fix: patch parse_tool_call bug that returns string args instead of dict

Root cause identified: langchain_core's parse_tool_call() sometimes returns
tool_calls with 'args' as a JSON string instead of parsed dict object.

This violates AIMessage's Pydantic schema which expects args to be dict.

Solution: Wrapper now detects when parse_tool_call returns string args
and immediately converts them to dict using json.loads().

This is a workaround for what appears to be a LangChain bug where
parse_tool_call's json.loads() call either:
1. Fails silently without raising exception, or
2. Succeeds but result is not being assigned to args field

The fix ensures AIMessage always receives properly parsed dict args,
resolving Pydantic validation errors for all DeepSeek tool calls.

agent/chat_model_wrapper.py

@@ -32,22 +32,30 @@ class ToolCallArgsParsingWrapper:
             # Model doesn't have this method (e.g., MockChatModel), skip patching
             return
 
-        # CRITICAL: Also patch parse_tool_call to see what it's returning
-        from langchain_core.output_parsers import openai_tools
-        original_parse_tool_call = openai_tools.parse_tool_call
+        # CRITICAL: Patch parse_tool_call in base.py's namespace (not in openai_tools module!)
+        from langchain_openai.chat_models import base as langchain_base
+        original_parse_tool_call = langchain_base.parse_tool_call
 
         def patched_parse_tool_call(raw_tool_call, *, partial=False, strict=False, return_id=True):
-            """Patched parse_tool_call to log what it returns"""
+            """Patched parse_tool_call to fix string args bug and add logging"""
             result = original_parse_tool_call(raw_tool_call, partial=partial, strict=strict, return_id=return_id)
             if result:
                 args_type = type(result.get('args', None)).__name__
                 print(f"[DIAGNOSTIC] parse_tool_call returned: args type = {args_type}")
                 if args_type == 'str':
-                    print(f"[DIAGNOSTIC] ⚠️ BUG FOUND! parse_tool_call returned STRING args: {result['args']}")
+                    print(f"[DIAGNOSTIC] ⚠️ BUG FOUND! parse_tool_call returned STRING args, fixing...")
+                    # FIX: parse_tool_call sometimes returns string args instead of dict
+                    # This happens when it fails to parse but doesn't raise an exception
+                    try:
+                        result['args'] = json.loads(result['args'])
+                        print(f"[DIAGNOSTIC] ✓ Fixed! Converted string args to dict")
+                    except (json.JSONDecodeError, TypeError) as e:
+                        print(f"[DIAGNOSTIC] ❌ Failed to parse args: {e}")
+                        # Leave as string if we can't parse it
             return result
 
-        # Replace globally
-        openai_tools.parse_tool_call = patched_parse_tool_call
+        # Replace in base.py's namespace (where _convert_dict_to_message uses it)
+        langchain_base.parse_tool_call = patched_parse_tool_call
 
         original_create_chat_result = self.wrapped_model._create_chat_result