313 lines
16 KiB
Python
313 lines
16 KiB
Python
import torch
|
|
import torch.nn.functional as F
|
|
import json
|
|
import numpy as np
|
|
import folder_paths
|
|
from PIL import Image, ImageOps
|
|
import os
|
|
import http.server
|
|
import socketserver
|
|
import threading
|
|
import math
|
|
|
|
# --- LOGIQUE DU SERVEUR D'APERÇU ---
|
|
preview_server_thread = None
|
|
PREVIEW_SERVER_PORT = 8189
|
|
|
|
def start_preview_server():
|
|
global preview_server_thread
|
|
if preview_server_thread is None or not preview_server_thread.is_alive():
|
|
class SecureHandler(http.server.SimpleHTTPRequestHandler):
|
|
def __init__(self, *args, **kwargs):
|
|
super().__init__(*args, directory=folder_paths.get_temp_directory(), **kwargs)
|
|
def end_headers(self):
|
|
self.send_header('Access-Control-Allow-Origin', '*')
|
|
super().end_headers()
|
|
def do_GET(self):
|
|
if self.path == '/' or self.path.endswith('/'):
|
|
self.send_error(403, "Directory listing is not allowed")
|
|
return
|
|
super().do_GET()
|
|
def log_message(self, format, *args):
|
|
return
|
|
address = ("127.0.0.1", PREVIEW_SERVER_PORT)
|
|
socketserver.TCPServer.allow_reuse_address = True
|
|
httpd = socketserver.TCPServer(address, SecureHandler)
|
|
thread = threading.Thread(target=httpd.serve_forever)
|
|
thread.daemon = True
|
|
thread.start()
|
|
preview_server_thread = thread
|
|
print(f"\n[Layer System] INFO: Démarrage du serveur d'aperçu local sur http://127.0.0.1:{PREVIEW_SERVER_PORT}")
|
|
|
|
# --- FONCTIONS UTILITAIRES ---
|
|
def tensor_to_pil(tensor):
|
|
return Image.fromarray(np.clip(255. * tensor.cpu().numpy().squeeze(), 0, 255).astype(np.uint8))
|
|
|
|
def pil_to_tensor(image):
|
|
return torch.from_numpy(np.array(image).astype(np.float32) / 255.0).unsqueeze(0)
|
|
|
|
def prepare_layer(top_image, base_image, resize_mode, scale, offset_x, offset_y):
|
|
B, base_H, base_W, C = base_image.shape
|
|
_, top_H, top_W, top_C = top_image.shape
|
|
if scale != 1.0:
|
|
new_H, new_W = int(top_H * scale), int(top_W * scale)
|
|
if new_H > 0 and new_W > 0:
|
|
top_image = F.interpolate(top_image.permute(0, 3, 1, 2), size=(new_H, new_W), mode='bilinear', align_corners=False).permute(0, 2, 3, 1)
|
|
top_H, top_W = new_H, new_W
|
|
|
|
canvas = torch.zeros(B, base_H, base_W, top_C, device=base_image.device)
|
|
if resize_mode == 'stretch':
|
|
return F.interpolate(top_image.permute(0, 3, 1, 2), size=(base_H, base_W), mode='bilinear', align_corners=False).permute(0, 2, 3, 1)
|
|
elif resize_mode == 'fit':
|
|
if top_W == 0 or top_H == 0: return canvas
|
|
ratio = min(base_W / top_W, base_H / top_H)
|
|
fit_H, fit_W = int(top_H * ratio), int(top_W * ratio)
|
|
resized_top = F.interpolate(top_image.permute(0, 3, 1, 2), size=(fit_H, fit_W), mode='bilinear', align_corners=False).permute(0, 2, 3, 1)
|
|
y_start, x_start = (base_H - fit_H) // 2, (base_W - fit_W) // 2
|
|
canvas[:, y_start:y_start+fit_H, x_start:x_start+fit_W, :] = resized_top
|
|
return canvas
|
|
elif resize_mode == 'cover':
|
|
if top_W == 0 or top_H == 0: return canvas
|
|
ratio = max(base_W / top_W, base_H / top_H)
|
|
cover_H, cover_W = int(top_H * ratio), int(top_W * ratio)
|
|
resized_top = F.interpolate(top_image.permute(0, 3, 1, 2), size=(cover_H, cover_W), mode='bilinear', align_corners=False)
|
|
y_start, x_start = (cover_H - base_H) // 2, (cover_W - base_W) // 2
|
|
src_y_end = min(y_start + base_H, cover_H)
|
|
src_x_end = min(x_start + base_W, cover_W)
|
|
canvas_permuted = resized_top[:, :, y_start:src_y_end, x_start:src_x_end]
|
|
return canvas_permuted.permute(0, 2, 3, 1)
|
|
elif resize_mode == 'crop':
