Merge remote-tracking branch 'origin/main' into feature/upload-compatibility

app/service/chat/gemini_chat_service.py

@@ -50,6 +50,33 @@ def _extract_file_references(contents: List[Dict[str, Any]]) -> List[str]:
                 logger.info(f"Found file reference: {file_id}")
     return file_names
 
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
 def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
     """构建工具"""
     
@@ -62,7 +89,15 @@ def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
             for k, v in item.items():
                 if k == "functionDeclarations" and v and isinstance(v, list):
                     functions = record.get("functionDeclarations", [])
-                    functions.extend(v)
+                    # 清理每个函数声明中的不支持字段
+                    cleaned_functions = []
+                    for func in v:
+                        if isinstance(func, dict):
+                            cleaned_func = _clean_json_schema_properties(func)
+                            cleaned_functions.append(cleaned_func)
+                        else:
+                            cleaned_functions.append(func)
+                    functions.extend(cleaned_functions)
                     record["functionDeclarations"] = functions
                 else:
                     record[k] = v
@@ -136,6 +171,10 @@ def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
         "systemInstruction": request_dict.get("systemInstruction"),
     }
 
+    # 确保 generationConfig 不为 None
+    if payload["generationConfig"] is None:
+        payload["generationConfig"] = {}
+
     if model.endswith("-image") or model.endswith("-image-generation"):
         payload.pop("systemInstruction")
         payload["generationConfig"]["responseModalities"] = ["Text", "Image"]
@@ -153,7 +192,13 @@ def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
         if model.endswith("-non-thinking"):
             payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
         elif model in settings.THINKING_BUDGET_MAP:
-            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
+            if settings.SHOW_THINKING_PROCESS:
+                payload["generationConfig"]["thinkingConfig"] = {
+                    "thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000),
+                    "includeThoughts": True
+                }
+            else:
+                payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
 
     return payload
 

app/service/chat/openai_chat_service.py

@@ -26,16 +26,43 @@ from app.service.key.key_manager import KeyManager
 logger = get_openai_logger()
 
 
-def _has_media_parts(contents: List[Dict[str, Any]]) -> bool:
-    """判断消息是否包含图片、音频或视频部分 (inline_data)"""
-    for content in contents:
-        if content and "parts" in content and isinstance(content["parts"], list):
-            for part in content["parts"]:
-                if isinstance(part, dict) and "inline_data" in part:
+def _has_media_parts(messages: List[Dict[str, Any]]) -> bool:
+    """判断消息是否包含多媒体部分"""
+    for message in messages:
+        if "parts" in message:
+            for part in message["parts"]:
+                if "image_url" in part or "inline_data" in part:
                     return True
     return False
 
 
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
 def _build_tools(
     request: ChatRequest, messages: List[Dict[str, Any]]
 ) -> List[Dict[str, Any]]:
@@ -76,6 +103,8 @@ def _build_tools(
                 ):
                     function.pop("parameters", None)
 
+                # 清理函数中的不支持字段
+                function = _clean_json_schema_properties(function)
                 function_declarations.append(function)
 
         if function_declarations:
@@ -137,9 +166,13 @@ def _build_payload(
     if request.model.endswith("-non-thinking"):
         payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0}
     if request.model in settings.THINKING_BUDGET_MAP:
-        payload["generationConfig"]["thinkingConfig"] = {
-            "thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000)
-        }
+        if settings.SHOW_THINKING_PROCESS:
+            payload["generationConfig"]["thinkingConfig"] = {
+                "thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000),
+                "includeThoughts": True
+            }
+        else:
+            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000)}
 
     if (
         instruction

app/service/chat/vertex_express_chat_service.py

@@ -28,6 +28,33 @@ def _has_image_parts(contents: List[Dict[str, Any]]) -> bool:
     return False
 
 
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
 def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
     """构建工具"""
     
@@ -40,7 +67,15 @@ def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
             for k, v in item.items():
                 if k == "functionDeclarations" and v and isinstance(v, list):
                     functions = record.get("functionDeclarations", [])
-                    functions.extend(v)
+                    # 清理每个函数声明中的不支持字段
+                    cleaned_functions = []
+                    for func in v:
+                        if isinstance(func, dict):
+                            cleaned_func = _clean_json_schema_properties(func)
+                            cleaned_functions.append(cleaned_func)
+                        else:
+                            cleaned_functions.append(func)
+                    functions.extend(cleaned_functions)
                     record["functionDeclarations"] = functions
                 else:
                     record[k] = v
@@ -98,10 +133,26 @@ def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
         payload.pop("systemInstruction")
         payload["generationConfig"]["responseModalities"] = ["Text", "Image"]
         
-    if model.endswith("-non-thinking"):
-        payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
-    if model in settings.THINKING_BUDGET_MAP:
-        payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
+    # 处理思考配置：优先使用客户端提供的配置，否则使用默认配置
+    client_thinking_config = None
+    if request.generationConfig and request.generationConfig.thinkingConfig:
+        client_thinking_config = request.generationConfig.thinkingConfig
+    
+    if client_thinking_config is not None:
+        # 客户端提供了思考配置，直接使用
+        payload["generationConfig"]["thinkingConfig"] = client_thinking_config
+    else:
+        # 客户端没有提供思考配置，使用默认配置    
+        if model.endswith("-non-thinking"):
+            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
+        elif model in settings.THINKING_BUDGET_MAP:
+            if settings.SHOW_THINKING_PROCESS:
+                payload["generationConfig"]["thinkingConfig"] = {
+                    "thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000),
+                    "includeThoughts": True
+                }
+            else:
+                payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
 
     return payload