#!/usr/bin/python2

"""
This script reads all the spec files in the folders specified and extract
the date of the most recent entry that wasn't created by rel-eng or Dennis
Gilmore.

"""


import argparse
import collections
import re
import os
import sys

import requests


_date_regex = '(?:mon|tue|wed|thu|fri|sat|sun)\s+'\
              '(?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\s+'\
              '\d\d?\s+'\
              '\d\d\d\d'

_months = {
    'jan': 1,
    'feb': 2,
    'mar': 3,
    'apr': 4,
    'may': 5,
    'jun': 6,
    'jul': 7,
    'aug': 8,
    'sep': 9,
    'oct': 10,
    'nov': 11,
    'dec': 12
}


def _get_retired_pkgs():
    pdc_url = 'https://pdc.fedoraproject.org/rest_api/v1/'\
        'component-branches/?active=false&type=rpm&name=master'
    pkgs = set()
    while pdc_url:
        print(pdc_url)
        req = requests.get(pdc_url)
        data = req.json()
        for res in data['results']:
            pkgs.add(res['global_component'])

        pdc_url = data['next']

    print('%s packages retired found' % len(pkgs))
    return pkgs


def parse_args():
    parser = argparse.ArgumentParser(
        description='Get stats about most recent human-action')
    parser.add_argument(
        'target', help='Folder where are present the spec files')

    return parser.parse_args()


def get_most_recent_entry(specfile):
    """ Return the date of the most recent entry in the specified specfiled
    where this entry was not created by rel-eng or Dennis Gilmore.
    """
    with open(specfile, 'r') as stream:
        spec = [row.strip() for row in stream.readlines()]

    changelog = False
    output = None
    for row in spec:
        if '%changelog' in row:
            changelog = True
            continue

        if changelog:
            dates = re.findall(_date_regex, row, flags=re.I)
            if dates:
                if 'fedora release engineering' in row.lower():
                    continue
                if 'dennis gilmore' in row.lower():
                    continue
                date = dates[0]
                month = '%s-%s-1' % (
                    date.split()[3], _months[date.split()[1].lower()])
                output = (month, '%s -- %s' % (specfile, row))
                break

        if output:
            break

    return output


def main():
    args = parse_args()

    retired_pkgs = _get_retired_pkgs()

    stats = collections.defaultdict(list)

    for dirpath, dirnames, filenames in os.walk(args.target):
        for filename in filenames:
            if filename.endswith('.spec'):
                if filename.split('.spec')[0] in retired_pkgs:
                    print('  - retired: %s' % filename)
                    continue
                specfile = os.path.join(dirpath, filename)
                entry = get_most_recent_entry(specfile)
                if entry is None:
                    stats['unknown'].append(specfile)
                else:
                    simple_date, row = entry
                    stats[simple_date].append(row)

    # print(stats.keys())
    for date in sorted(stats):
        print(date, len(stats[date]))

    with open('spec_stats.csv', 'w') as stream:
        stream.write('Date, entry\n')
        for date in sorted(stats):
            stream.write('%s,%s\n' % (date, len(stats[date])))

    with open('spec_stats_details.csv', 'w') as stream:
        stream.write('Date, entry\n')
        for date in sorted(stats):
            for entry in stats[date]:
                stream.write('%s,%s\n' % (date, entry))


if __name__ == '__main__':
    sys.exit(main())