Update pyproject.toml

route change to avoid conflicts

.github/workflows/publish_action.yml

@@ -7,14 +7,19 @@ on:
     paths:
       - "pyproject.toml"
 
+permissions:
+  issues: write
+
 jobs:
   publish-node:
     name: Publish Custom Node to registry
     runs-on: ubuntu-latest
+    if: ${{ github.repository_owner == 'if-ai' }}
     steps:
       - name: Check out code
         uses: actions/checkout@v4
       - name: Publish Custom Node
-        uses: Comfy-Org/publish-node-action@main
+        uses: Comfy-Org/publish-node-action@v1
         with:
-          personal_access_token: ${{ secrets.REGISTRY_ACCESS_TOKEN }} ## Add your own personal access token to your Github Repository secrets and reference it here.
+          ## Add your own personal access token to your Github Repository secrets and reference it here.
+          personal_access_token: ${{ secrets.REGISTRY_ACCESS_TOKEN }}

IFLoadImagesNodeS.py

@@ -377,16 +377,17 @@ class ImageManager:
             return sorted(files)
 
 class IFLoadImagess:
+    _color_channels = ["alpha", "red", "green", "blue"]
+    
     def __init__(self):
-        self.path_cache = {}  # Cache for path mapping
+        self.path_cache = {}
         
     @classmethod
     def INPUT_TYPES(s):
         input_dir = folder_paths.get_input_directory()
-        # Count available thumbnails
         available_images = len([f for f in os.listdir(input_dir) 
                             if f.startswith(ImageManager.THUMBNAIL_PREFIX)])
-        available_images = max(1, available_images)  # Ensure at least 1
+        available_images = max(1, available_images)
         
         files = [f for f in os.listdir(input_dir) 
                 if f.startswith(ImageManager.THUMBNAIL_PREFIX)]
@@ -396,7 +397,7 @@ class IFLoadImagess:
                 "image": (sorted(files), {"image_upload": True}),
                 "input_path": ("STRING", {"default": ""}),
                 "start_index": ("INT", {"default": 0, "min": 0, "max": 9999}),
-                "stop_index": ("INT", {"default": 10, "min": 1, "max": 9999}),  # Changed to stop_index
+                "stop_index": ("INT", {"default": 10, "min": 1, "max": 9999}),
                 "load_limit": (["10", "100", "1000", "10000", "100000"], {"default": "1000"}),
                 "image_selected": ("BOOLEAN", {"default": False}),
                 "available_image_count": ("INT", {
@@ -408,19 +409,21 @@ class IFLoadImagess:
                 "include_subfolders": ("BOOLEAN", {"default": True}),
                 "sort_method": (["alphabetical", "numerical", "date_created", "date_modified"],),
                 "filter_type": (["none", "png", "jpg", "jpeg", "webp", "gif", "bmp"],),
+                "channel": (s._color_channels, {"default": "alpha"}),
             }
         }
 
-    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "STRING", "STRING", "INT")
-    RETURN_NAMES = ("images", "masks", "image_paths", "filenames", "count_str", "count_int")
-    OUTPUT_IS_LIST = (True, True, True, True, True, True)
+    RETURN_TYPES = ("IMAGE", "MASK", "STRING", "STRING", "STRING", "INT", "IMAGE", "MASK")
+    RETURN_NAMES = ("images", "masks", "image_paths", "filenames", "count_str", "count_int", "images_batch", "masks_batch")
+    OUTPUT_IS_LIST = (True, True, True, True, True, True, False, False)
     FUNCTION = "load_images"
-    CATEGORY = "ImpactFrames💥🎞️"
+    CATEGORY = "ImpactFrames💥🎞️/images"
 
