写在前面:代码写于2013年
老大布置的任务,要分析一个5G大小的nginx log file,因为我的python也是刚学,所以摸索了很久,才实现了这个需求,废话不多话,简单粗暴,直接上代码!
功能介绍:
(1)统计Top 100 访问次数最多的ip,并显示地理位置信息!这个是用的淘宝的地址库返回的ip地理位置及运营商信息
注:这地方说明一下,log里记录的文件有的是分段发送给客户端,所以同一个ip可能只是访问一次,但在log里显示了多条记录,在这里我就简单粗暴的把每一次都算作一个访问记录!有待改进,其他同学也可以修改下,告诉我应该怎么识别多少条记录是一次完整的访问!
(2)统计Top 100 流量最高ip,并显示地理位置信息!
(3)统计Top 100 访问流量最高url列表!
(4)log文件记录的总流量!
下面上代码,有需要的同学直接拿去!这个脚本分析一个4G的log用时13分左右,系统配置(16G内存)!
(1)ip_location.py文件:利用淘宝ip地址库,返回ip所在国家,区域(省份),城市,运营商
ip_location.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# the script is used to query the location of every ip
import urllib
import json
#淘宝ip库接口
url = "http://ip.taobao.com/service/getIpInfo.php?ip="
def ip_location(ip):
data = urllib.urlopen(url + ip).read()
datadict=json.loads(data)
for oneinfo in datadict:
if "code" == oneinfo:
if datadict[oneinfo] == 0:
return datadict["data"]["country"] + datadict["data"]["region"] + datadict["data"]["city"] + "\t\t" + datadict["data"]["isp"]
(2)logparser.py文件:完成统计功能,具体见代码内注释!
#!/usr/local/python
# -*- coding: utf-8 -*-
import os
import time
import re
import sys
import ip_location
"""定义一个时间类,可以选取要分析的时间段,如果没有指定时间段,则分析全部log"""
class TimeParser(object):
def __init__(self, re_time, str_time, period):
self.__re_time = re.compile(re_time)
self.__str_time = str_time
self.__period = period
def __get(self, line):
t= re.search(self.__re_time, line).group(0)
return time.mktime(time.strptime(t, self.__str_time))
def inPeriod(self, line):
t = self.__get(line)
return (t > time.mktime(time.strptime(self.__period[0], self.__str_time))
and t < time.mktime(time.strptime(self.__period[1], self.__str_time)))
class ParseLog(object):
def __init__(self, file, re_time, str_time, period):
self.ip_dict = {}
self.url_dict = {}
try:
self.domain, self.parsetime, self.suffix = file.split("_")
except:
self.domain = file.split(".")[0]
self.parsetime = "unknown time"
#定义一个函数,用来统计数量和总流量,并存入到相应字典中
def Count(self):
#用TimeParser实例化CountTime
CountTime = TimeParser(re_time, str_time, period)
self.total_traffic = []
"""
以下for循环分析每一行,如果这一行不包含时间,就跳过,如果包含时间信息,且在所分析时间段内,
则统计ip和traffic,没有http_refer信息的行只记录ip,然后跳过!
"""
with open(file) as f:
for i, line in enumerate(f):
try:
if CountTime.inPeriod(line):
ip = line.split()[0]
try:
traffic = re.findall(r'\d{3}\ [^0]\d+', line)[0].split()[1]
except IndexError:
traffic = 0
try:
url = re.findall(r'GET\ .*\.*\ ', line)[0].split()[1]
except IndexError:
url = "unknown"
else:
continue
except AttributeError:
continue
self.ip_dict.setdefault(ip, {'number':0, 'traffic':0})['number'] += 1
self.ip_dict.setdefault(ip, {'number':0, 'traffic':0})['traffic'] += int(traffic)
self.url_dict.setdefault(url, 0)
self.url_dict[url] += int(traffic)
if not i % 1000000:
print "have processed " + str(i) + " lines !"
#统计总流量
self.total_traffic.append(int(traffic))
total = sum(self.total_traffic)
#打印总流量大小
print "******************************************************************"
print self.domain + " all the traffic in " + self.parsetime + " is below:"
print "total_traffic: %s" % str(total/1024/1024)+"MB"
"""定义两个字典,分别存储ip的数量和流量信息"""
def TopIp(self, number):
self.Count()
TopNumberIp = {}
TopTrafficIp = {}
#对字典赋值
for ip in self.ip_dict.keys():
TopNumberIp[ip] = self.ip_dict[ip]['number']
TopTrafficIp[ip] = self.ip_dict[ip]['traffic']
#按值从大到小的顺序排序键
SortIpNo = sorted(TopNumberIp.items(), key=lambda e: e[1], reverse=True)
SortIpTraffic = sorted(TopTrafficIp.items(), key=lambda e: e[1], reverse=True)
#输出连接数top 100 ip的相关信息到文件TopIpNo.txt中
ipno = open('TopIpNo.txt', 'w+')
ipno.write(u"ip地址\t\t\t访问次数\t\t国家/区域/城市\t\t\t运营商\n")
ipno.write("-------------------------------------------------------------------------------------------------\n")
for i in range(number):
try:
ipno.write(SortIpNo[i][0]+"\t\t"+str(SortIpNo[i][1])+"\t\t\t"+ip_location.ip_location(SortIpNo[i][0])+"\n")
except:
continue
ipno.write("-------------------------------------------------------------------------------------------------\n")
ipno.close()
#输出流量top 100 ip的相关信息到文件iptraffic.txt中
iptr = open('iptraffic.txt', 'w+')
iptr.write(u"ip地址\t\t\t总流量(MB)\t\t国家/区域/城市\t\t\t运营商\n")
iptr.write("-------------------------------------------------------------------------------------------------\n")
for i in range(number):
try:
iptr.write(SortIpTraffic[i][0]+"\t\t"+str(SortIpTraffic[i][1]/1024/1024))
#记入地理信息
iptr.write("\t\t\t"+ip_location.ip_location(SortIpTraffic[i][0])+"\n")
except:
continue
iptr.write("-------------------------------------------------------------------------------------------------\n")
iptr.close()
def TopUrl(self, number):
SortUrlTraffic = sorted(self.url_dict.items(), key=lambda e: e[1], reverse=True)
#输出流量top 100 url相关信息到urltraffic.txt文件中
urtr = open('urltraffic.txt', 'w+')
urtr.write("Filename".ljust(75)+u"TotalTraffic(MB)"+"\n")
urtr.write("-----------------------------------------------------------------------------------------\n")
for i in range(number):
try:
urtr.write(SortUrlTraffic[i][0].ljust(80)+str(SortUrlTraffic[i][1]/1024/1024)+"\n")
except:
continue
urtr.write("-----------------------------------------------------------------------------------------\n")
urtr.close()
#时间的正则和格式,一般不需要更改
re_time='\d{2}\/\w{3}\/\d{4}:\d{2}:\d{2}:\d{2}'
str_time='%d/%b/%Y:%H:%M:%S'
#定义分析的时间段
period=("16/Nov/2000:16:00:00", "16/Nov/2015:17:00:00")
#定义输出top number
number = 100
if __name__ == '__main__':
if len(sys.argv) < 2:
print 'no logfile specified!'
print "Usage: python logParser.py filename"
time.sleep(2)
sys.exit()
else:
file = sys.argv[1]
lp = ParseLog(file, re_time, str_time, period)
print
print "Start to parse the " + file + " struggling! please wait patiently!"
print
print "******************************************************************"
time.sleep(2)
lp.TopIp(number)
lp.TopUrl(number)