宜配屋

解析url用的类库:

python2版本：

from urlparse import urlparse
import urllib

python3版本：

from urllib.parse import urlparse
import urllib.request

研究了不同的url规则发现：只要在搜索关键字是用=嫁接的，查询的关键在解析后的query里

如果不是用=嫁接，查询的关键在解析后的path里。

解析的规则都是一样的，正则如下：（6中不同情况的组合）

另外host为‘s.weibo.com'的url编码与其他不同要另做处理。

代码如下：有些网站的规则还不是很清楚，需要花大量时间找规则，规则越清晰，关键字就越清楚，如下规则已适合绝大部分网站，酌情参考。

# -*- coding:utf-8 -*-
 
from urlparse import urlparse
import urllib
import re
 
# url
source_txt = "E:\\python_Anaconda_code\\url.txt"
# 规则
regular = r'(\w+(%\w\w)+\w+|(%\w\w)+\w+(%\w\w)+|\w+(%\w\w)+|(%\w\w)+\w+|(%\w\w)+|\w+)'
 
# 存放关键字
kw_list = list()
 
# key为要研究网站的host，value为关键字的嫁接标识符
dict = {
 "www.baidu.com": "wd=",
 "news.baidu.com": "word=",
 "www.sogou.com": "query=",
 "tieba.baidu.com": "kw=",
 "wenku.baidu.com": "word=",
 "music.sina.com.cn": "k=",
 "www.haosou.com": "q=",
 "www.lagou.com": "list_",
 "www.chunyuyisheng.com": "query=",
 "s.weibo.com": "weibo/"
}
 
def Main():
 with open(source_txt, 'r') as f_source_txt:
  for url in f_source_txt:
   host = url.split("//")[1].split("/")[0]
   if host in dict:
    flag = dict[host]
    if flag.find("=") != -1:
     query = urlparse(url).query.replace('+', '')
     kw = re.search(flag + regular, query, re.I) # .group(0)
     if kw:
      kw = urllib.unquote(kw.group(0).split(flag)[1])
      print(kw)
    else:
     path = urlparse(url).path.replace('+', '')
     kw = re.search(flag + regular, path.replace("%25", "%"), re.I)
     if kw:
      kw = urllib.unquote(kw.group(0).split(flag)[1])
      print(kw)
if __name__ == '__main__':
 Main()

url.txt的内容如下：

https://www.baidu.com/s?ie=utf-8&f=8&rsv_bp=0&rsv_idx=1&ch=&tn=baidu&bar=&wd=python&rn=&oq=&rsv_pq=ece0867c0002c793&rsv_t=edeaQq7DDvZnxq%2FZVra5K%2BEUanlTIUXhGIhvuTaqdfOECLuXR25XKDp%2Bi0I&rqlang=cn&rsv_enter=1&inputT=218
https://www.baidu.com/s?ie=utf-8&f=8&rsv_bp=1&rsv_idx=1&ch=&tn=baidu&bar=&wd=python%E9%87%8C%E7%9A%84%E5%AD%97%E5%85%B8dict&oq=python&rsv_pq=96c160e70003f332&rsv_t=0880NkOvMIr3TvOdDP1t8EbloD8qwr4yeP6CfPjQihQNNhdExfuwyOFMrx0&rqlang=cn&rsv_enter=0&inputT=10411 
https://www.baidu.com/s?ie=utf-8&f=8&rsv_bp=1&rsv_idx=1&ch=&tn=baidu&bar=&wd=python%E9%87%8C%E7%9A%84urlprese&oq=python%25E9%2587%258C%25E7%259A%2584re%25E9%2587%258C%25E7%259A%2584%257C%25E6%2580%258E%25E4%25B9%2588%25E7%2594%25A8&rsv_pq=d1d4e7b90003d391&rsv_t=5ff4Vok4EELK1PgJ4oSk8L0VvKAn51%2BL8ns%2FjSubexg7Lb7znKcTvnVtn8M&rqlang=cn&rsv_enter=1&inputT=2797 
https://www.baidu.com/s?ie=utf-8&f=8&rsv_bp=1&rsv_idx=1&ch=&tn=baidu&bar=&wd=python++wo+%E7%88%B1urlprese&oq=python%25E9%2587%258C%25E7%259A%2584urlprese&rsv_pq=eecf45e900033e87&rsv_t=1c70xAYhrvw5JOZA7lpVgt4pw%2BW1TO8hqTejTh67JgEQfqAGyDydd25HAmU&rqlang=cn&rsv_enter=0&inputT=10884 
http://news.baidu.com/ns?word=%E8%B6%B3%E7%90%83&tn=news&from=news&cl=2&rn=20&ct=1 
http://news.baidu.com/ns?ct=1&rn=20&ie=utf-8&bs=%E8%B6%B3%E7%90%83&rsv_bp=1&sr=0&cl=2&f=8&prevct=no&tn=news&word=++++++%E8%B6%B3++%E7%90%83+++++%E4%BD%A0%E5%A5%BD+%E5%98%9B%EF%BC%9F&rsv_sug3=14&rsv_sug4=912&rsv_sug1=4&inputT=8526 
http://tieba.baidu.com/f?ie=utf-8&kw=%E7%BA%A2%E6%B5%B7%E8%A1%8C%E5%8A%A8&fr=search&red_tag=q0224393377 
https://www.sogou.com/web?query=ni+zai+%E6%88%91+%E5%BF%83li&_asf=www.sogou.com&_ast=1520388441&w=01019900&p=40040100&ie=utf8&from=index-nologin&s_from=index&sut=9493&sst0=1520388440692&lkt=8%2C1520388431200%2C1520388436842&sugsuv=1498714959961744&sugtime=1520388440692 
https://www.lagou.com/jobs/list_python%E5%A4%A7%E6%95%B0%E6%8D%AEmr?labelWords=&fromSearch=true&suginput= 
https://www.chunyuyisheng.com/pc/search/?query=%E6%85%A2%E6%80%A7%E4%B9%99%E8%82%9D% 
http://s.weibo.com/weibo/%25E5%2594%2590%25E4%25BA%25BA%25E8%25A1%2597%25E6%258E%25A2%25E6%25A1%25882&Refer=index
http://s.weibo.com/weibo/%25E4%25BD%25A0%25E5%25A5%25BD123mm%2520%25E5%2597%25AF%2520mm11&Refer=STopic_box

结果如下：

Python解析、提取url关键字