新增提问分类器

web_search, retrieve_knowledge, generate_report

app/api/v1/zhipu_controller.py

@@ -104,4 +104,12 @@ def alltool_stream():
         for chunk in zhipu_alltool_service.web_search_sse(message):
             if chunk:
                 yield chunk
-    return Response(event_stream(), content_type='text/event-stream')
+    return Response(event_stream(), content_type='text/event-stream')
+
+@zhipu_controller.route('/zhipu/alltool/classify/non-stream', methods=['POST'])
+def alltool_classify_non_stream():
+    data = request.json
+    message = data.get('message', '')
+    response = zhipu_alltool_service.func_call_classify(message)
+    print(f'response: {response}')
+    return response

app/services/zhipu_alltool_service.py

@@ -10,60 +10,44 @@ class ZhipuAlltoolService:
         self.app_secret_key = "d54f764a1d67c17d857bd3983b772016.GRjowY0fyiMNurLc"
         logger.info("ZhipuAlltoolService initialized with model: %s", self.model_name)
 
-    def func_call(self, message):
+    def func_call_classify(self, message):
         logger.info("Starting web_search call")
         start_time = time.time()
         client = ZhipuAI(api_key=self.app_secret_key)
+        tools = [
+            {
+                "type": "function",
+                "function": {
+                    "name": "classify_user_input",
+                    "description": "根据用户输入，判断用户意图，返回意图类型",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {
+                            "category": {
+                                "type": "string",
+                                "description": "用户的意图有以下选项:web_search,retrive_knowledge,generate_report",
+                            }
+                        },
+                        "required": ["category"],
+                    },
+                }
+            }
+        ]
+        messages = [
+            {
+                "role": "user",
+                "content": f"判断以下用户输入的意图并将其分类返回意图类型：{message}"
+            }   
+        ]
         try:
             response = client.chat.completions.create(
-                model="glm-4-alltools",  # 填写需要调用的模型名称
-                messages=[
-                    {
-                        "role": "user",
-                        "content":[
-                            {
-                                "type":"text",
-                                "text":"帮我查询2018年至2024年，每年五一假期全国旅游出行数据，并绘制成柱状图展示数据趋势。"
-                            }
-                        ]
-                    }
-                ],
-                stream=True,
-                tools=[
-                {
-                    "type": "function",
-                    "function": {
-                        "name": "get_tourist_data_by_year",
-                        "description": "用于查询每一年的全国出行数据，输入年份范围(from_year,to_year)，返回对应的出行数据，包括总出行人次、分交通方式的人次等。",
-                        "parameters": {
-                            "type": "object",
-                            "properties": {
-                                "type": {
-                                    "description": "交通方式，默认为by_all，火车=by_train，飞机=by_plane，自驾=by_car",
-                                    "type": "string"
-                                },
-                                "from_year": {
-                                    "description": "开始年份，格式为yyyy",
-                                    "type": "string"
-                                },
-                                "to_year": {
-                                    "description": "结束年份，格式为yyyy",
-                                    "type": "string"
-                                }
-                            },
-                            "required": ["from_year","to_year"]
-                        }
-                    }
-                },
-                {
-                    "type": "code_interpreter"
-                }
-                ]
+                model="glm-4-flash",  # 填写需要调用的模型名称
+                messages= messages,
+                tools= tools,
+                tool_choice="auto"
             )
-
-            for chunk in response:
-                print(chunk)
-            return response
+            print(response)
+            return response.choices[0].message.tool_calls[0].function.arguments
         except Exception as e:
             logger.error("Error in web_search call: %s", str(e))
             raise e