|
|
x_start, y_start = int(offset_x), int(offset_y)
|
|
src_x_start, src_y_start = int(max(0, -x_start)), int(max(0, -y_start))
|
|
dst_x_start, dst_y_start = int(max(0, x_start)), int(max(0, y_start))
|
|
copy_W = int(min(base_W - dst_x_start, top_W - src_x_start))
|
|
copy_H = int(min(base_H - dst_y_start, top_H - src_y_start))
|
|
if copy_W > 0 and copy_H > 0:
|
|
src_slice = top_image[:, src_y_start:src_y_start+copy_H, src_x_start:src_x_start+copy_W, :]
|
|
canvas[:, dst_y_start:dst_y_start+copy_H, dst_x_start:dst_x_start+copy_W, :] = src_slice
|
|
return canvas
|
|
return canvas
|
|
|
|
|
|
class LayerSystem:
|
|
OUTPUT_NODE = True
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(cls):
|
|
header_anchors = {}
|
|
for i in range(1, 11):
|
|
header_anchors[f"header_anchor_{i}"] = ("STRING", {"multiline": True, "default": ""})
|
|
|
|
optional_inputs = {
|
|
"_properties_json": ("STRING", {"multiline": True, "default": "{}"}),
|
|
"_preview_anchor": ("STRING", {"multiline": True, "default": "PREVIEW_ANCHOR"}),
|
|
}
|
|
optional_inputs.update(header_anchors)
|
|
|
|
return {
|
|
"required": {
|
|
"base_image": ("IMAGE",),
|
|
},
|
|
"optional": optional_inputs
|
|
}
|
|
|
|
RETURN_TYPES = ("IMAGE",)
|
|
FUNCTION = "composite_layers"
|
|
CATEGORY = "Layer System"
|
|
DESCRIPTION = "This custom node for ComfyUI provides a powerful and flexible dynamic layering system, similar to what you would find in image editing software like Photoshop."
|
|
|
|
@classmethod
|
|
def IS_CHANGED(cls, **kwargs):
|
|
return float("NaN")
|
|
|
|
def _blend(self, base, top, mode):
|
|
if mode == 'normal': return top
|
|
if mode == 'multiply': return base * top
|
|
if mode == 'screen': return 1.0 - (1.0 - base) * (1.0 - top)
|
|
if mode == 'overlay': return torch.where(base < 0.5, 2.0 * base * top, 1.0 - 2.0 * (1.0 - base) * (1.0 - top))
|
|
if mode == 'soft_light': return torch.where(top < 0.5, 2.0 * base * top + base.pow(2.0) * (1.0 - 2.0 * top), torch.sqrt(base) * (2.0 * top - 1.0) + 2.0 * base * (1.0 - top))
|
|
if mode == 'hard_light': return torch.where(top < 0.5, 2.0 * top * base, 1.0 - 2.0 * (1.0 - top) * (1.0 - base))
|
|
if mode == 'difference': return torch.abs(base - top)
|
|
if mode == 'color_dodge':
|
|
denominator = 1.0 - top
|
|
return torch.where(denominator < 1e-6, torch.ones_like(base), torch.clamp(base / (denominator + 1e-6), 0, 1))
|
|
if mode == 'color_burn':
|
|
return torch.where(top < 1e-6, torch.zeros_like(base), 1.0 - torch.clamp((1.0 - base) / (top + 1e-6), 0, 1))
|
|
return top
|
|
|
|
def composite_layers(self, base_image, _properties_json="{}", **kwargs):
|
|
start_preview_server()
|
|
final_image = base_image.clone()
|
|
previews_data = {}
|
|
temp_dir = folder_paths.get_temp_directory()
|
|
B, base_H, base_W, C = base_image.shape
|
|
|
|
base_pil = tensor_to_pil(base_image)
|
|
base_filename = "layersys_base.png"
|
|
base_pil.save(os.path.join(temp_dir, base_filename))
|
|
previews_data["base_image"] = f"http://127.0.0.1:{PREVIEW_SERVER_PORT}/{base_filename}"
|
|
|
|
if final_image.shape[-1] == 4:
|
|
final_image = final_image[..., :3]
|
|
|
|
properties = json.loads(_properties_json)
|
|
layers = {k: v for k, v in kwargs.items() if k.startswith("layer_")}
|
|
masks = {k: v for k, v in kwargs.items() if k.startswith("mask_")}
|
|
sorted_layer_names = sorted(layers.keys())
|
|
|
|
for layer_name in sorted_layer_names:
|
|
layer_image_full = layers.get(layer_name)
|
|
if layer_image_full is None: continue
|
|
|
|
layer_pil = tensor_to_pil(layer_image_full)
|
|
layer_filename = f"layersys_{layer_name}.png"
|
|
layer_pil.save(os.path.join(temp_dir, layer_filename))
|
|
previews_data[layer_name] = f"http://127.0.0.1:{PREVIEW_SERVER_PORT}/{layer_filename}"
|
|
|
|
mask_name = layer_name.replace("layer_", "mask_")
|
|
mask = masks.get(mask_name)
|
|
|
|
if mask is not None:
|
|
mask_for_preview = mask
|
|
if mask_for_preview.dim() == 3:
|
|
mask_for_preview = mask_for_preview.unsqueeze(-1)
|
|
mask_for_preview_rgb = mask_for_preview.repeat(1, 1, 1, 3)
|
|
mask_pil = tensor_to_pil(mask_for_preview_rgb)
|
|
mask_filename = f"layersys_{mask_name}.png"
|
|
mask_pil.save(os.path.join(temp_dir, mask_filename))
|
|
previews_data[mask_name] = f"http://127.0.0.1:{PREVIEW_SERVER_PORT}/{mask_filename}"
|
|
|
|
props = properties.get(layer_name, {})
|
|
if not props.get("enabled", True): continue
|
|
|
|
resize_mode = props.get("resize_mode", "fit")
|
|
scale = props.get("scale", 1.0)
|
|
offset_x = props.get("offset_x", 0)
|
|
offset_y = props.get("offset_y", 0)
|
|
rotation = props.get("rotation", 0.0)
|
|
|
|
prepared_layer = None
|
|
content_alpha_mask = None
|
|
|
|
if resize_mode == 'crop':
|
|
pil_layer = tensor_to_pil(layer_image_full)
|
|
if pil_layer.mode != 'RGBA':
|
|
pil_layer = pil_layer.convert('RGBA')
|
|
|
|
new_w = int(pil_layer.width * scale)
|
|
new_h = int(pil_layer.height * scale)
|
|
if new_w > 0 and new_h > 0:
|
|
pil_layer = pil_layer.resize((new_w, new_h), Image.Resampling.BICUBIC)
|
|
|
|
if rotation != 0.0:
|
|
pil_layer = pil_layer.rotate(-rotation, resample=Image.Resampling.BICUBIC, expand=True)
|
|
|
|
final_canvas_pil = Image.new('RGBA', (base_W, base_H), (0, 0, 0, 0))
|
|
paste_x = (base_W // 2) + offset_x - (pil_layer.width // 2)
|
|
paste_y = (base_H // 2) + offset_y - (pil_layer.height // 2)
|
|
|
|
final_canvas_pil.paste(pil_layer, (paste_x, paste_y), pil_layer)
|
|
|
|
prepared_tensor = pil_to_tensor(final_canvas_pil)
|
|
prepared_layer = prepared_tensor[..., :3]
|
|
content_alpha_mask = prepared_tensor[..., 3:4]
|
|
else:
|
|
if layer_image_full.shape[-1] == 4:
|
|
layer_alpha_channel = layer_image_full[..., 3:4]
|
|
layer_image = layer_image_full[..., :3]
|
|
content_alpha_mask = prepare_layer(layer_alpha_channel, final_image, resize_mode, scale, offset_x, offset_y)
|
|
else:
|
|
layer_image = layer_image_full
|
|
|
|
prepared_layer = prepare_layer(layer_image, final_image, resize_mode, scale, offset_x, offset_y)
|
|
|
|
brightness = props.get("brightness", 0.0)
|
|
if brightness != 0.0: prepared_layer = torch.clamp(prepared_layer + brightness, 0.0, 1.0)
|
|
contrast = props.get("contrast", 0.0)
|
|
if contrast != 0.0:
|
|
contrast_factor = 1.0 + contrast
|
|
prepared_layer = torch.clamp((prepared_layer - 0.5) * contrast_factor + 0.5, 0.0, 1.0)
|
|
color_r, color_g, color_b = props.get("color_r", 1.0), props.get("color_g", 1.0), props.get("color_b", 1.0)
|
|
if color_r != 1.0 or color_g != 1.0 or color_b != 1.0:
|
|
prepared_layer[..., 0] = torch.clamp(prepared_layer[..., 0] * color_r, 0.0, 1.0)
|
|
prepared_layer[..., 1] = torch.clamp(prepared_layer[..., 1] * color_g, 0.0, 1.0)
|
|
prepared_layer[..., 2] = torch.clamp(prepared_layer[..., 2] * color_b, 0.0, 1.0)
|
|
saturation = props.get("saturation", 1.0)
|
|
if saturation != 1.0:
|
|
grayscale = prepared_layer[..., 0] * 0.299 + prepared_layer[..., 1] * 0.587 + prepared_layer[..., 2] * 0.114
|
|
grayscale = grayscale.unsqueeze(-1)
|
|
prepared_layer = torch.clamp(grayscale * (1.0 - saturation) + prepared_layer * saturation, 0.0, 1.0)
|
|
|
|
mode = props.get("blend_mode", "normal").replace('-', '_')
|
|
opacity = props.get("opacity", 1.0)
|
|
blended_image = self._blend(final_image, prepared_layer, mode)
