python 正则分析nginx日志
#!/usr/bin/env python#-*- coding:utf-8 –*-
#Author:xiaoluo
#QQ:942729042
#date:2015:05:12
import re
import sys
log = sys.argv
ip = r"?P<ip>[\d.]*"
date = r"?P<date>\d+"
month = r"?P<month>\w+"
year = r"?P<year>\d+"
log_time = r"?P<time>\S+"
timezone = r"""?P<timezone>
[^\"]*
"""
name = r"""?P<name>\"
[^\"]*\"
"""
method = r"?P<method>\S+"
request = r"?P<request>\S+"
protocol = r"?P<protocol>\S+"
status = r"?P<status>\d+"
bodyBytesSent = r"?P<bodyBytesSent>\d+"
refer = r"""?P<refer>\"
[^\"]*\"
"""
userAgent=r"""?P<userAgent>
.*
"""
#f = open('access1.log','r')
#for logline in f.readlines():
p = re.compile(r"(%s)\ \[(%s)/(%s)/(%s)\:(%s)\ (%s)\ (%s)\ (%s)\ (%s)\ (%s)\ (%s)\ (%s)\ (%s)\ (%s)" %(ip, date, month, year, log_time,timezone,name,method,request,protocol,status,bodyBytesSent,refer,userAgent), re.VERBOSE)
def getcode():
codedic={}
f = open(log,'r')
for logline in f.readlines():
matchs = p.match(logline)
if matchs !=None:
allGroups =matchs.groups()
status= allGroups
codedic=codedic.get(status,0) +1
return codedic
f.close()
def getIP():
f = open(log,'r')
IPdic={}
for logline in f.readlines():
matchs = p.match(logline)
if matchs !=None:
allGroups =matchs.groups()
IP=allGroups
IPdic = IPdic.get(IP,0) +1
IPdic=sorted(IPdic.iteritems(),key=lambda c:c,reverse=True)
IPdic=IPdic
return IPdic
f.close()
def getURL():
f = open(log,'r')
URLdic={}
for logline in f.readlines():
matchs = p.match(logline)
if matchs !=None:
allGroups =matchs.groups()
urlname = allGroups
URLdic = URLdic.get(urlname,0) +1
URLdic=sorted(URLdic.iteritems(),key=lambda c:c,reverse=True)
URLdic=URLdic
return URLdic
def getpv():
f = open(log,'r')
pvdic={}
for logline in f.readlines():
matchs = p.match(logline)
if matchs !=None:
allGroups =matchs.groups()
timezone=allGroups
time = timezone.split(':')
minute = time+":"+time
pvdic=pvdic.get(minute,0) +1
pvdic=sorted(pvdic.iteritems(),key=lambda c:c,reverse=True)
pvdic=pvdic
return pvdic
if __name__=='__main__':
print "网站监控状况检查状态码"
print getcode()
print "网站访问量最高的20个IP地址"
print getIP()
print "网站访问最多的20个站点名"
print getURL()
print getpv()
页:
[1]