#!/usr/bin/python
# emacs: -*- mode: python; py-indent-offset: 4; indent-tabs-mode: nil -*-
# vi: set ft=python sts=4 ts=4 sw=4 et:
#
import fileinput
import sys
import time
from datetime import datetime
import re
import json
import operator


releases = {
    'etch': 'Debian GNU/Linux 4.0 (etch)',
    'lenny': 'Debian GNU/Linux 5.0 (lenny)',
    'squeeze': 'Debian GNU/Linux 6.0 (squeeze)',
    'wheezy': 'Debian GNU/Linux 7.0 (wheezy)',
    'jessie': 'Debian testing (jessie)',
    'sid': 'Debian unstable (sid)',
    'hardy': 'Ubuntu 08.04 LTS "Hardy Heron" (hardy)',
    'jaunty': 'Ubuntu 09.04 "Jaunty Jackalope" (jaunty)',
    'karmic': 'Ubuntu 09.10 "Karmic Koala" (karmic)',
    'lucid': 'Ubuntu 10.04 LTS "Lucid Lynx" (lucid)',
    'maverick': 'Ubuntu 10.10 "Maverick Meerkat" (maverick)',
    'natty': 'Ubuntu 11.04 "Natty Narwhal" (natty)',
    'oneiric': 'Ubuntu 11.10 "Oneiric Ocelot" (oneiric)',
    'precise': 'Ubuntu 12.04 LTS "Precise Pangolin" (precise)',
    'quantal': 'Ubuntu 12.10 "Quantal Quetzal" (quantal)',
    'raring': 'Ubuntu 13.04 "Raring Ringtail" (raring)',
    'saucy': 'Ubuntu 13.10 "Saucy Salamander" (saucy)',
    'trusty': 'Ubuntu 14.04 "Trusty Tahr" (trusty)',
}


if __name__ == '__main__':
    data = {}
    # get the IP, date and target release
    # the date is truncated to a month/year combo
    listget = re.compile(r'^([0-9.:]*) .*\[([^:]*).*GET /lists/([a-z]*)')
    for line in fileinput.FileInput(openhook=fileinput.hook_compressed):
        match = listget.match(line)
        if not match:
            continue
        addr, date, release = match.groups()
        if not release in releases:
            # ignore fantasy names
            continue
        date = datetime.strptime(date, '%d/%b/%Y')
        # truncate to a week
        try:
            date = datetime(date.year, date.month, date.day / 7 * 7 + 1)
        except ValueError:
            # only on Feb28...
            date = datetime(date.year, date.month, date.day / 7 * 7)
        # microseconds since epoch
        date = int(time.mktime(date.timetuple()) * 1000)
        rstats = data.setdefault(releases[release], {})
        rtime = rstats.setdefault(date, 0)
        rtime += 1
        rstats[date] = rtime
        data[releases[release]] = rstats
    # determine the union of all timestamps
    timestamps = set()
    for codename, stats in data.iteritems():
        timestamps.update(stats.keys())
    export = [{'key': release,
               'values': [[ts, float(data[release].setdefault(ts, 0)) / 7]
                        for ts in sorted(timestamps)]}
                                    for release in sorted(data)]
    print json.dumps(export)