feat: added helpers for various exceptions including a prompt truncation error

[adubovik]

Introduced error helpers for the prompt truncation procedure which is triggered by request.max_prompt_tokens parameter.

A corresponding code is introduced for such an error called truncate_prompt_error:

def _truncate_prompt_error(message: str, **kwargs) -> HTTPException:
    return invalid_request_error(
        code="truncate_prompt_error",
        param="max_prompt_tokens",
        message=message,
        **kwargs,
    )

Compare it with the error message thrown by OpenAI itself when context length is exceeded:

    {
        "message": "This model's maximum context length is 128000 tokens. However, your messages resulted in 458759 tokens. Please reduce the length of the messages.",
        "type": "invalid_request_error",
        "param": "messages",
        "code": "context_length_exceeded"
    }