improve mcp docstring descriptions

app.py

@@ -179,23 +179,7 @@ def ocr_demo(image, task, ocr_type, ocr_box, ocr_color):
     Main OCR demonstration function that processes images and returns results.
     
     Args:
-        image (Union[dict, np.ndarray, str, PIL.Image]): Input image in one of these formats:
-            - dict: Image component state with keys:
-                - path: str | None (Path to local file)
-                - url: str | None (Public URL or base64 image)
-                - size: int | None (Image size in bytes)
-                - orig_name: str | None (Original filename)
-                - mime_type: str | None (Image MIME type)
-                - is_stream: bool (Always False)
-                - meta: dict(str, Any)
-            - dict: ImageEditor component state with keys:
-                - background: filepath | None
-                - layers: list[filepath]
-                - composite: filepath | None
-                - id: str | None
-            - np.ndarray: Raw image array
-            - str: Path to image file
-            - PIL.Image: PIL Image object
+        image (Union[dict, np.ndarray, str, PIL.Image]): Input image in one of these formats: Image component state with keys: path: str | None (Path to local file) url: str | None (Public URL or base64 image) size: int | None (Image size in bytes) orig_name: str | None (Original filename) mime_type: str | None (Image MIME type) is_stream: bool (Always False) meta: dict(str, Any) OR  dict: ImageEditor component state with keys: background: filepath | None layers: list[filepath] composite: filepath | None id: str | None OR np.ndarray: Raw image array str: Path to image file PIL.Image: PIL Image object
         task (Literal['Plain Text OCR', 'Format Text OCR', 'Fine-grained OCR (Box)', 
               'Fine-grained OCR (Color)', 'Multi-crop OCR', 'Render Formatted OCR']): 
             Selected OCR task type