Parsing log file

parse_logfile

#!/usr/bin/env python
#
# log_format traffic '$http_host|$status|$bytes_sent|$msec|$remote_addr|$request_uri';

import sys
import itertools, operator
from datetime import datetime

access_log = '/var/log/nginx/access.log'

def sort_uniq(sequence):
  return itertools.imap( operator.itemgetter(0), itertools.groupby(sorted(sequence)))

def get_unique_hosts(lines):
  hosts = []
  for line in lines:
    hosts.append(line.split("|")[0])
  return list(sort_uniq(hosts))

def parse_log_file(logfile):
  print 'Parsing %s' % logfile
  f = open(logfile)
  lines = f.readlines()
  hosts = get_unique_hosts(lines)
  print hosts
  for line in lines:
    fullline = line.split('|')
    host = fullline[0]
    status_code = fullline[1]
    bytes_sent = fullline[2]
    date_time = float(fullline[3])
    request_uri = fullline[4]
    print datetime.fromtimestamp(date_time)
    print "request uri %s with bytes sent %s" % ( request_uri, bytes_sent )
try:
   with open(access_log):
     parse_log_file(access_log)
except IOError:
   print 'Unable to open %s' % access_log
   exit(1)