Docker容器中Python代码无法连接OpenAI且运行OpenAI Agent代码报参数错误的解决求助

开发者阿泽

2026-4-14

我正在尝试创建一个OpenAI代理，基于用户输入对运行在Docker容器（docker-vscode-1）里的VS Code Server进行认证。我在容器内运行了以下代码：

import os
import requests
import logging
from requests.exceptions import RequestException
from dotenv import load_dotenv
from langchain_community.llms import OpenAI
from langchain.agents import initialize_agent, Tool
from langchain.agents.agent_types import AgentType
import warnings

warnings.simplefilter("ignore")
load_dotenv()

# Configuration
CODE_SERVER_URL = os.getenv("CODE_SERVER_URL", "http://172.20.0.2:8080")
PASSWORD = os.getenv("CODE_SERVER_PASSWORD", "yourpassword")
LOGIN_URL = f"{CODE_SERVER_URL}/login"
SESSION = requests.Session()
DEBUG = os.getenv("DEBUG", "False").lower() == "true"
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")

# Set up logging
logging.basicConfig(
    level=logging.DEBUG if DEBUG else logging.INFO,
    format="%(asctime)s [%(levelname)s]: %(message)s"
)


def authenticate():
    """Authenticates with the code-server."""
    if not PASSWORD:
        logging.error("Password not set. Please set the 'CODE_SERVER_PASSWORD' environment variable.")
        return "Password not set."

    try:
        # Get the login page to retrieve CSRF token if needed (optional for code-server)
        response = SESSION.get(LOGIN_URL)
        response.raise_for_status()

        # Post the login credentials
        payload = {"password": PASSWORD}
        auth_response = SESSION.post(LOGIN_URL, data=payload)
        auth_response.raise_for_status()

        # Check if login was successful
        if auth_response.status_code == 200 and "code-server" in auth_response.text:
            logging.info("Authentication successful!")
            return "Authentication successful!"
        else:
            logging.warning("Authentication failed. Check your password or URL.")
            return "Authentication failed. Check your password or URL."

    except Exception as e:
        error_message = f"An unexpected error occurred: {e}"
        logging.error(error_message)
        return error_message


# Define the LangChain agent tools
tools = [
    Tool(
        name="Authenticate with Code-Server",
        func=authenticate,
        description="Authenticate with the VS code-server. \
            This is a one-time action and does not require further input. \
            This action completes immediately."
    )
]

# Initialize the LangChain agent with OpenAI's LLM
def main():
    
    llm = OpenAI(openai_api_key=OPENAI_API_KEY, temperature=0)

    agent = initialize_agent(
        tools,
        llm,
        agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
        verbose=True
    )

    # Agent loop to take user input
    print("Simple AI Agent is ready! Type 'exit' to quit.")
    while True:
        user_input = input("\nEnter your instruction (type 'exit' to quit): ")
        if user_input.lower() in ["exit", "quit"]:
            print("Goodbye!")
            break

        try:
            # Get the agent's response for a single action
            response = agent.run(user_input)
            # Ensure that no further actions are taken after a single response
            if "Action:" in response:
                print("Ashwani Kumar Shamlodhiya: The AI has finished processing your request.")
            else:
                print(response)
    
        except Exception as e:
            logging.error(f"An error occurred: {e}")

if __name__ == "__main__":
    main()

运行后收到以下错误：

Traceback (most recent call last):
  File "/home/coder/project/temp/aiagent-vscode-using-openaiTODO.py", line 113, in <module>
    main()
  File "/home/coder/project/temp/aiagent-vscode-using-openaiTODO.py", line 83, in main
    llm = OpenAI(openai_api_key=OPENAI_API_KEY, temperature=0)
  File "/home/coder/.local/lib/python3.9/site-packages/langchain_core/_api/deprecation.py", line 216, in warn_if_direct_instance
    return wrapped(self, *args, **kwargs)
  File "/home/coder/.local/lib/python3.9/site-packages/langchain_core/load/serializable.py", line 125, in __init__
    super().__init__(*args, **kwargs)
  File "/usr/local/lib/python3.9/dist-packages/pydantic/main.py", line 214, in __init__
    validated_self = self.__pydantic_validator__.validate_python(data, self_instance=self)
  File "/usr/local/lib/python3.9/dist-packages/pydantic/_internal/_decorators_v1.py", line 148, in _wrapper1
    return validator(values)
  File "/home/coder/.local/lib/python3.9/site-packages/langchain_core/utils/pydantic.py", line 219, in wrapper
    return func(cls, values)
  File "/home/coder/.local/lib/python3.9/site-packages/langchain_community/llms/openai.py", line 322, in validate_environment
    values["client"] = openai.OpenAI(**client_params).completions
  File "/usr/local/lib/python3.9/dist-packages/openai/_client.py", line 123, in __init__
    super().__init__(
  File "/usr/local/lib/python3.9/dist-packages/openai/_base_client.py", line 844, in __init__
    self._client = http_client or SyncHttpxClientWrapper(
  File "/usr/local/lib/python3.9/dist-packages/openai/_base_client.py", line 742, in __init__
    super().__init__(**kwargs)
TypeError: __init__() got an unexpected keyword argument 'proxies'

我的操作系统是Windows，请问有什么办法可以解决这个问题？

解决方案分析

这个错误核心是OpenAI SDK与LangChain系列包的版本不兼容导致的。新版本的OpenAI Python SDK（v1.x及以上）已经移除了proxies这个初始化参数，但旧版本的LangChain仍在尝试传递该参数，从而触发了类型错误。下面给你几个实用的解决办法：

方法一：安装兼容的版本组合
如果你不想改动代码，可以直接安装经过验证的兼容版本：
1. 先卸载现有冲突包：
```
pip uninstall -y openai langchain langchain-community
```
2. 安装指定版本：
```
pip install openai==0.28.1 langchain==0.1.10 langchain-community==0.0.23
```
  这个版本组合能完美适配你当前的代码写法，不会出现参数冲突问题。
方法二：升级到最新兼容版本（推荐）
如果想使用新版OpenAI SDK（v1.x+），需要同步升级LangChain并调整代码：
1. 升级所有相关包：
```
pip install --upgrade langchain langchain-community openai
```
2. 修改代码中的LLM初始化部分：
  原来的导入和初始化要替换成：
```
# 替换旧的from langchain_community.llms import OpenAI
from langchain_openai import OpenAI
llm = OpenAI(api_key=OPENAI_API_KEY, temperature=0)
```
  新版LangChain已经把OpenAI相关的实现迁移到了langchain_openai包中，需要安装这个包的话可以执行pip install langchain-openai。

方法三：调整代理配置（如果容器需要代理访问OpenAI）
如果你在Docker容器内需要通过代理访问OpenAI，不要依赖LangChain的旧代理参数，而是直接通过环境变量或requests会话设置：

# 在代码开头添加环境变量代理（可选）
os.environ['HTTP_PROXY'] = 'http://你的代理地址:端口'
os.environ['HTTPS_PROXY'] = 'http://你的代理地址:端口'

# 或者给requests.Session设置代理
SESSION = requests.Session()
SESSION.proxies = {
    'http': 'http://你的代理地址:端口',
    'https': 'http://你的代理地址:端口',
}

另外，别忘了检查Docker容器的网络连通性，可以在容器内执行curl https://api.openai.com/v1/models测试是否能正常访问OpenAI API，避免因网络问题引发其他错误。

备注：内容来源于stack exchange，提问作者Ashwani Shamlodhiya

Docker容器中Python代码无法连接OpenAI且运行OpenAI Agent代码报参数错误的解决求助

解决方案分析 #

解决方案分析