task1

#!/usr/bin/python

# http://regexr.com/

import re
import os
import time
import datetime

text = r'''123.124.554.123 this is 192.168.5.20 a
195 ip adress 200.5.3.125 and not an ip 344.45.12.145,
244.244.244.0 244.244.50.1
time: 12:15:11 5:11:30 10:30:22
date: 31/12/2015 8/13/2010 9/7/2010 1/1/2050
bytes: 4444

to test sum:

192.168.5.20
200.5.3.125
5556
'''
# date = %d/%m/%Y
# time = %H:%M:%S

#with open(file_name) as f:
#    text = f.read()

ips = re.compile(r'\b(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\b', re.MULTILINE)
times = re.compile(r'\b(?:[0-5]?[0-9]\:){2}[0-5]?[0-9]\b', re.MULTILINE)
dates = re.compile(r'\b[0-3]?[0-9]\/(?:1[0-2]|[0-9])\/\d{4}\b', re.MULTILINE)


bits = [int(i) for i in text.split() if i.isdigit()]

def make_ip_tuples(lst):
    """
    get list of all ips
    return tuples from each line

    """
    try:
        assert len(lst) % 2 == 0
    except AssertionError:
        print 'lst length not even, try to add some IPs to make couples'
        exit(1)

    result = []
    for i in range(len(lst))[::2]:
        ip_tuple = (lst[i], lst[i+1])
        result.append(ip_tuple)
    return result

def dump(ip_tuples, bts):
    """
    get list of ip tuples and list of bits for every occurence
    return D: .keys = set uf tuples
               .values = sum of bits
    """
    result = {}
    for i in range(len(ip_tuples)):
        ip_tuple = ip_tuples[i]
        if ip_tuple not in result:
            result[ip_tuple] = result.get(ip_tuple, bts[i])
        else:
            result[ip_tuple] += bts[i]
    return result

def time_stamp(d, t):
    ts = '%s %s' % (d, t)
    result = time.mktime(datetime.datetime.strptime(ts, '%d/%m/%Y %H:%M:%S').timetuple())
    return result

def make_output(input_list):
    """
    return dict with all columns in sorted order without dublicates
    """
    sorted_list = sorted(input_list, reverse=True, key=lambda x: time_stamp(x[0], x[1]))

    result = []
    for i in sorted_list:
        if i[2] not in [row[2] for row in result]:
            result.append(list(i))
        else:
            idx = [row[2] for row in result].index(i[2])
            result[idx][3] += i[3]
    return result


ips = ips.findall(text)
time_list = times.findall(text)
date_list = dates.findall(text)
ip_tuples = make_ip_tuples(ips)
ip_dump = dump(ip_tuples, bits)

inp = zip(date_list, time_list, ip_tuples, bits)

print make_output(inp)