aiohttp 异步http请求-6.ClientTimeout 整个会话超时

硬核知识 / 作者：

前言

ClientTimeout 是设置整个会话的超时时间，默认情况下是300秒（5分钟）超时。

ClientTimeout

默认情况下，aiohttp使用总共300 秒（5 分钟）超时，这意味着整个操作应该在 5 分钟内完成。

timeout = aiohttp.ClientTimeout(total=60)
async with aiohttp.ClientSession(timeout=timeout) as session:
    ...

对于如下对session会话发单个请求，超时可能会被覆盖ClientSession.get()：

async with session.get(url, timeout=timeout) as resp:
    ...

ClientTimeout 可以支持的字段

  • total 整个操作的最大秒数，包括建立连接、发送请求和读取响应。
  • connect 如果超出池连接限制，则建立新连接或等待池中的空闲连接的最大秒数。
  • sock_connect 为新连接连接到对等点的最大秒数，不是从池中给出的。
  • sock_read 从对等点读取新数据部分之间允许的最大秒数。

默认超时为：

aiohttp.ClientTimeout(total=5*60, connect=None,
                      sock_connect=None, sock_read=None)

实例

timeout参数是对整个aiohttp.ClientSession 会话的超时时间，比如我创建一个会话，里面有10个请求。
设置超时时间为3秒，那么从创建会话开始，在3秒内完成所有的请求就不会报错，当大于3秒还没完成10个请求就会抛异常

import aiohttp
import asyncio
from pathlib import Path


async def down_img(session, url):
    """下载图片"""
    name = url.split('/')[-1]  # 获得图片名字
    img = await session.get(url)
    # 触发到await就切换，等待get到数据
    content = await img.read()
    # 读取内容
    with open('./down_img/'+str(name), 'wb') as f:
        # 写入至文件
        f.write(content)
        print(f'{name} 下载完成！')

    return str(url)


async def main(URL):
    # 建立会话session
    timeout = aiohttp.ClientTimeout(total=3)
    conn = aiohttp.TCPConnector(ssl=False)  # 防止ssl报错
    async with aiohttp.ClientSession(timeout=timeout, connector=conn) as session:
        # 建立所有任务
        tasks = [asyncio.create_task(down_img(session, img_url)) for img_url in URL]
        # 触发await，等待任务完成
        done, pending = await asyncio.wait(tasks)
        # all_results = [done_task.result() for done_task in done]
        # # 获取所有结果
        # print("ALL RESULT:"+str(all_results))


URL = [
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
    'https://cdn.pixabay.com/photo/2018/03/12/20/07/maldives-3220702_960_720.jpg',
    'https://cdn.pixabay.com/photo/2017/08/04/17/56/dolomites-2580866_960_720.jpg',
    'https://cdn.pixabay.com/photo/2016/06/20/03/15/pier-1467984_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/07/30/02/00/iceberg-404966_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/11/02/10/41/plane-513641_960_720.jpg',
    'https://cdn.pixabay.com/photo/2015/10/30/20/13/sea-1014710_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
]

fp = Path('./down_img')
if not fp.exists():
    fp.mkdir()


loop = asyncio.get_event_loop()
loop.run_until_complete(main(URL))

由于3秒内没完成所有的请求，运行结果会抛出异常：asyncio.TimeoutError

sea-1014710_960_720.jpg 下载完成！
railroad-163518_960_720.jpg 下载完成！
plane-513641_960_720.jpg 下载完成！
iceberg-404966_960_720.jpg 下载完成！
Task exception was never retrieved
future: <Task finished name='Task-2' coro=<down_img() done, defined at D:/demo/demo/new/xuexi/d1.py:6> exception=TimeoutError()>
Traceback (most recent call last):
File "D:/demo/demo/new/xuexi/d1.py", line 12, in down_img
    content = await img.read()
...
    raise asyncio.TimeoutError from None
asyncio.exceptions.TimeoutError

给单个请求添加timeout

如果我们的需求是每个请求设置超时，单个请求大于3秒就超时，把timeout参数放到get请求上

    timeout = aiohttp.ClientTimeout(total=3)
    img = await session.get(url, timeout=timeout)

