Function run_request

fastdeploy/entrypoints/openai/run_batch.py:294–329

(
    serving_engine_func: Callable,
    request: BatchRequestInput,
    tracker: BatchProgressTracker,
    semaphore: asyncio.Semaphore,
)

async def run_request(
    serving_engine_func: Callable,
    request: BatchRequestInput,
    tracker: BatchProgressTracker,
    semaphore: asyncio.Semaphore,
) -> BatchRequestOutput:
    async with semaphore:
        try:
            response = await serving_engine_func(request.body)

            if isinstance(response, ChatCompletionResponse):
                batch_output = BatchRequestOutput(
                    id=f"fastdeploy-{random_uuid()}",
                    custom_id=request.custom_id,
                    response=BatchResponseData(
                        status_code=200, body=response, request_id=f"fastdeploy-batch-{random_uuid()}"
                    ),
                    error=None,
                )
            elif isinstance(response, ErrorResponse):
                batch_output = BatchRequestOutput(
                    id=f"fastdeploy-{random_uuid()}",
                    custom_id=request.custom_id,
                    response=BatchResponseData(status_code=400, request_id=f"fastdeploy-batch-{random_uuid()}"),
                    error=response,
                )
            else:
                batch_output = make_error_request_output(request, error_msg="Request must not be sent in stream mode")

            tracker.completed()
            return batch_output

        except Exception as e:
            console_logger.error(f"Request {request.custom_id} processing failed: {str(e)}")
            tracker.completed()
            return make_error_request_output(request, error_msg=f"Request processing failed: {str(e)}")


def determine_process_id() -> int: