# SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import json import logging from components.processor import Processor from fastapi import FastAPI from fastapi.responses import JSONResponse, StreamingResponse from utils.args import parse_vllm_args from utils.protocol import MultiModalRequest from dynamo.sdk import DYNAMO_IMAGE, api, depends, service logger = logging.getLogger(__name__) @service( dynamo={ "namespace": "dynamo", }, resources={"cpu": "10", "memory": "20Gi"}, workers=1, image=DYNAMO_IMAGE, app=FastAPI(title="Multimodal Example"), ) class Frontend: processor = depends(Processor) def __init__(self): class_name = self.__class__.__name__ self.engine_args = parse_vllm_args(class_name, "") @api(name="v1/chat/completions") async def generate(self, request: MultiModalRequest): if self.engine_args.model != request.model: return JSONResponse( {"error": f"Model '{request.model}' not found"}, status_code=404, ) async def content_generator(): async for response in self.processor.generate(request.model_dump_json()): try: s = json.loads(response) yield s except json.JSONDecodeError as e: raise RuntimeError(f"Failed to parse JSON response: {e}") return StreamingResponse(content_generator(), media_type="text/event-stream")