|
|
|
|
# --- START OF THE CORRECTION ---
|
|
# `blended_image` contains the blended result, but also the blended black background.
|
|
# We fix this by compositing it onto the `final_image` using the layer's own content shape.
|
|
# This creates a correct "top layer" before applying the user mask.
|
|
if content_alpha_mask is None and resize_mode != 'stretch':
|
|
# If the layer had no alpha (e.g., a JPG), create a mask from its content.
|
|
content_alpha_mask = (prepared_layer.sum(dim=-1, keepdim=True) > 0.001).float()
|
|
|
|
if content_alpha_mask is not None:
|
|
# Ensure the mask is the correct size
|
|
if content_alpha_mask.shape[1:3] != final_image.shape[1:3]:
|
|
content_alpha_mask = F.interpolate(content_alpha_mask.permute(0, 3, 1, 2), size=(base_H, base_W), mode='bilinear', align_corners=False).permute(0, 2, 3, 1)
|
|
|
|
# Composite the blended result over the background using the content's alpha.
|
|
# This removes the black borders from `blended_image`.
|
|
blended_image = final_image * (1.0 - content_alpha_mask) + blended_image * content_alpha_mask
|
|
# --- END OF THE CORRECTION ---
|
|
|
|
final_mask = None # This will now only hold the final user-provided mask.
|
|
if mask is not None:
|
|
if mask.dim() == 3: mask = mask.unsqueeze(-1)
|
|
|
|
if resize_mode == 'crop':
|
|
pil_mask = tensor_to_pil(mask)
|
|
if pil_mask.mode != 'L': pil_mask = pil_mask.convert('L')
|
|
|
|
mask_w = int(pil_mask.width * scale)
|
|
mask_h = int(pil_mask.height * scale)
|
|
if mask_w > 0 and mask_h > 0:
|
|
pil_mask = pil_mask.resize((mask_w, mask_h), Image.Resampling.BICUBIC)
|
|
|
|
if rotation != 0.0:
|
|
pil_mask = pil_mask.rotate(-rotation, resample=Image.Resampling.BICUBIC, expand=True)
|
|
|
|
mask_canvas_pil = Image.new('L', (base_W, base_H), 0)
|
|
mask_paste_x = (base_W // 2) + offset_x - (pil_mask.width // 2)
|
|
mask_paste_y = (base_H // 2) + offset_y - (pil_mask.height // 2)
|
|
mask_canvas_pil.paste(pil_mask, (mask_paste_x, mask_paste_y))
|
|
final_mask = pil_to_tensor(mask_canvas_pil)
|
|
else:
|
|
final_mask = prepare_layer(mask, final_image, resize_mode, scale, offset_x, offset_y)
|
|
|
|
if final_mask is not None:
|
|
if final_mask.dim() == 3:
|
|
final_mask = final_mask.unsqueeze(-1)
|
|
|
|
if props.get("invert_mask", False):
|
|
final_mask = 1.0 - final_mask
|
|
|
|
if final_mask.shape[1:3] != final_image.shape[1:3]:
|
|
final_mask = F.interpolate(final_mask.permute(0, 3, 1, 2), size=(base_H, base_W), mode='bilinear', align_corners=False).permute(0, 2, 3, 1)
|
|
|
|
final_mask_with_opacity = final_mask * opacity
|
|
# `blended_image` is now pre-composited, so it acts as the "top" layer.
|
|
final_image = final_image * (1.0 - final_mask_with_opacity) + blended_image * final_mask_with_opacity
|
|
else:
|
|
# No user mask. `blended_image` is already the correct final state of the layer.
|
|
# We just need to apply the overall opacity.
|
|
final_image = (1.0 - opacity) * final_image + blended_image * opacity
|
|
|
|
return {
|
|
"result": (final_image,),
|
|
"ui": {
|
|
"layer_previews": [previews_data]
|
|
}
|
|
}
|
|
|
|
NODE_CLASS_MAPPINGS = { "LayerSystem": LayerSystem, }
|
|
NODE_DISPLAY_NAME_MAPPINGS = { "LayerSystem": "Layer System (Dynamic)" } |