Backups weeding script

weeding.py

#!/usr/bin/python
# -*- coding: utf8 -*-

import os
import re
import datetime
from calendar import Calendar
from itertools import chain

KEEP_DAY = 0

def make_test_files(num=200):
    now = datetime.datetime.now()

    for i in range(num):
        date = datetime.datetime.now() - datetime.timedelta(days=i)
        date = date.strftime("%Y-%m-%d")
        open('file.%s.ext' % date, 'w').close()

def cleanup(directory='.'):
    today = datetime.date.today()
    week_ago = today - datetime.timedelta(days=7)
    month_ago = today - datetime.timedelta(days=31)

    calendar = Calendar()

    re_file = r'^(.+)[-.]([0-9]{4})-([0-9]{2})-([0-9]{2})\.(.+)$'
    _files = os.listdir(directory)

    file_dates = {}
    oldest_files = {}
    files = []

    for file_name in _files:
        match = re.search(re_file, file_name)
        if not match:
            continue

        year, month, day =  ( int(x) for x
                                in ( match.group(x) for x in (2, 3, 4)))
        date = datetime.date(year, month, day)
        group = tuple(match.group(x) for x in (1, 5))

        file_dates.setdefault(group, []).append(date)

        files.append((file_name, date, group))

    for group, dates in file_dates.items():
        oldest_files[group] = min(dates)

    for file_name, date, group in files:
        if date == oldest_files[group]:
            # skip oldest file
            continue

        if date > week_ago:
            # skip this week files
            continue

        if date > month_ago:
            # process this month files
            if date.weekday() != KEEP_DAY :
                os.unlink(file_name)
            continue

        keep_day = min([ x for x
                            in list(chain(*calendar.monthdatescalendar(date.year, date.month)))
                            if x.month==date.month and x.weekday() == KEEP_DAY])

        if date != keep_day:
            os.unlink(file_name)

#make_test_files()
cleanup()