LLMClient

LLM adapter providing unified async call interface.

Class

LLMClient

class LLMClient:
    def __init__(
        self,
        api_key: str,
        base_url: str = "https://api.deepseek.com",
        model: str = "deepseek-chat",
        timeout: int = 30,
        max_retries: int = 3
    )

Parameters:

• api_key (str): API key
• base_url (str): API base URL
• model (str): Model name
• timeout (int): Timeout in seconds
• max_retries (int): Maximum retry count

Methods

chat_completion

async def chat_completion(
    self,
    system_prompt: str,
    user_prompt: str,
    temperature: float = 0.7,
    max_tokens: int = 2000
) -> str

Send chat completion request.

Parameters:

• system_prompt (str): System prompt
• user_prompt (str): User prompt
• temperature (float): Temperature parameter, controls randomness
• max_tokens (int): Maximum generation tokens

Returns:

• str: LLM response content

Exceptions:

• LLMTimeoutError: Request timeout
• LLMParseError: Response parsing failed

close

async def close(self)

Close client connection.

Features

• Auto Retry: Automatically retry on failure, up to max_retries times
• Timeout Control: Support custom timeout
• Async Support: Fully async implementation, supports high concurrency