from __future__ import with_statement

import re

counter = {}
total_lines = 0

# finds the requester's IP address from a log line such as:
# 2009-07-07 19:39:39 78.101.206.196 - W3SVC126 NT49 216.81.77.197 80 GET /index.asp - 200 0 21899 256 375 HTTP/1.1 Mozilla/4.0+(compatible;+MSIE+6.0;+Windows+NT+5.1;+SV1;+.NET+CLR+1.1.4322) - -
regex = re.compile("\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2} (?P<ip_address>[0-9\.]{7,15}) .*")

with open("~/Downloads/ex090707.log") as current_file:
	for line in current_file.readlines():
		ip_list = regex.findall(line)
		
		for ip_address in ip_list:
			if counter.has_key(ip_address):
				counter[ip_address] += 1
			else:
				counter[ip_address] = 1
			total_lines += 1
			
result = open("./sort_results.txt", 'w')

counter_lines = sorted(counter.iteritems(), key=lambda (k,v): (v,k))
counter_lines.reverse()

result.write(str(total_lines) + ' total\n\n\n')

for k, c in counter_lines:
	result.write(k + '\t:\t' + str(c) + '\n')

result.close()