Python22_httpx网络请求

Python22_httpx网络请求

httpx.AsyncClient，这是 Python 中处理异步 HTTP 请求的核心类。

httpx.AsyncClient 概览

httpx.AsyncClient是HTTPX库的异步客户端，基于asyncio构建，支持非阻塞的 HTTP/1.1 和 HTTP/2 请求，性能远超同步的requests库。

基本使用

1. 安装

pipinstallhttpx# 如需 HTTP/2 支持pipinstallhttpx[http2]

2. 简单请求

importhttpximportasyncioasyncdefmain():# 创建客户端（推荐用 async with 管理生命周期）asyncwithhttpx.AsyncClient()asclient:# GET 请求response=awaitclient.get("https://api.github.com")print(response.status_code)# 200print(response.json())# 解析 JSONasyncio.run(main())

核心特性详解

3. 请求方法

asyncwithhttpx.AsyncClient()asclient:# 各类 HTTP 方法r1=awaitclient.get(url,params={"key":"value"})r2=awaitclient.post(url,json={"data":"value"})# JSON 体r3=awaitclient.post(url,data={"form":"field"})# 表单数据r4=awaitclient.put(url,content=b"raw bytes")# 原始字节r5=awaitclient.patch(url,files={"file":open("a.jpg","rb")})r6=awaitclient.delete(url)r7=awaitclient.head(url)r8=awaitclient.options(url)

4. 高级配置

client=httpx.AsyncClient(# 超时设置（连接、读取、写入）timeout=httpx.Timeout(10.0,connect=5.0),# 请求头headers={"User-Agent":"MyApp/1.0"},# 基础 URL（后续请求可写相对路径）base_url="https://api.example.com/v1",# 启用 HTTP/2http2=True,# 跟随重定向follow_redirects=True,# 验证 SSL（开发时可关闭）verify=True,# 或 verify=False（不推荐生产环境）# 代理proxies="http://localhost:8080",# 或按协议区分：proxies={"http://": "...", "https://": "..."}# 认证auth=("username","password"),# Basic Auth# 或 auth=httpx.BearerToken("token")# Cookie 持久化cookies={"session":"abc123"},# 限制连接池limits=httpx.Limits(max_connections=100,max_keepalive_connections=20))# 使用 base_url 后response=awaitclient.get("/users")# 实际请求 https://api.example.com/v1/users

5. 并发请求（核心优势）

importasyncioimporthttpxasyncdeffetch(client,url):response=awaitclient.get(url)returnresponse.json()asyncdefmain():urls=["https://api.github.com/users/octocat","https://api.github.com/users/torvalds","https://api.github.com/users/gvanrossum"]asyncwithhttpx.AsyncClient()asclient:# 并发执行所有请求tasks=[fetch(client,url)forurlinurls]results=awaitasyncio.gather(*tasks)print(results)asyncio.run(main())

6. 流式响应（大文件下载）

asyncwithhttpx.AsyncClient()asclient:# 流式读取，避免内存溢出asyncwithclient.stream("GET","https://example.com/large-file.zip")asresponse:asyncforchunkinresponse.aiter_bytes():# 处理每个数据块print(f"Received{len(chunk)}bytes")# 或流式文本asyncwithclient.stream("GET",url)asresponse:asyncforlineinresponse.aiter_lines():print(line)

7. 请求/响应钩子

deflog_request(request):print(f"→ Request:{request.method}{request.url}")deflog_response(response):print(f"← Response:{response.status_code}")asyncwithhttpx.AsyncClient(event_hooks={"request":[log_request],"response":[log_response]})asclient:awaitclient.get("https://httpbin.org/get")

与同步 Client 对比

最佳实践

importhttpximportasyncioclassAPIClient:"""封装示例：带重试和错误处理的异步客户端"""def__init__(self):self.client=httpx.AsyncClient(base_url="https://api.example.com",timeout=30.0,http2=True,limits=httpx.Limits(max_connections=50))asyncdef__aenter__(self):returnselfasyncdef__aexit__(self,*args):awaitself.client.aclose()# 确保关闭asyncdefget_user(self,user_id:int):try:response=awaitself.client.get(f"/users/{user_id}")response.raise_for_status()# 自动抛出 4xx/5xx 异常returnresponse.json()excepthttpx.HTTPStatusErrorase:print(f"HTTP error:{e.response.status_code}")raiseexcepthttpx.RequestErrorase:print(f"Request failed:{e}")raise# 使用asyncdefmain():asyncwithAPIClient()asapi:user=awaitapi.get_user(123)print(user)asyncio.run(main())

常见异常处理

httpx.AsyncClient是现代 Python 异步编程中处理 HTTP 请求的首选工具，特别适合FastAPI/Starlette等异步 Web 框架的后端服务调用。