91 lines
2.5 KiB
Python
91 lines
2.5 KiB
Python
"""
|
||
This code is supported by the website: https://www.guanjihuan.com
|
||
The newest version of this code is on the web page: https://www.guanjihuan.com/archives/47925
|
||
"""
|
||
|
||
import os
|
||
import dotenv
|
||
from langchain_openai import ChatOpenAI
|
||
from langchain_core.tools import tool
|
||
from langchain.agents import create_agent
|
||
from langchain_core.globals import set_debug
|
||
|
||
# 开启调试日志输出
|
||
set_debug(True)
|
||
dotenv.load_dotenv()
|
||
|
||
# -----------------------
|
||
# 定义工具
|
||
# -----------------------
|
||
@tool
|
||
def execute_python(code: str) -> str:
|
||
"""执行 Python 代码并返回输出"""
|
||
import sys, io
|
||
old_stdout = sys.stdout
|
||
sys.stdout = mystdout = io.StringIO()
|
||
try:
|
||
exec(code, {})
|
||
except Exception as e:
|
||
return f"Error: {e}"
|
||
finally:
|
||
sys.stdout = old_stdout
|
||
return mystdout.getvalue().strip()
|
||
|
||
tools = [execute_python]
|
||
|
||
# -----------------------
|
||
# 创建 LLM
|
||
# -----------------------
|
||
llm = ChatOpenAI(
|
||
api_key=os.getenv("OPENAI_API_KEY"),
|
||
base_url=os.getenv("DASHSCOPE_BASE_URL"),
|
||
model="qwen-plus",
|
||
temperature=0,
|
||
)
|
||
|
||
# -----------------------
|
||
# 创建 Agent
|
||
# -----------------------
|
||
agent = create_agent(llm, tools)
|
||
|
||
# -----------------------
|
||
# 系统消息,模拟 ReAct 风格
|
||
# -----------------------
|
||
system_message = """
|
||
你是一个严谨的 Python 编程助手,必须通过工具执行代码来验证逻辑。
|
||
每次只能执行一个 Action,所有 Python 代码必须通过工具执行,不得假设输出结果。
|
||
Action Input 必须是纯 Python 代码,不要包含 Markdown 代码块符号。
|
||
如果测试失败(如 assert 报错),请根据错误信息修正代码并重新测试。
|
||
输出必须严格遵循以下格式:
|
||
Thought: <你的思考>
|
||
Action: <工具名称>
|
||
Action Input: <纯 Python 代码>
|
||
Observation: <工具返回结果>
|
||
Final Answer: <最终答案>
|
||
"""
|
||
|
||
# -----------------------
|
||
# 用户问题:编写 Fibonacci 函数
|
||
# -----------------------
|
||
user_message = """
|
||
请编写一个 Python 函数 fibonacci(n),返回第 n 个斐波那契数。
|
||
先测试代码正确性,如果测试通过再进行优化,并给出最终代码。
|
||
"""
|
||
|
||
# -----------------------
|
||
# 调用 Agent
|
||
# -----------------------
|
||
response = agent.invoke(
|
||
{
|
||
"messages": [
|
||
{"role": "system", "content": system_message},
|
||
{"role": "user", "content": user_message}
|
||
]
|
||
}
|
||
)
|
||
|
||
# -----------------------
|
||
# 输出结果
|
||
# -----------------------
|
||
print("\n--- Agent 输出 ---\n")
|
||
print(response["messages"][-1].content) |