     @classmethod
-    def IS_CHANGED(cls, image, input_path="", start_index=0, max_images=1,
-                include_subfolders=True, sort_method="numerical", 
-                filter_type="none", image_name="", unique_id=None):
+    def IS_CHANGED(cls, image, input_path="", start_index=0, stop_index=0, max_images=1,
+                include_subfolders=True, sort_method="numerical", image_selected=False, 
+                filter_type="none", image_name="", unique_id=None, load_limit="1000", 
+                available_image_count=0, channel="alpha"  ):
         """
         Properly handle all input parameters and return NaN to force updates
         This matches the input parameters from INPUT_TYPES
@@ -442,20 +445,21 @@ class IFLoadImagess:
             return float("NaN")
 
     def load_images(self, image="", input_path="", start_index=0, stop_index=10,
-               load_limit="1000", image_selected=False, available_image_count=0, 
-               include_subfolders=True, sort_method="numerical", 
-               filter_type="none", image_name="", unique_id=None):
+                   load_limit="1000", image_selected=False, available_image_count=0, 
+                   include_subfolders=True, sort_method="numerical", 
+                   filter_type="none", channel="alpha"):
         try:
             # Process input path
             abs_path = os.path.abspath(input_path if os.path.isabs(input_path) 
                                     else os.path.join(folder_paths.get_input_directory(), input_path))
                     
-            # Get all valid images first
+            # Get all valid images
             all_files = ImageManager.get_image_files(abs_path, include_subfolders, filter_type)
             if not all_files:
                 logger.warning(f"No valid images found in {abs_path}")
                 img_tensor, mask = self.load_placeholder()
-                return ([img_tensor], [mask], [""], [""], ["0/0"], [0])
+                return ([img_tensor], [mask], [""], [""], ["0/0"], [0], 
+                       img_tensor.unsqueeze(0), mask.unsqueeze(0))
 
             # Sort files
             all_files = ImageManager.sort_files(all_files, sort_method)
@@ -478,49 +482,80 @@ class IFLoadImagess:
                 start_index = image_order[image]
                 num_images = 1
 
-            # Create path mapping
-            self.path_cache = {
-                thumb: orig for thumb, orig in zip(all_thumbnails, all_files[start_index:start_index + num_images])
-            }
-            
             # Process selected range
             selected_files = all_files[start_index:start_index + num_images]
-            selected_thumbnails = all_thumbnails[:num_images]  
-                
-            # Process selected files
+            
+            # Lists to store outputs
             images = []
             masks = []
             paths = []
             filenames = []
             count_strs = []
             count_ints = []
-
-            for idx, (file_path, thumb_name) in enumerate(zip(selected_files, selected_thumbnails)):
+            
+            # Track max dimensions for resizing
+            max_height = 0
+            max_width = 0
+            
+            # First pass to determine max dimensions
+            for file_path in selected_files:
                 try:
                     with Image.open(file_path) as img:
                         img = ImageOps.exif_transpose(img)
+                        max_height = max(max_height, img.height)
+                        max_width = max(max_width, img.width)
+                except Exception as e:
+                    logger.error(f"Error checking dimensions of {file_path}: {e}")
+                    continue
+
+            # Second pass to load and resize images
+            for idx, file_path in enumerate(selected_files):
+                try:
+                    img = Image.open(file_path)
+                    img = ImageOps.exif_transpose(img)
+                    
+                    if img.mode == 'I':
+                        img = img.point(lambda i: i * (1 / 255))
+                    
+                    # Resize to match max dimensions
+                    image = img.convert('RGB')
+                    if image.size != (max_width, max_height):
+                        image = image.resize((max_width, max_height), Image.Resampling.LANCZOS)
+                    
+                    # Convert to numpy array and normalize
+                    image_array = np.array(image).astype(np.float32) / 255.0
+                    image_tensor = torch.from_numpy(image_array).unsqueeze(0)  # [1, H, W, 3]
+                    images.append(image_tensor)
+                    
+                    # Handle mask based on selected channel
+                    if img.mode not in ('RGB', 'L'):
+                        img = img.convert('RGBA')
                         
