nginx日志简单分析工具

自己有个tony6.com的服务器，上面挂着我的博客，web服务器是nginx。

由于最近一直在折腾python，所以简单写了个nginx日志分析工具，它可以分析出每个IP的点击数量和IP所在地。

 1 #!/usr/bin/env pypy
 2 #coding:utf8
 3 import re
 4 import sys
 5 import urllib
 6 import json
 7 contents = sys.argv[1]
 8 url = "http://ip.taobao.com/service/getIpInfo.php?ip="
 9 def ip_location(ip):
10     data = urllib.urlopen(url + ip).read()
11     datadict=json.loads(data)
12     for oneinfo in datadict:
13         if "code" == oneinfo:
14             if datadict[oneinfo] == 0:
15                 return datadict["data"]["country"] + datadict["data"]["region"] + datadict["data"]["city"] + "\t" + datadict["data"]["isp"]
16 
17 def NginxIpHite(logfile_path):
18     ipadd = r'\.'.join([r'\d{1,3}']*4)
19     re_ip = re.compile(ipadd)
20     iphitlisting = {}
21     for line in open(contents):
22         match = re_ip.match(line)
23         if match:
24             ip = match.group()
25             iphitlisting[ip] = iphitlisting.get(ip, 0) + 1
26     return iphitlisting
27 
28 for k in NginxIpHite(contents):
29     dili = ip_location(k)
30     print "%s " % k,NginxIpHite(contents)[k],dili