agent_framework.py

# agent_framework.py —— 可复用 Agent 框架
import os, json
from openai import OpenAI

# ================================================================
#  工具函数（业务代码，和框架无关）
# ================================================================
def get_weather(city):
    data = {
        "北京": "晴, 25°C", "上海": "小雨, 22°C", "广州": "多云, 30°C"
    }
    return data.get(city, f"{city}天气未收录")

def get_order(order_id):
    orders = {"A001": "iPhone 16, 已发货", "B002": "机械键盘, 打包中"}
    return orders.get(order_id, f"订单{order_id}未找到")

def calculate(expression):
    try:
        return f"{expression} = {eval(expression)}"
    except:
        return "计算失败"

# ================================================================
#  Agent 类
# ================================================================
class Agent:
    def __init__(self, name, system_prompt, tool_map, tools_config):
        self.name = name
        self.system_prompt = system_prompt
        self.tool_map = tool_map
        self.tools_config = tools_config
        self.messages = []
        self.client = OpenAI(
            api_key=os.environ.get("DEEPSEEK_API_KEY"),
            base_url="https://api.deepseek.com"
        )
        self.max_rounds = 5

    def reset(self):
        """重置对话历史"""
        self.messages = [{"role": "system", "content": self.system_prompt}]

    def _think_and_act(self):
        """ReAct 内层循环：思考 → 行动 → 观察 → 再思考"""
        round_num = 0
        while round_num < self.max_rounds:
            round_num += 1

            response = self.client.chat.completions.create(
                model="deepseek-chat",
                messages=self.messages,
                tools=self.tools_config,
                temperature=0
            )
            msg = response.choices[0].message

            if msg.tool_calls:
                self.messages.append(msg)
                for tool_call in msg.tool_calls:
                    func_name = tool_call.function.name
                    args = json.loads(tool_call.function.arguments)

                    result = self.tool_map[func_name](**args)

                    print(f"  🧠 Thought → ⚡ {func_name}({args})")
                    print(f"  👁️ Observation: {result}")

                    self.messages.append({
                        "role": "tool",
                        "tool_call_id": tool_call.id,
                        "content": result
                    })
            else:
                self.messages.append({"role": "assistant", "content": msg.content})
                return msg.content

        return "⚠️ 达到最大推理轮数，请简化问题后重试。"

    def run(self):
        """启动对话循环"""
        self.reset()
        print(f"=== {self.name}（输入 quit 退出）===")
        while True:
            user_input = input("\n👤 你: ")
            if user_input.lower() == "quit":
                print("👋 再见！")
                break
            self.messages.append({"role": "user", "content": user_input})
            answer = self._think_and_act()
            print(f"🤖 {self.name}: {answer}")

# ================================================================
#  使用示例
# ================================================================
if __name__ == "__main__":
    # 几行代码创建一个 Agent
    tool_map = {
        "get_weather": get_weather,
        "get_order": get_order,
        "calculate": calculate
    }

    tools_config = [
        {
            "type": "function",
            "function": {
                "name": "get_weather",
                "description": "查询城市天气",
                "parameters": {
                    "type": "object",
                    "properties": {"city": {"type": "string"}},
                    "required": ["city"]
                }
            }
        },
        {
            "type": "function",
            "function": {
                "name": "get_order",
                "description": "查询订单状态",
                "parameters": {
                    "type": "object",
                    "properties": {"order_id": {"type": "string"}},
                    "required": ["order_id"]
                }
            }
        },
        {
            "type": "function",
            "function": {
                "name": "calculate",
                "description": "数学计算",
                "parameters": {
                    "type": "object",
                    "properties": {"expression": {"type": "string"}},
                    "required": ["expression"]
                }
            }
        }
    ]

    assistant = Agent(
        name="小智",
        system_prompt="你是一个全能助手，语气友善简洁。可以查天气、查订单、算数学。",
        tool_map=tool_map,
        tools_config=tools_config
    )

    assistant.run()