Merge pull request #1754 from better629/main

simplify log if messages contain base64 data

Author: better629

metagpt/provider/base_llm.py

@@ -142,6 +142,38 @@ def get_costs(self) -> Costs:
             return Costs(0, 0, 0, 0)
         return self.cost_manager.get_costs()
 
+    def mask_base64_data(self, msg: dict) -> dict:
+        """Process the base64 image data in the message, replacing it with placeholders for easier logging
+
+        Args:
+            msg (dict): A dictionary of messages in OpenAI format
+
+        Returns:
+            dict: This is the processed message dictionary with the image data replaced with placeholders
+        """
+        if not isinstance(msg, dict):
+            return msg
+
+        new_msg = msg.copy()
+        content = new_msg.get("content")
+        img_base64_prefix = "data:image/"
+
+        if isinstance(content, list):
+            # Handling multimodal content (like gpt-4v format)
+            new_content = []
+            for item in content:
+                if isinstance(item, dict) and item.get("type") == "image_url":
+                    image_url = item.get("image_url", {}).get("url", "")
+                    if image_url.startswith(img_base64_prefix):
+                        item = item.copy()
+                        item["image_url"] = {"url": "<Image base64 data has been omitted>"}
+                new_content.append(item)
+            new_msg["content"] = new_content
+        elif isinstance(content, str) and img_base64_prefix in content:
+            # Process plain text messages containing base64 image data
+            new_msg["content"] = "<Messages containing image base64 data have been omitted>"
+        return new_msg
+
     async def aask(
         self,
         msg: Union[str, list[dict[str, str]]],
@@ -165,7 +197,11 @@ async def aask(
             message.extend(msg)
         if stream is None:
             stream = self.config.stream
-        logger.debug(message)
+
+        # the image data is replaced with placeholders to avoid long output
+        masked_message = [self.mask_base64_data(m) for m in message]
+        logger.debug(masked_message)
+
         compressed_message = self.compress_messages(message, compress_type=self.config.compress_type)
         rsp = await self.acompletion_text(compressed_message, stream=stream, timeout=self.get_timeout(timeout))
         # rsp = await self.acompletion_text(message, stream=stream, timeout=self.get_timeout(timeout))

metagpt/provider/constant.py

@@ -34,4 +34,12 @@
 MULTI_MODAL_MODELS = [
     "gpt-4o",
     "gpt-4o-mini",
+    "openai/gpt-4o",
+    "gemini-2.0-flash-exp",
+    "gemini-2.0-pro-exp-02-05",
+    "claude-3-5-sonnet-v2",
+    "google/gemini-2.0-flash-exp:free",
+    "google/gemini-2.0-pro-exp-02-05:free",
+    "anthropic/claude-3.5-sonnet",
+    "anthropic/claude-3.7-sonnet",
 ]