新增 ImageWorkflowExtractor 類別，實現圖像資訊提取與轉換功能

__init__.py

@@ -3,6 +3,7 @@ import folder_paths
 from .loader_node import DynamicTagLoaderJS
 from .saver_node import DynamicTagSaver
 from .iterator_node import DynamicTagIterator
+from .image_info_node import ImageWorkflowExtractor
 
 # ==============================================================================
 # 模組導入與環境檢查
@@ -74,13 +75,15 @@ if PromptServer:
 NODE_CLASS_MAPPINGS = {
     "DynamicTagLoaderJS": DynamicTagLoaderJS,
     "DynamicTagSaver": DynamicTagSaver,
-    "DynamicTagIterator": DynamicTagIterator
+    "DynamicTagIterator": DynamicTagIterator,
+    "WorkflowMetadataReader": ImageWorkflowExtractor,
 }
 
 NODE_DISPLAY_NAME_MAPPINGS = {
     "DynamicTagLoaderJS": "⚡Dynamic Tag Loader",
     "DynamicTagSaver": "💾 Dynamic Tag Saver",
-    "DynamicTagIterator": "🔄 Dynamic Tag Iterator"
+    "DynamicTagIterator": "🔄 Dynamic Tag Iterator",
+    "WorkflowMetadataReader": "🔍 Workflow Metadata Reader",
 }
 
 WEB_DIRECTORY = "./web"

image_info_node.py

@@ -0,0 +1,97 @@
+import os
+import json
+import torch
+import numpy as np
+from PIL import Image, ImageOps
+
+class ImageWorkflowExtractor:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "image_or_dir": ("STRING", {"default": "C:/ComfyUI/output"}),
+                "search_by": (["ID", "Type"], {"default": "ID"}),
+                "search_query": ("STRING", {"default": "71"}),
+                "seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+            },
+        }
+
+    # 修改輸出類型：將原本的 raw_info (STRING) 改為 image (IMAGE)
+    RETURN_TYPES = ("JSON", "STRING", "STRING", "IMAGE")
+    RETURN_NAMES = ("workflow_json", "clean_text", "selected_path", "image")
+    FUNCTION = "extract_info"
+    CATEGORY = "DynamicTags"
+
+    def extract_info(self, image_or_dir, search_by, search_query, seed):
+        target_path = image_or_dir.strip()
+        selected_file = ""
+
+        # ==========================================
+        # 1. 檔案選取邏輯
+        # ==========================================
+        if os.path.isfile(target_path):
+            selected_file = target_path
+        elif os.path.isdir(target_path):
+            valid_extensions = ('.png', '.webp', '.jpg', '.jpeg')
+            files = sorted([
+                os.path.join(target_path, f) 
+                for f in os.listdir(target_path) 
+                if f.lower().endswith(valid_extensions)
+            ])
+
+            if not files:
+                # 若無檔案，回傳一個空的黑色張量避免系統崩潰
+                return ({}, "No images found", target_path, torch.zeros((1, 64, 64, 3)))
+
+            index = seed % len(files)
+            selected_file = files[index]
+        else:
+            return ({}, "Invalid path", target_path, torch.zeros((1, 64, 64, 3)))
+
+        # ==========================================
+        # 2. 資訊提取與圖像轉換輸出
+        # ==========================================
+        try:
+            with Image.open(selected_file) as img:
+                # --- A. 提取 Workflow Metadata ---
+                info = img.info
+                workflow = info.get("workflow", "{}")
+                wf_data = json.loads(workflow) if isinstance(workflow, str) else workflow
+                nodes = wf_data.get("nodes", [])
+
+                clean_results = []
+                for node in nodes:
+                    match = False
+                    if search_by == "ID":
+                        if str(node.get("id")) == search_query.strip():
+                            match = True
+                    else:
+                        if node.get("type") == search_query.strip():
+                            match = True
+
+                    if match:
+                        values = node.get("widgets_values", [])
+                        for val in values:
+                            if isinstance(val, list):
+                                for sub_val in val:
+                                    clean_results.append(str(sub_val))
+                            else:
+                                clean_results.append(str(val))
+
+                final_text = "\n".join(clean_results) if clean_results else f"No nodes match {search_query}"
+                
+                # --- B. 將圖片轉為 ComfyUI 格式 (IMAGE Tensor) ---
+                # 修正圖片轉向 (Exif 資訊)
+                img = ImageOps.exif_transpose(img)
+                # 統一轉為 RGB
+                image_rgb = img.convert("RGB")
+                # 轉為 numpy 陣列並正規化至 0.0 ~ 1.0
+                image_np = np.array(image_rgb).astype(np.float32) / 255.0
+                # 轉為 PyTorch Tensor 並調整維度為 [Batch, Height, Width, Channel]
+                image_tensor = torch.from_numpy(image_np)[None,]
+
+                return (wf_data, final_text, selected_file, image_tensor)
+        
+        except Exception as e:
+            # 發生錯誤時回傳錯誤訊息與空圖片
+            return ({}, f"Error: {str(e)}", selected_file, torch.zeros((1, 64, 64, 3)))