Windows 通过代理服务器调用 openai 的 api

1. 查看网络是否联通

通过代理访问一个 HTTPS 网站来验证代理服务器是否工作正常

curl -x socks5h://127.0.0.1:1080 https://www.google.com

2. curl 访问openai的api

curl -x socks5h://127.0.0.1:1080 -s https://api.openai.com/v1/models/gpt-3.5-turbo  
-H "Authorization: Bearer $OPENAI_API_KEY"

结果：

{
  "id": "gpt-3.5-turbo",
  "object": "model",
  "created": 1677610602,
  "owned_by": "openai"
}

再次访问：

curl -k -x socks5h://127.0.0.1:1080 -X POST https://api.openai.com/v1/chat/completions ^
-H "Content-Type: application/json" ^
-H "Authorization: Bearer $OPENAI_API_KEY" ^
-d "{\"model\":\"gpt-3.5-turbo\",\"messages\":[{\"role\":\"user\",\"content\":\"Translate the following English text to French: 'Hello, world!'\"}],\"max_tokens\":100}"

其中：

• -x http://proxy.example.com:8080：指定代理服务器及其端口。
• -X POST：指定请求方法为 POST。
• https://api.openai.com/v1/chat/completions：目标 API 的 URL。
• -H “Content-Type: application/json”：指定请求头，表明发送的数据类型为 JSON。
• -H “Authorization: Bearer YOUR_API_KEY”：指定请求头，包含你的 API 密钥。
• -d “{“model”:“gpt-3.5-turbo”,“messages”:[{“role”:“user”,“content”:“Translate the following English text to French: ‘Hello, world!’”}],“max_tokens”:100}”：发送的数据，包含一个 JSON 对象。

特别注意：在 Windows 命令行中，双引号内的双引号需要用反斜杠转义。

3. python访问OpenAI的api

python 访问代码：

import openai
import os
import httpx
import time

# 设置你的 OpenAI API 密钥
openai.api_key = os.getenv("OPENAI_API_KEY")

# 设置代理服务器
proxy_url = 'socks5://127.0.0.1:1080'

# 创建一个 HTTPX 客户端并配置代理
client = httpx.Client(proxies={
    "http://": proxy_url,
    "https://": proxy_url,
})

# 自定义发送请求的函数
def list_models(max_retries=10, retry_delay=5):
    url = "https://api.openai.com/v1/models"
    headers = {
        "Authorization": f"Bearer {openai.api_key}"
    }

    for attempt in range(max_retries):
        try:
            response = client.get(url, headers=headers)
            response.raise_for_status()  # 如果请求失败，抛出异常
            return response.json()
        except httpx.HTTPStatusError as e:
            if e.response.status_code == 429:
                print(f"Received 429 Too Many Requests. Retrying in {retry_delay} seconds...")
                time.sleep(retry_delay)
                retry_delay *= 2  # 指数退避策略：等待时间加倍
            else:
                raise e
    raise Exception("Max retries exceeded")

def print_models(models_response):
    print("Available Models:")
    print("=================")
    for model in models_response['data']:
        model_id = model.get('id', 'N/A')
        created_at = model.get('created', 'N/A')
        owned_by = model.get('owned_by', 'N/A')
        print(f"Model ID: {model_id}")
        print(f"Created At: {created_at}")
        print(f"Owned By: {owned_by}")
        print("-----------------")

# 列出模型并打印结果
models_response = list_models()
print_models(models_response)

输出结果：

Available Models:
=================
Model ID: whisper-1
Created At: 1677532384
Owned By: openai-internal
-----------------
Model ID: babbage-002
Created At: 1692634615
Owned By: system
-----------------
Model ID: dall-e-2
Created At: 1698798177
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-16k
Created At: 1683758102
Owned By: openai-internal
-----------------
Model ID: tts-1-hd-1106
Created At: 1699053533
Owned By: system
-----------------
Model ID: tts-1-hd
Created At: 1699046015
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-instruct-0914
Created At: 1694122472
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-instruct
Created At: 1692901427
Owned By: system
-----------------
Model ID: dall-e-3
Created At: 1698785189
Owned By: system
-----------------
Model ID: text-embedding-3-small
Created At: 1705948997
Owned By: system
-----------------
Model ID: tts-1
Created At: 1681940951
Owned By: openai-internal
-----------------
Model ID: text-embedding-3-large
Created At: 1705953180
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-0125
Created At: 1706048358
Owned By: system
-----------------
Model ID: gpt-3.5-turbo
Created At: 1677610602
Owned By: openai
-----------------
Model ID: gpt-3.5-turbo-0301
Created At: 1677649963
Owned By: openai
-----------------
Model ID: tts-1-1106
Created At: 1699053241
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-16k-0613
Created At: 1685474247
Owned By: openai
-----------------
Model ID: gpt-3.5-turbo-0613
Created At: 1686587434
Owned By: openai
-----------------
Model ID: text-embedding-ada-002
Created At: 1671217299
Owned By: openai-internal
-----------------
Model ID: davinci-002
Created At: 1692634301
Owned By: system
-----------------
Model ID: gpt-3.5-turbo-1106
Created At: 1698959748
Owned By: system
-----------------

