API调用基础

{
  "headers": {
    "Authorization": "Bearer sk-...",  // API密钥
    "X-API-Version": "2025-01",        // 版本控制
    "X-Request-ID": "uuid-v4"          // 请求追踪
  }
}

{
  "model": "gpt-4-turbo-2025",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant"},
    {"role": "user", "content": "Hello, how are you?"}
  ],
  "temperature": 0.7,
  "max_tokens": 1000,
  "response_format": {
    "type": "json_schema",
    "json_schema": {...}  // 2025新增：结构化输出
  }
}

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1709000000,
  "choices": [{
    "message": {"role": "assistant", "content": "..."},
    "finish_reason": "stop"
  }],
  "usage": {"prompt_tokens": 10, "completion_tokens": 20}
}

data: {"choices":[{"delta":{"content":"Hello"}}]}
data: {"choices":[{"delta":{"content":" there"}}]}
data: [DONE]

# 指数退避重试
def retry_with_backoff(func, max_retries=3):
    for i in range(max_retries):
        try:
            return func()
        except RateLimitError as e:
            if i == max_retries - 1:
                raise
            wait_time = (2 ** i) + random.uniform(0, 1)
            time.sleep(wait_time)