完整的代码

import aiohttp
import asyncio
from pathlib import Path


async def down_img(session, url):
    """下载图片"""
    name = url.split('/')[-1]  # 获得图片名字
    timeout = aiohttp.ClientTimeout(total=3)
    img = await session.get(url, timeout=timeout)
    # 触发到await就切换，等待get到数据
    content = await img.read()
    # 读取内容
    with open('./down_img/'+str(name), 'wb') as f:
        # 写入至文件
        f.write(content)
        print(f'{name} 下载完成！')

    return str(url)


async def main(URL):
    # 建立会话session
    conn = aiohttp.TCPConnector(ssl=False)  # 防止ssl报错
    async with aiohttp.ClientSession(connector=conn) as session:
        # 建立所有任务
        tasks = [asyncio.create_task(down_img(session, img_url)) for img_url in URL]
        # 触发await，等待任务完成
        done, pending = await asyncio.wait(tasks)
        # all_results = [done_task.result() for done_task in done]
        # # 获取所有结果
        # print("ALL RESULT:"+str(all_results))


URL = [
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
    'https://cdn.pixabay.com/photo/2018/03/12/20/07/maldives-3220702_960_720.jpg',
    'https://cdn.pixabay.com/photo/2017/08/04/17/56/dolomites-2580866_960_720.jpg',
    'https://cdn.pixabay.com/photo/2016/06/20/03/15/pier-1467984_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/07/30/02/00/iceberg-404966_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/11/02/10/41/plane-513641_960_720.jpg',
    'https://cdn.pixabay.com/photo/2015/10/30/20/13/sea-1014710_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
]

fp = Path('./down_img')
if not fp.exists():
    fp.mkdir()


loop = asyncio.get_event_loop()
loop.run_until_complete(main(URL))

这样写其实并不会达到我们的预期，官方文档上说对于如下请求，超时可能会被覆盖ClientSession.get()

那你timeout不管放在ClientSession() 还是单个get/post请求里，其实效果都一样，都是针对整个会话超时。

捕获超时异常 asyncio.TimeoutError

触发的超时异常是 asyncio.TimeoutError，从Traceback可以看到是content = await img.read() 这里报错

import aiohttp
import asyncio
from pathlib import Path


async def down_img(session, url):
    """下载图片"""
    name = url.split('/')[-1]  # 获得图片名字
    try:
        img = await session.get(url)
        # 触发到await就切换，等待get到数据
        content = await img.read()
        # 读取内容
        with open('./down_img/'+str(name), 'wb') as f:
            # 写入至文件
            f.write(content)
            print(f'{name} 下载完成！')
            return str(url)
    except asyncio.TimeoutError as msg:
        print('timeout: '+str(url))
        return str(url)



async def main(URL):
    # 建立会话session
    timeout = aiohttp.ClientTimeout(total=3)
    conn = aiohttp.TCPConnector(ssl=False)  # 防止ssl报错
    async with aiohttp.ClientSession(timeout=timeout, connector=conn) as session:
        # 建立所有任务
        tasks = [asyncio.create_task(down_img(session, img_url)) for img_url in URL]
        # 触发await，等待任务完成
        done, pending = await asyncio.wait(tasks)
        # all_results = [done_task.result() for done_task in done]
        # # 获取所有结果
        # print("ALL RESULT:"+str(all_results))


URL = [
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
    'https://cdn.pixabay.com/photo/2018/03/12/20/07/maldives-3220702_960_720.jpg',
    'https://cdn.pixabay.com/photo/2017/08/04/17/56/dolomites-2580866_960_720.jpg',
    'https://cdn.pixabay.com/photo/2016/06/20/03/15/pier-1467984_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/07/30/02/00/iceberg-404966_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/11/02/10/41/plane-513641_960_720.jpg',
    'https://cdn.pixabay.com/photo/2015/10/30/20/13/sea-1014710_960_720.jpg',
    'https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg',
    'https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg',
]

fp = Path('./down_img')
if not fp.exists():
    fp.mkdir()


loop = asyncio.get_event_loop()
loop.run_until_complete(main(URL))

运行结果:

plane-513641_960_720.jpg 下载完成！
maldives-3220702_960_720.jpg 下载完成！
sea-1014710_960_720.jpg 下载完成！
dolomites-2580866_960_720.jpg 下载完成！
iceberg-404966_960_720.jpg 下载完成！
mountain-477832_960_720.jpg 下载完成！
timeout: https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg
timeout: https://cdn.pixabay.com/photo/2016/06/20/03/15/pier-1467984_960_720.jpg
timeout: https://cdn.pixabay.com/photo/2013/07/18/10/56/railroad-163518_960_720.jpg
timeout: https://cdn.pixabay.com/photo/2014/10/07/13/48/mountain-477832_960_720.jpg