最后一个模型用 curl 验证：

$ curl -x socks5h://127.0.0.1:1080 -s https://api.openai.com/v1/models/gpt-3.5-turbo-1106  -H "Authorization: Bearer $OPENAI_API_KEY"
{
  "id": "gpt-3.5-turbo-1106",
  "object": "model",
  "created": 1698959748,
  "owned_by": "system"
}

检索特定的模型 gpt-3.5-turbo-1106：

import openai
import os
import httpx

# 设置你的 OpenAI API 密钥
openai.api_key = os.getenv("OPENAI_API_KEY")

# 设置代理服务器
proxy_url = 'socks5://127.0.0.1:1080'

# 创建一个 HTTPX 客户端并配置代理
client = httpx.Client(proxies={
    "http://": proxy_url,
    "https://": proxy_url,
})

# 自定义发送请求的函数以检索模型信息
def retrieve_model(model_id, max_retries=10, retry_delay=5):
    url = f"https://api.openai.com/v1/models/{model_id}"
    headers = {
        "Authorization": f"Bearer {openai.api_key}"
    }

    for attempt in range(max_retries):
        try:
            response = client.get(url, headers=headers)
            response.raise_for_status()  # 如果请求失败，抛出异常
            return response.json()
        except httpx.HTTPStatusError as e:
            if e.response.status_code == 429:
                print(f"Received 429 Too Many Requests. Retrying in {retry_delay} seconds...")
                time.sleep(retry_delay)
                retry_delay *= 2  # 指数退避策略：等待时间加倍
            else:
                raise e
    raise Exception("Max retries exceeded")

# 检索 gpt-3.5-turbo 模型的信息
model_id = "gpt-3.5-turbo-1106"
model_info = retrieve_model(model_id)

# 输出模型信息
print(f"Model ID: {model_info.get('id', 'N/A')}")
print(f"Object: {model_info.get('object', 'N/A')}")
print(f"Created At: {model_info.get('created', 'N/A')}")
print(f"Owned By: {model_info.get('owned_by', 'N/A')}")

运行结果：

Model ID: gpt-3.5-turbo-1106
Object: model
Created At: 1698959748
Owned By: system

还有代码：

import openai
import os
import httpx
import time

# 设置你的 OpenAI API 密钥
openai.api_key = os.getenv("OPENAI_API_KEY")

# 设置代理服务器
proxy_url = 'socks5://127.0.0.1:1080'

# 创建一个 HTTPX 客户端并配置代理
client = httpx.Client(proxies={
    "http://": proxy_url,
    "https://": proxy_url,
})

# 自定义发送请求的函数
def create_chat_completion(messages, max_retries=10, retry_delay=10):
    url = "https://api.openai.com/v1/chat/completions"
    headers = {
        "Content-Type": "application/json",
        "Authorization": f"Bearer {openai.api_key}"
    }
    payload = {
        "model": "gpt-3.5-turbo",
        "messages": messages
    }

    for attempt in range(max_retries):
        try:
            response = client.post(url, headers=headers, json=payload)
            response.raise_for_status()  # 如果请求失败，抛出异常
            return response.json()
        except httpx.HTTPStatusError as e:
            if e.response.status_code == 429:
                print(f"Received 429 Too Many Requests. Retrying in {retry_delay} seconds...")
                time.sleep(retry_delay)
            else:
                raise e
    raise Exception("Max retries exceeded")

# 测试请求
messages = [
    {"role": "user", "content": "Translate the following English text to French: 'Hello, world!'"}
]
response = create_chat_completion(messages)

# 输出响应
print(response['choices'][0]['message']['content'])

参考：

1. Open Ai 常见接口参数说明以及常见报错总结