Update tesseract/mcp_output/mcp_plugin/mcp_service.py

tesseract/mcp_output/mcp_plugin/mcp_service.py

@@ -7,6 +7,7 @@ sys.path.insert(0, source_path)
 from fastmcp import FastMCP
 from PIL import Image
 import pytesseract
+import io
 
 mcp = FastMCP("tesseract_service")
 
@@ -51,24 +52,34 @@ def set_image(image_data: bytes, width: int, height: int) -> dict:
         return {"success": False, "result": None, "error": str(e)}
 
 @mcp.tool(name="recognize_text", description="Perform OCR on the provided image and return recognized text.")
-def recognize_text(image_path: str) -> dict:
+def recognize_text(image_path: str = None, image_base64: str = None) -> dict:
     """
     Perform OCR on the provided image and return recognized text.
 
     Args:
         image_path (str): Path to the image to recognize.
+        image_base64 (str): Base64 encoded image data.
 
     Returns:
         dict: A dictionary containing recognized text or error message.
     """
     try:
-        # 检查图片路径是否存在
-        if not os.path.exists(image_path):
-            return {"success": False, "result": None, "error": f"Image not found: {image_path}"}
-
-        # 使用 PIL 打开图片
-        image = Image.open(image_path)
-        print(f"Processing image: {image_path}, size: {image.size}")
+        # 处理base64数据或文件路径
+        if image_base64:
+            # 从base64数据创建图片
+            import base64
+            image_data = base64.b64decode(image_base64)
+            image = Image.open(io.BytesIO(image_data))
+            print(f"Processing base64 image, size: {image.size}")
+        elif image_path:
+            # 检查图片路径是否存在
+            if not os.path.exists(image_path):
+                return {"success": False, "result": None, "error": f"Image not found: {image_path}"}
+            # 使用 PIL 打开图片
+            image = Image.open(image_path)
+            print(f"Processing image: {image_path}, size: {image.size}")
+        else:
+            return {"success": False, "result": None, "error": "Either image_path or image_base64 must be provided"}
 
         # 尝试多种语言配置进行OCR识别
         recognized_text = ""