服务器之家

服务器之家 > 正文

Python 爬虫性能相关总结

时间:2020-08-04 00:07     来源/作者:syncd

这里我们通过请求网页例子来一步步理解爬虫性能

当我们有一个列表存放了一些url需要我们获取相关数据,我们首先想到的是循环

简单的循环串行

这一种方法相对来说是最慢的,因为一个一个循环,耗时是最长的,是所有的时间总和
代码如下:

?
1
2
3
4
5
6
7
8
9
10
11
import requests
 
url_list = [
  'http://www.baidu.com',
  'http://www.pythonsite.com',
  'http://www.cnblogs.com/'
]
 
for url in url_list:
  result = requests.get(url)
  print(result.text)

通过线程池

通过线程池的方式访问,这样整体的耗时是所有连接里耗时最久的那个,相对循环来说快了很多

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
import requests
from concurrent.futures import ThreadPoolExecutor
 
def fetch_request(url):
  result = requests.get(url)
  print(result.text)
 
url_list = [
  'http://www.baidu.com',
  'http://www.bing.com',
  'http://www.cnblogs.com/'
]
pool = ThreadPoolExecutor(10)
 
for url in url_list:
  #去线程池中获取一个线程,线程去执行fetch_request方法
  pool.submit(fetch_request,url)
 
pool.shutdown(True)

线程池+回调函数

这里定义了一个回调函数callback

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
from concurrent.futures import ThreadPoolExecutor
import requests
 
 
def fetch_async(url):
  response = requests.get(url)
 
  return response
 
 
def callback(future):
  print(future.result().text)
 
 
url_list = [
  'http://www.baidu.com',
  'http://www.bing.com',
  'http://www.cnblogs.com/'
]
 
pool = ThreadPoolExecutor(5)
 
for url in url_list:
  v = pool.submit(fetch_async,url)
  #这里调用回调函数
  v.add_done_callback(callback)
 
pool.shutdown()

通过进程池

通过进程池的方式访问,同样的也是取决于耗时最长的,但是相对于线程来说,进程需要耗费更多的资源,同时这里是访问url时IO操作,所以这里线程池比进程池更好

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
import requests
from concurrent.futures import ProcessPoolExecutor
 
def fetch_request(url):
  result = requests.get(url)
  print(result.text)
 
url_list = [
  'http://www.baidu.com',
  'http://www.bing.com',
  'http://www.cnblogs.com/'
]
pool = ProcessPoolExecutor(10)
 
for url in url_list:
  #去进程池中获取一个线程,子进程程去执行fetch_request方法
  pool.submit(fetch_request,url)
 
pool.shutdown(True)

进程池+回调函数

这种方式和线程+回调函数的效果是一样的,相对来说开进程比开线程浪费资源

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
from concurrent.futures import ProcessPoolExecutor
import requests
 
 
def fetch_async(url):
  response = requests.get(url)
 
  return response
 
 
def callback(future):
  print(future.result().text)
 
 
url_list = [
  'http://www.baidu.com',
  'http://www.bing.com',
  'http://www.cnblogs.com/'
]
 
pool = ProcessPoolExecutor(5)
 
for url in url_list:
  v = pool.submit(fetch_async, url)
  # 这里调用回调函数
  v.add_done_callback(callback)
 
pool.shutdown()

主流的单线程实现并发的几种方式

  1. asyncio
  2. gevent
  3. Twisted
  4. Tornado

下面分别是这四种代码的实现例子:

asyncio例子1:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
import asyncio
 
 
@asyncio.coroutine #通过这个装饰器装饰
def func1():
  print('before...func1......')
  # 这里必须用yield from,并且这里必须是asyncio.sleep不能是time.sleep
  yield from asyncio.sleep(2)
  print('end...func1......')
 
 
tasks = [func1(), func1()]
 
loop = asyncio.get_event_loop()
loop.run_until_complete(asyncio.gather(*tasks))
loop.close()

上述的效果是同时会打印两个before的内容,然后等待2秒打印end内容
这里asyncio并没有提供我们发送http请求的方法,但是我们可以在yield from这里构造http请求的方法。

asyncio例子2:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
import asyncio
 
 
@asyncio.coroutine
def fetch_async(host, url='/'):
  print("----",host, url)
  reader, writer = yield from asyncio.open_connection(host, 80)
 
  #构造请求头内容
  request_header_content = """GET %s HTTP/1.0\r\nHost: %s\r\n\r\n""" % (url, host,)
  request_header_content = bytes(request_header_content, encoding='utf-8')
  #发送请求
  writer.write(request_header_content)
  yield from writer.drain()
  text = yield from reader.read()
  print(host, url, text)
  writer.close()
 
tasks = [
  fetch_async('www.cnblogs.com', '/zhaof/'),
  fetch_async('dig.chouti.com', '/pic/show?nid=4073644713430508&lid=10273091')
]
 
loop = asyncio.get_event_loop()
results = loop.run_until_complete(asyncio.gather(*tasks))
loop.close()

asyncio + aiohttp 代码例子:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
import aiohttp
import asyncio
 