-                        if img.mode == 'I':
-                            img = img.point(lambda i: i * (1 / 255))
-                        image = img.convert('RGB')
-                        
-                        image_array = np.array(image).astype(np.float32) / 255.0
-                        image_tensor = torch.from_numpy(image_array)[None,]
-                        
-                        if 'A' in img.getbands():
-                            mask = np.array(img.getchannel('A')).astype(np.float32) / 255.0
-                            mask = 1. - torch.from_numpy(mask)
-                        else:
-                            mask = torch.zeros((image_array.shape[0], image_array.shape[1]), 
-                                        dtype=torch.float32, device="cpu")
-                        
-                        images.append(image_tensor)
-                        masks.append(mask.unsqueeze(0))
-                        paths.append(file_path)
-                        filenames.append(os.path.basename(file_path))
-                        count_str = f"{start_index + idx + 1}/{total_files}"  # Update count to show global position
-                        count_strs.append(count_str)
-                        count_ints.append(start_index + idx + 1)
+                    c = channel[0].upper()
+                    if c in img.getbands():
+                        mask = np.array(img.getchannel(c)).astype(np.float32) / 255.0
+                        mask = torch.from_numpy(mask)
+                        if c == 'A':
+                            mask = 1. - mask
+                    else:
+                        mask = torch.zeros((max_height, max_width), 
+                                       dtype=torch.float32, device="cpu")
+                    
+                    # Resize mask if needed
+                    if mask.shape != (max_height, max_width):
+                        mask = torch.nn.functional.interpolate(
+                            mask.unsqueeze(0).unsqueeze(0),
+                            size=(max_height, max_width),
+                            mode='bilinear',
+                            align_corners=False
+                        ).squeeze(0).squeeze(0)
+                    
+                    masks.append(mask.unsqueeze(0))  # Add batch dimension to mask [1, H, W]
+                    
+                    paths.append(file_path)
+                    filenames.append(os.path.basename(file_path))
+                    count_strs.append(f"{start_index + idx + 1}/{total_files}")
+                    count_ints.append(start_index + idx + 1)
 
                 except Exception as e:
                     logger.error(f"Error processing image {file_path}: {e}")
@@ -528,36 +563,29 @@ class IFLoadImagess:
 
             if not images:
                 img_tensor, mask = self.load_placeholder()
-                return ([img_tensor], [mask], [""], [""], ["0/0"], [0])
+                return ([img_tensor], [mask], [""], [""], ["0/0"], [0], 
+                       img_tensor.unsqueeze(0), mask.unsqueeze(0))
 
-            ui_data = {
-                "images": all_thumbnails,
-                "current_thumbnails": selected_thumbnails,
-                "total_images": total_files,
-                "path_mapping": self.path_cache,
-                "available_image_count": total_files,
-                "image_order": image_order,
-                "start_index": start_index,
-                "stop_index": start_index + num_images
-            }
-            
-            return {
-                "ui": {"values": ui_data},
-                "result": (images, masks, paths, filenames, count_strs, count_ints)
-            }
+            # Create batched version - now all images are the same size
+            images_batch = torch.cat(images, dim=0)  # [B, H, W, 3]
+            masks_batch = torch.cat(masks, dim=0)    # [B, H, W]
+
+            return (images, masks, paths, filenames, count_strs, count_ints,
+                   images_batch, masks_batch)
 
         except Exception as e:
             logger.error(f"Error in load_images: {e}", exc_info=True)
             img_tensor, mask = self.load_placeholder()
-            return ([img_tensor], [mask], [""], [""], ["error"], [0])
+            return ([img_tensor], [mask], [""], [""], ["error"], [0],
+                   img_tensor.unsqueeze(0), mask.unsqueeze(0))
 
     def load_placeholder(self):
         """Creates and returns a placeholder image tensor and mask"""
         img = Image.new('RGB', (512, 512), color=(73, 109, 137))
         image_array = np.array(img).astype(np.float32) / 255.0
-        image_tensor = torch.from_numpy(image_array)[None,]
-        mask = torch.zeros((1, image_array.shape[0], image_array.shape[1]), 
-                       dtype=torch.float32, device="cpu")
+        image_tensor = torch.from_numpy(image_array)  # [H, W, 3]
+        mask = torch.zeros((image_array.shape[0], image_array.shape[1]), 
+                       dtype=torch.float32, device="cpu")  # [H, W]
         return image_tensor, mask
 
     def process_single_image(self, image_path: str):
@@ -588,7 +616,7 @@ class IFLoadImagess:
             img_tensor, mask = self.load_placeholder()
             return ([img_tensor], [mask], [""], [""], ["error"], [0])
     
