爬虫性能分析
对于爬虫，python进行并发抓取的实现方式主要有以下几种：进程，线程，协程。
性能的消耗主要在IO请求中，当单进程单线程模式下请求URL时必然会引起等待，从而使得请求整体变慢。
一 多进程执行
可以实现并发，但是，请求发送出去后和返回之前，中间时期进程空闲
编写方式：
1- 多进程直接返回处理
1 from concurrent.futures import ProcessPoolExecutor 2 import requests 3 import time 4 5 def task(url): 6 response = requests.get(url) 7 print(url,response) 8 # 写正则表达式 9 return response 10 11 pool = ProcessPoolExecutor(7) 12 url_list = [ 13 'http://www.cnblogs.com/wupeiqi', 14 'http://huaban.com/favorite/beauty/', 15 'http://www.bing.com', 16 'http://www.zhihu.com', 17 'http://www.sina.com', 18 'http://www.baidu.com', 19 'http://www.autohome.com.cn', 20 ] 21 22 for url in url_list: 23 pool.submit(task,url) 24 25 pool.shutdown(wait=True)
View Code
2- 多进程通过回调函数处理
1 from concurrent.futures import ProcessPoolExecutor 2 import requests 3 import time 4 5 def task(url): 6 response = requests.get(url) 7 return response 8 9 def done(future,*args,**kwargs): 10 response = future.result() 11 print(response.status_code,response.content) 12 13 pool = ProcessPoolExecutor(7) 14 url_list = [ 15 'http://www.cnblogs.com/wupeiqi', 16 'http://huaban.com/favorite/beauty/', 17 'http://www.bing.com', 18 'http://www.zhihu.com', 19 'http://www.sina.com', 20 'http://www.baidu.com', 21 'http://www.autohome.com.cn', 22 ] 23 for url in url_list: 24 v = pool.submit(task,url) 25 v.add_done_callback(done) 26 27 pool.shutdown(wait=True)
View Code
二 多线程执行
爬虫可以实现并发，但是，请求发送出去后和返回之前，中间时期线程空闲。
编写方式：
1 多线程直接返回处理
from concurrent.futures import ThreadPoolExecutor import requests import time def task(url): response = requests.get(url) print(url,response) # 写正则表达式 pool = ThreadPoolExecutor(7) url_list = [ 'http://www.cnblogs.com/wupeiqi', 'http://huaban.com/favorite/beauty/', 'http://www.bing.com', 'http://www.zhihu.com', 'http://www.sina.com', 'http://www.baidu.com', 'http://www.autohome.com.cn', ] for url in url_list: pool.submit(task,url) pool.shutdown(wait=True)
2 多线程通过回调函数处理
from concurrent.futures import ThreadPoolExecutor import requests import time def task(url): """ 下载页面 :param url: :return: """ response = requests.get(url) return response def done(future,*args,**kwargs): response = future.result() print(response.status_code,response.content) pool = ThreadPoolExecutor(7) url_list = [ 'http://www.cnblogs.com/wupeiqi', 'http://huaban.com/favorite/beauty/', 'http://www.bing.com', 'http://www.zhihu.com', 'http://www.sina.com', 'http://www.baidu.com', 'http://www.autohome.com.cn', ] for url in url_list: v = pool.submit(task,url) v.add_done_callback(done) pool.shutdown(wait=True)
通过上述代码均可以完成对请求性能的提高，对于多线程和多进行的缺点是在IO阻塞时会造成了线程和进程的浪费
三 异步非阻塞模块
协程（微线程） + 异步IO—> 1个线程发送N个Http请求
import asyncio @asyncio.coroutine def task(): print('before...task......') yield from asyncio.sleep(5) # 发送Http请求，支持TCP获取结果.. print('end...task......') tasks = [task(), task()] loop = asyncio.get_event_loop() loop.run_until_complete(asyncio.gather(*tasks)) loop.close()
one
1 import asyncio 2 @asyncio.coroutine 3 def task(host, url='/'): 4 print('start',host,url) 5 reader, writer = yield from asyncio.open_connection(host, 80) 6 7 request_header_content = "GET %s HTTP/1.0\r\nHost: %s\r\n\r\n" % (url, host,) 8 request_header_content = bytes(request_header_content, encoding='utf-8') 9 10 writer.write(request_header_content) 11 yield from writer.drain() 12 text = yield from reader.read() 13 print('end',host, url, text) 14 writer.close() 15 16 tasks = [ 17 task('www.cnblogs.com', '/gregoryli/'), 18 task('dig.chouti.com', '/pic/show?nid=4073644713430508&lid=10273091') 19 ] 20 21 loop = asyncio.get_event_loop() 22 results = loop.run_until_complete(asyncio.gather(*tasks)) 23 loop.close()
