out,)}
|
||
|
||
|
||
NODE_CLASS_MAPPINGS = {
|
||
**MULTI_REF_NODES,
|
||
**UNIFIED_NODES,
|
||
**SAMPLER_NODES,
|
||
**POSE_NODES,
|
||
"ToggleEmptyLatent": ToggleEmptyLatent,
|
||
"SaveLoadImage": SaveLoadImage,
|
||
"ToggleLoadImage": ToggleLoadImage,
|
||
"TogglePreviewImage": TogglePreviewImage,
|
||
"ToggleImageCompare": ToggleImageCompare,
|
||
}
|
||
|
||
NODE_DISPLAY_NAME_MAPPINGS = {
|
||
**MULTI_REF_NAMES,
|
||
**UNIFIED_NAMES,
|
||
**SAMPLER_NAMES,
|
||
**POSE_NAMES,
|
||
"ToggleEmptyLatent": "Toggle Empty Latent",
|
||
"SaveLoadImage": "Save/Load Image",
|
||
"ToggleLoadImage": "Toggle Load Image",
|
||
"TogglePreviewImage": "Toggle Preview Image",
|
||
"ToggleImageCompare": "图片对比",
|
||
}
|
||
|
||
print(f"[Toggle-Pass] v{__version__} loaded — {len(NODE_CLASS_MAPPINGS)} nodes")
|