rusofobica generator

gistfile1.py

import unicodedata as ud

caron = '\N{COMBINING CARON}'
cedilla = '\N{COMBINING CEDILLA}'
umlaut = '\N{COMBINING DIAERESIS}' # Really?
circumflex = '\N{COMBINING CIRCUMFLEX ACCENT}'

consonants = {
    'б': 'b',
    'в': 'v',
    'г': 'g',
    'д': 'd',
    'ж': 'z' + caron,
    'з': 'z',
    'й': 'y',
    'к': 'k',
    'л': 'l',
    'м': 'm',
    'н': 'n',
    'п': 'p',
    'р': 'r',
    'с': 's',
    'т': 't',
    'ф': 'f',
    'х': 'h',
    'ц': 'c',
    'ч': 'c' + caron,
    'ш': 's' + caron,
    'щ': 's' + caron + cedilla,
}

vowels = {
    'а': 'a',
    'у': 'u',
    'о': 'o',
    'ы': 'i' + circumflex,
    'и': 'i',
    'э': 'e',
    'я': 'a' + umlaut,
    'ю': 'u' + umlaut,
    'ё': 'o' + umlaut,
    'е': 'e' + umlaut,
    # Not really vowels, but
    'ъ': '',
    'ь': ''
}

soft_hard_pairs = (
    ('я', 'а'),
    ('ю', 'у'),
    ('ё', 'о'),
    ('е', 'э'),
    # Some kind of hack
    ('ь', 'ь'),
)
    

norm = lambda c: ud.normalize('NFKC', c) # Maybe NFC?


def make_pair(soft, hard):
    """Return function that creates pair of consonant + vowel with cedilla."""
    def inner(cyr, rus):
        cyrillic = '{}{}'.format(cyr, soft)
        if cyr != 'щ':
            rusofobic = '{}{}'.format(norm(rus + cedilla), vowels[hard])
        else:
            rusofobic = '{}{}'.format(norm(rus), vowels[soft])
        return [cyrillic, rusofobic]
    return inner


def make_rules():
    """Create mapping between cyrillic and rusofobic."""
    rules = {}
    # Normalize consonants...
    for cyr, rus in consonants.items():
        rules[cyr] = norm(rus)
    # ...and vowels
    for cyr, rus in vowels.items():
        rules[cyr] = norm(rus)

    # And Ъ
    for cyr, rus in consonants.items():
        rules['{}ъ'.format(cyr)] = norm(rus)

    # Create pairs of consonant + vowel with cedilla
    for pair in soft_hard_pairs:
        func = make_pair(*pair)
        result = [func(cyr, rus) for cyr, rus in consonants.items()]
        for cyr, rus in consonants.items():
            pair = func(cyr, rus)
            rules[pair[0]] = pair[1]
    return rules


def make_table(rules):
    """Create transliteration mapping for all combinations of 
    capitalized and non capitalized characters."""
    table = {}
    for cyr, rus in rules.items():
        table[cyr] = rus
        table[cyr.upper()] = rus.upper()
        if len(cyr) == 2 and len(rus) == 2:
            table[cyr.capitalize()] = rus.capitalize()
            table['{}{}'.format(cyr[0], cyr[1].upper())] = \
                    '{}{}'.format(rus[0], rus[1].upper())
    return table


if __name__ == '__main__':
    # Generate js object
    tbl = make_table(make_rules())
    # Sorted pairs
    pairs = ["'{}': '{}'".format(k, tbl[k]) for k in sorted(tbl.keys())]
    print('{{\n{}\n}}'.format(',\n'.join(pairs)))