-@PromptServer.instance.routes.post("/ifai/backup_input")
+@PromptServer.instance.routes.post("/IF_img/backup_input")
 async def backup_input_folder(request):
     try:
         success, message = ImageManager.backup_input_folder()
@@ -603,7 +631,7 @@ async def backup_input_folder(request):
             "error": str(e)
         }, status=500)
 
-@PromptServer.instance.routes.post("/ifai/restore_input")
+@PromptServer.instance.routes.post("/IF_img/restore_input")
 async def restore_input_folder(request):
     try:
         success, message = ImageManager.restore_input_folder()
@@ -618,7 +646,7 @@ async def restore_input_folder(request):
             "error": str(e)
         }, status=500)
 
-@PromptServer.instance.routes.post("/ifai/refresh_previews")
+@PromptServer.instance.routes.post("/IF_img/refresh_previews")
 async def refresh_previews(request):
     try:
         data = await request.json()
@@ -691,7 +719,7 @@ async def refresh_previews(request):
         }, status=500)
 
 # Add route for widget refresh
-@PromptServer.instance.routes.post("/ifai/refresh_widgets")
+@PromptServer.instance.routes.post("/IF_img/refresh_widgets")
 async def refresh_widgets(request):
     try:
         input_dir = folder_paths.get_input_directory()
@@ -716,4 +744,4 @@ NODE_CLASS_MAPPINGS = {
 
 NODE_DISPLAY_NAME_MAPPINGS = {
     "IF_LoadImagesS": "IF Load Images S 🖼️"
-}
+}

README.md

@@ -1,3 +1,7 @@
+Here’s the revised README section with the video tutorial link added:
+
+---
+
 # ComfyUI_IF_AI_LoadImages
 
 This tool enables you to load images from arbitrary folder selections, display previews of the images within subfolders, and output a list of images so you can work with multiple images simultaneously.
@@ -14,16 +18,20 @@ This tool enables you to load images from arbitrary folder selections, display p
 8. To use a single selected image, switch the option to true and queue the node (execute the workflow).
 
 > **Note:** This is a workaround solution, so it may feel a bit clunky but functions effectively.
-> 
+
+[![Tutorial Video](https://img.youtube.com/vi/6ylkSoJ-Tnw/0.jpg)](https://www.youtube.com/watch?v=6ylkSoJ-Tnw)  
+*[Watch the video tutorial here](https://www.youtube.com/watch?v=6ylkSoJ-Tnw)*
+
 ![thorium_WAUJJwgYI6](https://github.com/user-attachments/assets/c3f1a31a-9d11-4e6d-b619-bb0226760e05)
 
 ![Image Preview 1](https://github.com/user-attachments/assets/55c67132-f7f5-4755-afef-7f9d5679c1d0)
 
-
 ## TODO
-
-- [ ] Add a counter to enable loops.
-- [ ] Add support for video files.
+- [x] Fix image extension filters
+- [ ] Fix Masks from mask editor
+- [ ] Fix single image upload
+- [ ] Add a counter to enable loops
+- [ ] Add support for video files
 
