Update deprecated Python 3.8 typing (#13971)

vllm/entrypoints/openai/api_server.py

@@ -13,10 +13,11 @@ import socket
 import tempfile
 import uuid
 from argparse import Namespace
+from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
 from functools import partial
 from http import HTTPStatus
-from typing import Annotated, AsyncIterator, Dict, Optional, Set, Tuple, Union
+from typing import Annotated, Optional, Union
 
 import uvloop
 from fastapi import APIRouter, Depends, FastAPI, Form, HTTPException, Request
@@ -93,7 +94,7 @@ prometheus_multiproc_dir: tempfile.TemporaryDirectory
 # Cannot use __name__ (https://github.com/vllm-project/vllm/pull/4765)
 logger = init_logger('vllm.entrypoints.openai.api_server')
 
-_running_tasks: Set[asyncio.Task] = set()
+_running_tasks: set[asyncio.Task] = set()
 
 
 @asynccontextmanager
@@ -587,7 +588,7 @@ async def do_rerank_v2(request: RerankRequest, raw_request: Request):
     return await do_rerank(request, raw_request)
 
 
-TASK_HANDLERS: Dict[str, Dict[str, tuple]] = {
+TASK_HANDLERS: dict[str, dict[str, tuple]] = {
     "generate": {
         "messages": (ChatCompletionRequest, create_chat_completion),
         "default": (CompletionRequest, create_completion),
@@ -894,7 +895,7 @@ async def init_app_state(
     state.task = model_config.task
 
 
-def create_server_socket(addr: Tuple[str, int]) -> socket.socket:
+def create_server_socket(addr: tuple[str, int]) -> socket.socket:
     family = socket.AF_INET
     if is_valid_ipv6_address(addr[0]):
         family = socket.AF_INET6