python爬虫（5）——正则表达式（二）

1 import re
2 
3 str=‘192.168.1.1‘
4 
5 re.search(r‘(([01]{0,1}\d{0,1}\d|2[0-4]\d|25[0-5])\.){3}([01]{0,1}\d{0,1}\d|2[0-4]\d|25[0-5])‘,str)

 1 import urllib.request
 2 import re
 3 
 4 
 5 url="http://www.xicidaili.com/"
 6 useful_ip=[]
 7 def loadPage(url):
 8     headers = {"User-Agent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.84 Safari/537.36"}
 9     response=urllib.request.Request(url,headers=headers)
10     html=urllib.request.urlopen(response).read().decode("utf-8")
11     return html
12 
13 def getProxy():
14     html=loadPage(url)
15     pattern=re.compile(r‘(<td>\d+</td>)‘)
16     duankou=pattern.findall(html)
17     pattern=re.compile(r‘(?:(?:[01]{0,1}\d{0,1}\d|2[0-4]\d|25[0-5])\.){3}(?:[01]{0,1}\d{0,1}\d|2[0-4]\d|25[0-5])‘)
18     content_list=pattern.findall(html)
19     list_num=[]
20     for num in duankou:
21         list_num.append(num[4:-5])
22     for i in range(len(list_num)):
23         ip=content_list[i]+ ":"+list_num[i]
24         while True:
25             proxy_support=urllib.request.ProxyHandler({‘http‘:ip})
26             opener=urllib.request.build_opener(proxy_support)
27             opener.add_handler=[("User-Agent","Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.84 Safari/537.36")]
28             urllib.request.install_opener(opener)
29             try:
30                 print("正在尝试使用 %s 访问..." % ip)
31                 ip_filter="http://www.whatsmyip.org/"
32                 ip_response=urllib.request.urlopen(ip_filter)
33             except urllib.error.URLError:
34                 print("访问出错，这个IP不能用啦")
35                 break
36             else:
37                 print("访问成功！")
38                 print("可用IP为： %s " % ip)
39                 useful_ip.append(ip)
40                 if input("继续爬取？")=="N":
41                     print("有效IP如下：")
42                     for key in useful_ip:
43                         print(key)
44                         exit()
45                 else:
46                     break
47 
48 if __name__=="__main__":
49     getProxy()