two
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://www.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()
aiohttp
1 # -*- coding: utf-8 -*- 2 # 2017/11/17 14:04 3 import asyncio 4 import requests 5 6 @asyncio.coroutine 7 def task(func, *args): 8 print(func,args) 9 loop = asyncio.get_event_loop() 10 future = loop.run_in_executor(None, func, *args) # requests.get('http://www.cnblogs.com/wupeiqi/') 11 response = yield from future 12 print(response.url, response.content) 13 14 tasks = [ 15 task(requests.get, 'http://www.cnblogs.com/gregoryli/'), 16 task(requests.get, 'http://dig.chouti.com/pic/show?nid=4073644713430508&lid=10273091') 17 ] 18 19 loop = asyncio.get_event_loop() 20 results = loop.run_until_complete(asyncio.gather(*tasks)) 21 loop.close()
requests
import gevent import requests from gevent import monkey monkey.patch_all() def task(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(task, method='get', url='https://www.python.org/', req_kwargs={}), # gevent.spawn(task, method='get', url='https://www.yahoo.com/', req_kwargs={}), # gevent.spawn(task, 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(5) gevent.joinall([ pool.spawn(task, method='get', url='https://www.python.org/', req_kwargs={}), pool.spawn(task, method='get', url='https://www.yahoo.com/', req_kwargs={}), pool.spawn(task, method='get', url='https://www.github.com/', req_kwargs={}), ])
gevent+requests
1 import grequests 2 3 request_list = [ 4 grequests.get('http://httpbin.org/delay/1', timeout=0.001), 5 grequests.get('http://fakedomain/'), 6 grequests.get('http://httpbin.org/status/500') 7 ] 8 9 # ##### 执行并获取响应列表 ##### 10 response_list = grequests.map(request_list,size=5) 11 print(response_list)
grequests
#!/usr/bin/env python # -*- coding:utf-8 -*- from twisted.internet import defer from twisted.web.client import getPage from twisted.internet import reactor def one_done(arg): print(arg) def all_done(arg): print('done') reactor.stop() @defer.inlineCallbacks def task(url): res = getPage(bytes(url, encoding='utf8')) # 发送Http请求 res.addCallback(one_done) yield res url_list = [ 'http://www.cnblogs.com', 'http://www.cnblogs.com', 'http://www.cnblogs.com', 'http://www.cnblogs.com', ] defer_list = [] # [特殊，特殊，特殊(已经向url发送请求)] for url in url_list: v = task(url) defer_list.append(v) d = defer.DeferredList(defer_list) d.addBoth(all_done) reactor.run() # 死循环,事件循环
twisted
1 #!/usr/bin/env python 2 # -*- coding:utf-8 -*- 3 from tornado.httpclient import AsyncHTTPClient 4 from tornado.httpclient import HTTPRequest 5 from tornado import ioloop 6 7 COUNT = 0 8 def handle_response(response): 9 global COUNT 10 COUNT -= 1 11 if response.error: 12 print("Error:", response.error) 13 else: 14 print(response.body) 15 # 方法同twisted 16 # ioloop.IOLoop.current().stop() 17 18 if COUNT == 0: 19 ioloop.IOLoop.current().stop() 20 21 def func(): 22 url_list = [ 23 'http://www.baidu.com', 24 'http://www.bing.com', 25 ] 26 global COUNT 27 COUNT = len(url_list) 28 for url in url_list: 29 print(url) 30 http_client = AsyncHTTPClient() 31 http_client.fetch(HTTPRequest(url), handle_response) 32 33 ioloop.IOLoop.current().add_callback(func) 34 ioloop.IOLoop.current().start() # 死循环
tornado
– asyncio
– 示例1：asyncio.sleep(5)
– 示例2：自己封装Http数据包
– 示例3：asyncio+aiohttp
aiohttp模块：封装Http数据包 pip3 install aiohttp
– 示例4：asyncio+requests
requests模块：封装Http数据包 pip3 install requests
– gevent，greenlet+异步IO
pip3 install greenlet
pip3 install gevent