 ## Support
 

__init__.py

@@ -1,25 +1,25 @@
-import os
-import glob
-import shutil
-import sys
-import folder_paths
-
-from .IFLoadImagesNodeS import IFLoadImagess
-
-
-
-
-NODE_CLASS_MAPPINGS = {
-    "IF_LoadImagesS": IFLoadImagess,
-}
-
-NODE_DISPLAY_NAME_MAPPINGS = {
-    "IF_LoadImagesS": "IF Load Images S 🖼️",
-}
-
-WEB_DIRECTORY = "./web"
-__all__ = [
-    "NODE_CLASS_MAPPINGS", 
-    "NODE_DISPLAY_NAME_MAPPINGS", 
-    "WEB_DIRECTORY", 
-    ]
+import os
+import glob
+import shutil
+import sys
+import folder_paths
+
+from .IFLoadImagesNodeS import IFLoadImagess
+
+
+
+
+NODE_CLASS_MAPPINGS = {
+    "IF_LoadImagesS": IFLoadImagess,
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "IF_LoadImagesS": "IF Load Images S 🖼️",
+}
+
+WEB_DIRECTORY = "./web"
+__all__ = [
+    "NODE_CLASS_MAPPINGS", 
+    "NODE_DISPLAY_NAME_MAPPINGS", 
+    "WEB_DIRECTORY", 
+    ]

pyproject.toml

@@ -1,9 +1,9 @@
 [project]
 name = "comfyui_if_ai_loadimages"
 description = "It Load Images with subfolders form arbitrary folders previous on node outputs lists- convinient selection via file browser"
-version = "1.0.1"
-license = { file = "LICENSE.txt" } 
-dependencies = ["pillow"]
+version = "1.0.7"
+license = { file = "MIT License" }
+dependencies = ["pillow", "numpy"]
 
 [project.urls]
 Repository = "https://github.com/if-ai/ComfyUI_IF_AI_LoadImages"
@@ -11,5 +11,5 @@ Repository = "https://github.com/if-ai/ComfyUI_IF_AI_LoadImages"
 
 [tool.comfy]
 PublisherId = "impactframes"
-DisplayName = "ComfyUI_IF_LoadImages"
-Icon = ""
+DisplayName = "IF_LoadImages"
+Icon = "https://impactframes.ai/System/Icons/48x48/if.png"

requirements.txt

@@ -1,2 +1,2 @@
-
-
+pillow
+numpy

web/js/IFLoadImagesNodeS.js

@@ -170,7 +170,7 @@ app.registerExtension({
             const backupBtn = this.addWidget("button", "backup_input", "Backup Input 💾", 
                 async () => {
                     try {
-                        const response = await api.fetchApi("/if_ai/backup_input", {
+                        const response = await api.fetchApi("/IF_img/backup_input", {
                             method: "POST"
                         });
                         
@@ -192,7 +192,7 @@ app.registerExtension({
             const restoreBtn = this.addWidget("button", "restore_input", "Restore Input ♻️", 
                 async () => {
                     try {
-                        const response = await api.fetchApi("/if_ai/restore_input", {
+                        const response = await api.fetchApi("/IF_img/restore_input", {
                             method: "POST"
                         });
                         
@@ -245,7 +245,7 @@ app.registerExtension({
                         load_limit: parseInt(this.widgets.find(w => w.name === "load_limit")?.value || "1000")
                     };
             
-                    const response = await api.fetchApi("/if_ai/refresh_previews", {
+                    const response = await api.fetchApi("/IF_img/refresh_previews", {
                         method: "POST",
                         headers: { "Content-Type": "application/json" },
                         body: JSON.stringify(options)
@@ -342,7 +342,7 @@ app.registerExtension({
         nodeType.prototype.backupInputFolder = async function() {
             try {
                 this.showLoader();
-                const response = await fetch("/if_ai/backup_input", {
+                const response = await fetch("/IF_img/backup_input", {
                     method: "POST"
                 });
                 
@@ -365,7 +365,7 @@ app.registerExtension({
         nodeType.prototype.restoreInputFolder = async function() {
             try {
                 this.showLoader();
-                const response = await fetch("/if_ai/restore_input", {
+                const response = await fetch("/IF_img/restore_input", {
                     method: "POST"
                 });
                 
@@ -400,7 +400,7 @@ app.registerExtension({
 
                 this.showLoader();
 
-                const response = await fetch("/if_ai/refresh_previews", {
+                const response = await fetch("/IF_img/refresh_previews", {
                     method: "POST",
                     headers: {
                         "Content-Type": "application/json"