Format Bank Statement dates to use isoformat

format_statement_filenames.py

import os
import shutil
import re
import datetime
from calendar import monthrange
from collections import OrderedDict
from dateutil.relativedelta import relativedelta

PATH = '/home/lvisintini/Downloads/statement'

YEAR_REG = r'[0-9]{4}'
MONTH_REG = r'(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)'
DAY_REG = r'[0-9]{2}'


class DateParser(object):
    F1 = re.compile(f'^{DAY_REG} {MONTH_REG} - {DAY_REG} {MONTH_REG} {YEAR_REG}$')
    F2 = re.compile(f'^{DAY_REG} {MONTH_REG} {YEAR_REG} - {DAY_REG} {MONTH_REG} {YEAR_REG}$')
    F3 = re.compile(f'^Annual Summary - {DAY_REG} {MONTH_REG} {YEAR_REG} - {DAY_REG} {MONTH_REG} {YEAR_REG}$')

    @classmethod
    def parse(cls, value):
        mapping = OrderedDict()
        mapping[cls.F1] = cls.parse_f1
        mapping[cls.F2] = cls.parse_f2
        mapping[cls.F3] = cls.parse_f3
        for regex, parser in mapping.items():
            if regex.match(value):
                return parser(value)
        else:
            print('Failed to parse date "{}" '.format(value))
            return None, None

    @staticmethod
    def parse_f1(value):
        date0, date1 = value.split(' - ')
        date1 = datetime.datetime.strptime(date1, "%d %b %Y").date()
        date0 = '{} {}'.format(date0, date1.year)
        date0 = datetime.datetime.strptime(date0, "%d %b %Y").date()
        return date0, date1

    @staticmethod
    def parse_f2(value):
        dates = sorted([datetime.datetime.strptime(x, "%d %b %Y").date() for x in value.split(' - ')])
        return dates[0], dates[1]

    @staticmethod
    def parse_f3(value):
        value = value.replace('Annual Summary - ', '')
        dates = sorted([datetime.datetime.strptime(x, "%d %b %Y").date() for x in value.split(' - ')])
        return dates[0], dates[1]


def main():
    new_names = {}
    for dirpath, _, filenames in os.walk(PATH):
        for f in filenames:

            date_string, extension = f.split('.')
            start, end = DateParser.parse(date_string)
            new_name = '{} - {}'.format(start.isoformat(), end.isoformat())
            if 'Annual Summary' in date_string:
                new_name = '{} - {} - Annual Summary'.format(start.isoformat(), end.isoformat())
            new_names[os.path.join(dirpath, f)] = os.path.join(dirpath, f'{new_name}.{extension}')

    for old_name, new_name in new_names.items():
        shutil.copy2(old_name, new_name)


if __name__ == '__main__':
    main()