 
@asyncio.coroutine
def fetch_async(url):
  print(url)
  response = yield from aiohttp.request('GET', url)
  print(url, response)
  response.close()
 
 
tasks = [fetch_async('http://baidu.com/'), fetch_async('http://www.chouti.com/')]
 
event_loop = asyncio.get_event_loop()
results = event_loop.run_until_complete(asyncio.gather(*tasks))
event_loop.close()

asyncio+requests代码例子

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
import asyncio
import requests
 
 
@asyncio.coroutine
def fetch_async(func, *args):
  loop = asyncio.get_event_loop()
  future = loop.run_in_executor(None, func, *args)
  response = yield from future
  print(response.url, response.content)
 
 
tasks = [
  fetch_async(requests.get, 'http://www.cnblogs.com/wupeiqi/'),
  fetch_async(requests.get, 'http://dig.chouti.com/pic/show?nid=4073644713430508&lid=10273091')
]
 
loop = asyncio.get_event_loop()
results = loop.run_until_complete(asyncio.gather(*tasks))
loop.close()

gevent+requests代码例子

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
import gevent
 
import requests
from gevent import monkey
 
monkey.patch_all()
 
 
def fetch_async(method, url, req_kwargs):
  print(method, url, req_kwargs)
  response = requests.request(method=method, url=url, **req_kwargs)
  print(response.url, response.content)
 
# ##### 发送请求 #####
gevent.joinall([
  gevent.spawn(fetch_async, method='get', url='https://www.python.org/', req_kwargs={}),
  gevent.spawn(fetch_async, method='get', url='https://www.yahoo.com/', req_kwargs={}),
  gevent.spawn(fetch_async, method='get', url='https://github.com/', req_kwargs={}),
])
 
# ##### 发送请求(协程池控制最大协程数量) #####
# from gevent.pool import Pool
# pool = Pool(None)
# gevent.joinall([
#   pool.spawn(fetch_async, method='get', url='https://www.python.org/', req_kwargs={}),
#   pool.spawn(fetch_async, method='get', url='https://www.yahoo.com/', req_kwargs={}),
#   pool.spawn(fetch_async, method='get', url='https://www.github.com/', req_kwargs={}),
# ])

grequests代码例子
这个是讲requests+gevent进行了封装

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
import grequests
 
 
request_list = [
  grequests.get('http://httpbin.org/delay/1', timeout=0.001),
  grequests.get('http://fakedomain/'),
  grequests.get('http://httpbin.org/status/500')
]
 
 
# ##### 执行并获取响应列表 #####
# response_list = grequests.map(request_list)
# print(response_list)
 
 
# ##### 执行并获取响应列表(处理异常) #####
# def exception_handler(request, exception):
# print(request,exception)
#   print("Request failed")
 
# response_list = grequests.map(request_list, exception_handler=exception_handler)
# print(response_list)

twisted代码例子

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
#getPage相当于requets模块,defer特殊的返回值,rector是做事件循环
from twisted.web.client import getPage, defer
from twisted.internet import reactor
 
def all_done(arg):
  reactor.stop()
 
def callback(contents):
  print(contents)
 
deferred_list = []
 
url_list = ['http://www.bing.com', 'http://www.baidu.com', ]
for url in url_list:
  deferred = getPage(bytes(url, encoding='utf8'))
  deferred.addCallback(callback)
  deferred_list.append(deferred)
#这里就是进就行一种检测,判断所有的请求知否执行完毕
dlist = defer.DeferredList(deferred_list)
dlist.addBoth(all_done)
 
reactor.run()

tornado代码例子

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
from tornado.httpclient import AsyncHTTPClient
from tornado.httpclient import HTTPRequest
from tornado import ioloop
 
 
def handle_response(response):
  """
  处理返回值内容(需要维护计数器,来停止IO循环),调用 ioloop.IOLoop.current().stop()
  :param response:
  :return:
  """
  if response.error:
    print("Error:", response.error)
  else:
    print(response.body)
 
 
def func():
  url_list = [
    'http://www.baidu.com',
    'http://www.bing.com',
  ]
  for url in url_list:
    print(url)
    http_client = AsyncHTTPClient()
    http_client.fetch(HTTPRequest(url), handle_response)
 
 
ioloop.IOLoop.current().add_callback(func)
ioloop.IOLoop.current().start()

以上就是Python 爬虫性能相关总结的详细内容,更多关于Python 爬虫性能的资料请关注服务器之家其它相关文章!

原文链接:https://www.cnblogs.com/zhaof/p/7171148.html

标签:

相关文章

热门资讯

2022年最旺的微信头像大全 微信头像2022年最新版图片
2022年最旺的微信头像大全 微信头像2022年最新版图片 2022-01-10
蜘蛛侠3英雄无归3正片免费播放 蜘蛛侠3在线观看免费高清完整
蜘蛛侠3英雄无归3正片免费播放 蜘蛛侠3在线观看免费高清完整 2021-08-24
背刺什么意思 网络词语背刺是什么梗
背刺什么意思 网络词语背刺是什么梗 2020-05-22
yue是什么意思 网络流行语yue了是什么梗
yue是什么意思 网络流行语yue了是什么梗 2020-10-11
暖暖日本高清免费中文 暖暖在线观看免费完整版韩国
暖暖日本高清免费中文 暖暖在线观看免费完整版韩国 2021-05-08
返回顶部