– 示例1：gevent+requests
– 示例2：gevent（协程池，最多发多少个请求）+requests
– 示例3：gevent+requests => grequests
pip3 install grequests
效率：gevent > Twisted > Tornado > asyncio
四 socket
1. socket客户端、服务端
连接阻塞
setblocking(0): 无数据（连接无响应；数据未返回）就报错
——》http请求本质：阻塞
sk = socket.socket() # 1.连接 sk.connect(('www.baidu.com',80,)) # IO阻塞 print('连接成功了...') # 2. 连接成功发送消息 sk.send(b'GET / HTTP/1.0\r\nHost:www.baidu.com\r\n\r\n') # sk.send(b'POST / HTTP/1.0\r\nHost:www.baidu.com\r\n\r\nk1=v1&k2=v2') # 3. 等待着服务端响应 data = sk.recv(8096) # 响应头，响应体，IO阻塞 print(data) # 关闭连接 sk.close()
2. IO多路复用
客户端：
try:
socket对象1.connet()
socket对象2.connet()
socket对象3.connet()
except Exception as e:
pass
while True:
r,w,e = select.select([socket对象1,socket对象2,socket对象3,],[socket对象1,socket对象2,socket对象3,],[],0.05)
r = [socket对象1,] # 表示有人给我发送数据
xx = socket对象1.recv()
w = [socket对象1,] # 表示我已经和别人创建连接成功:
socket对象1.send(‘”””GET /index HTTP/1.0\r\nHost: baidu.com\r\n\r\n”””‘)
———>http请求本质：非阻塞
sk = socket.socket() sk.setblocking(False) # 1.连接 try: sk.connect(('www.baidu.com',80,)) # IO阻塞 print('连接成功了...') except BlockingIOError as e: print(e) # 2. 连接成功发送消息 sk.send(b'GET / HTTP/1.0\r\nHost:www.baidu.com\r\n\r\n') # sk.send(b'POST / HTTP/1.0\r\nHost:www.baidu.com\r\n\r\nk1=v1&k2=v2') # 3. 等待着服务端响应 data = sk.recv(8096) # IO阻塞 print(data) # 关闭连接 sk.close()
3.
class Foo:
·def fileno(self):
·obj = socket()
·return obj.fileno()
r,w,e = select.select([socket对象？,对象？,对象？,Foo()],[],[])
# 对象必须有： fileno方法，并返回一个文件描述符
要点：
a. select内部：对象.fileno()
b. Foo()内部封装socket文件描述符
IO多路复用： r,w,e = while 监听多个socket对象；
异步IO： 非阻塞的socket+IO多路复用
– 非阻塞socket
– select[自己对象],w,r
import socket import select class HttpRequest: def __init__(self,sk,host,callback): self.socket = sk self.host = host self.callback = callback def fileno(self): return self.socket.fileno() class HttpResponse: def __init__(self,recv_data): self.recv_data = recv_data self.header_dict = {} self.body = None self.initialize() def initialize(self): headers, body = self.recv_data.split(b'\r\n\r\n', 1) self.body = body header_list = headers.split(b'\r\n') for h in header_list: h_str = str(h,encoding='utf-8') v = h_str.split(':',1) if len(v) == 2: self.header_dict[v[0]] = v[1] class AsyncRequest: def __init__(self): self.conn = [] self.connection = [] # 用于检测是否已经连接成功 def add_request(self,host,callback): try: sk = socket.socket() sk.setblocking(0)#false sk.connect((host,80,)) except BlockingIOError as e: pass request = HttpRequest(sk,host,callback) self.conn.append(request) self.connection.append(request) def run(self): while True: rlist,wlist,elist = select.select(self.conn,self.connection,self.conn,0.05) for w in wlist: print(w.host,'连接成功...') # 只要能循环到，表示socket和服务器端已经连接成功 tpl = "GET / HTTP/1.0\r\nHost:%s\r\n\r\n" %(w.host,) w.socket.send(bytes(tpl,encoding='utf-8')) self.connection.remove(w) for r in rlist: # r,是HttpRequest recv_data = bytes() while True: try: chunck = r.socket.recv(8096) recv_data += chunck except Exception as e: break print(r.host,"有数据返回...",recv_data) response = HttpResponse(recv_data) r.callback(response) r.socket.close() self.conn.remove(r) if len(self.conn) == 0: break def f1(response): print('保存到文件',response.header_dict) def f2(response): print('保存到数据库', response.header_dict) url_list = [ {'host':'www.baidu.com','callback': f1}, {'host':'cn.bing.com','callback': f2}, {'host':'www.cnblogs.com','callback': f2}, ] req = AsyncRequest() for item in url_list: req.add_request(item['host'],item['callback']) req.run()
View Code
版权声明：本文为gregoryli原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。