apache日志分析(python)

#!/usr/bin/env python

import sys
import datetime
import socket
from file_backwards import *

MONTH = {
	‘Jan‘:1,
	‘Feb‘:2,
	‘Mar‘:3,
	‘Apr‘:4,
	‘May‘:5,
	‘Jue‘:6,
	‘Jul‘:7,
	‘Aug‘:8,
	‘Sep‘:9,
	‘Oct‘:10,
	‘Nov‘:11,
	‘Dec‘:12,
}

def parse_apache_date(datestr):
	day, month, yearandtime = datestr.split(‘/‘)
	year, hour, minute,second= yearandtime.split(‘:‘)
	return datetime.datetime(int(year),MONTH[month],int(day),int(hour),int(minute))

def countDict(d, k):
	if k in d:
		d[k] += 1
	else:
		d[k] = 1

def parse_apache_log(logfile,ten_m):
	result = {}
	with open(logfile) as fd:
		for line in filerev(fd):
			splited_line = line.split()
			datestr = splited_line[3][1:]
			apache_date = parse_apache_date(datestr)
			if apache_date > ten_m:
				countDict(result, apache_date.strftime(‘%s‘))
			else:
				return result

if __name__ == ‘__main__‘:
	now = datetime.datetime.now()
	timedelta = datetime.timedelta(minutes=10)
	ten_m_ago = now - timedelta
	key = ‘http.count‘
	data = parse_apache_log(sys.argv[1], ten_m_ago)
	sock = socket.socket()
	sock.connect((‘127.0.0.1‘, 2003))
	print data
	for k, v in data.items():
		sock.send("%s %d %s\n" % (key, v, k))