normalisename/normalisename.py

#!/usr/bin/env python3
from unidecode import unidecode
from os import rename
from os.path import basename, dirname
from os.path import join as joinpath
from argparse import ArgumentParser
from sys import exit


class normalisename:
    @property
    def separator(self):
        return self._separator

    @property
    def whitelist(self):
        return self._whitelist

    def __init__(self, separator, whitelist):
        self._separator = separator
        self._whitelist = set(whitelist)

    def normalise(self, files):
        for path in files:
            dir = dirname(path)
            file = basename(path)
            rename(path, joinpath(dir, self.normalname(file)))

    def normalname(self, filename):
        return unidecode(''.join(ch for ch in filename.replace(' ', self.separator)
                                 if ch.isalnum()
                                 or ch in self.whitelist))


if __name__ == '__main__':
    # parse arguments
    parser = ArgumentParser(description='Eliminate funky stuff from filenames.')
    parser.add_argument('--separator', '-s', type=str, nargs='?', default='_',
                        help='Set separator to use (to replace spaces with).')
    parser.add_argument('--allow', '-a', type=str, nargs='?', action='append',
                        help='Specify a special character to allow (it will not be removed).')
    parser.add_argument('--disallow', '-d', type=str, nargs='?', action='append',
                        help='Specify a special character to disallow (it will removed).')
    parser.add_argument('--whitelist', '-w', type=str,
                        help='Overwrite default whitelist (format: whitespace separated string).')
    parser.add_argument('files', type=str, nargs='+',
                        help='File(s) to normalise the name of (relative or absolute path).')
    args = parser.parse_args()

    # verify arguments
    if args.disallow:
        if args.separator in args.disallow: exit('Disallowing your chosen separator makes no sense!')

    # declare special characters that will not be removed (spaces are handled elsewhere)
    whitelist = {' ', '.', '-'}
    if args.whitelist: whitelist = set(args.whitelist.split())

    # modify whitelist based on arguments
    whitelist.add(args.separator)
    if args.allow: whitelist = whitelist.union(set(args.allow))
    if args.disallow: whitelist = whitelist.difference(set(args.disallow))


    normalisename(args.separator, whitelist).normalise(args.files)