代码示例

通过设置 tool_stream=True 参数，可以启用流式工具调用功能： Python SDK 安装 SDK # 安装最新版本 pip install zai-sdk # 或指定版本 pip install zai-sdk== 0.2.3 验证安装 import zai print (zai. __version__ ) 完整示例 from zai import ZhipuAiClient # 初始化客户端 client = ZhipuAiClient( api_key = 'YOUR_API_KEY' ) # 创建流式工具调用请求 response = client.chat.completions.create( model = "glm-5.3" , # 使用支持工具调用的模型 messages = [ { "role" : "user" , "content" : "北京天气怎么样" }, ], tools = [ { "type" : "function" , "function" : { "name" : "get_weather" , "description" : "获取指定地点当前的天气情况" , "parameters" : { "type" : "object" , "properties" : { "location" : { "type" : "string" , "description" : "城市，例如：北京、上海" }, "unit" : { "type" : "string" , "enum" : [ "celsius" , "fahrenheit" ]} }, "required" : [ "location" ] } } } ], stream = True , # 启用流式输出 tool_stream = True # 启用工具调用流式输出 ) # 初始化变量用于收集流式数据 reasoning_content = "" # 推理过程内容 content = "" # 回答内容 final_tool_calls = {} # 工具调用信息 reasoning_started = False # 推理过程开始标志 content_started = False # 内容输出开始标志 # 处理流式响应 for chunk in response: if not chunk.choices: continue delta = chunk.choices[ 0 ].delta # 处理流式推理过程输出 if hasattr (delta, 'reasoning_content' ) and delta.reasoning_content: if not reasoning_started and delta.reasoning_content.strip(): print ( " \n 🧠 思考过程：" ) reasoning_started = True reasoning_content += delta.reasoning_content print (delta.reasoning_content, end = "" , flush = True ) # 处理流式回答内容输出 if hasattr (delta, 'content' ) and delta.content: if not content_started and delta.content.strip(): print ( " \n\n 💬 回答内容：" ) content_started = True content += delta.content print (delta.content, end = "" , flush = True ) # 处理流式工具调用信息 if delta.tool_calls: for tool_call in delta.tool_calls: index = tool_call.index if index not in final_tool_calls: # 新的工具调用 final_tool_calls[index] = tool_call final_tool_calls[index].function.arguments = tool_call.function.arguments else : # 追加工具调用参数（流式构建） final_tool_calls[index].function.arguments += tool_call.function.arguments # 输出最终的工具调用信息 if final_tool_calls: print ( " \n 📋 命中 Function Calls :" ) for index, tool_call in final_tool_calls.items(): print ( f " { index } : 函数名: { tool_call.function.name } , 参数: { tool_call.function.arguments } " )