abydos.phonetic.soundex.soundex()

Complexity

Conditions

13

Size

Total Lines	146
Code Lines	68

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	31
CRAP Score	13

Importance

Changes

0

# -*- coding: utf-8 -*-
def soundex(word, max_length=4, var='American', reverse=False, zero_pad=True):
    """Return the Soundex code for a word.

    :param str word: the word to transform
    :param int max_length: the length of the code returned (defaults to 4)
    :param str var: the variant of the algorithm to employ (defaults to
        'American'):

        - 'American' follows the American Soundex algorithm, as described at
          :cite:`US:2007` and in :cite:`Knuth:1998`; this is also called
          Miracode
        - 'special' follows the rules from the 1880-1910 US Census
          retrospective re-analysis, in which h & w are not treated as blocking
          consonants but as vowels. Cf. :cite:`Repici:2013`.
        - 'Census' follows the rules laid out in GIL 55 :cite:`US:1997` by the
          US Census, including coding prefixed and unprefixed versions of some
          names

    :param bool reverse: reverse the word before computing the selected Soundex
        (defaults to False); This results in "Reverse Soundex", which is useful
        for blocking in cases where the initial elements may be in error.
    :param bool zero_pad: pad the end of the return value with 0s to achieve a
        max_length string
    :returns: the Soundex value
    :rtype: str

    >>> soundex("Christopher")
    'C623'
    >>> soundex("Niall")
    'N400'
    >>> soundex('Smith')
    'S530'
    >>> soundex('Schmidt')
    'S530'

    >>> soundex('Christopher', max_length=-1)
    'C623160000000000000000000000000000000000000000000000000000000000'
    >>> soundex('Christopher', max_length=-1, zero_pad=False)
    'C62316'

    >>> soundex('Christopher', reverse=True)
    'R132'

    >>> soundex('Ashcroft')
    'A261'
    >>> soundex('Asicroft')
    'A226'
    >>> soundex('Ashcroft', var='special')
    'A226'
    >>> soundex('Asicroft', var='special')
    'A226'
    """
    _soundex_translation = dict(
        zip(
            (ord(_) for _ in 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'),

The variable _ does not seem to be defined.

            '01230129022455012623019202',
        )
    )

    # Require a max_length of at least 4 and not more than 64
    if max_length != -1:
        max_length = min(max(4, max_length), 64)
    else:
        max_length = 64

    # uppercase, normalize, decompose, and filter non-A-Z out
    word = unicode_normalize('NFKD', text_type(word.upper()))
    word = word.replace('ß', 'SS')

    if var == 'Census':
        # TODO: Should these prefixes be supplemented? (VANDE, DELA, VON)

TODO and FIXME comments should generally be avoided.

        if word[:3] in {'VAN', 'CON'} and len(word) > 4:
            return (
                soundex(word, max_length, 'American', reverse, zero_pad),
                soundex(word[3:], max_length, 'American', reverse, zero_pad),
            )
        if word[:2] in {'DE', 'DI', 'LA', 'LE'} and len(word) > 3:
            return (
                soundex(word, max_length, 'American', reverse, zero_pad),
                soundex(word[2:], max_length, 'American', reverse, zero_pad),
            )
        # Otherwise, proceed as usual (var='American' mode, ostensibly)

    word = ''.join(
        c
        for c in word
        if c
        in {
            'A',
            'B',
            'C',
            'D',
            'E',
            'F',
            'G',
            'H',
            'I',
            'J',
            'K',
            'L',
            'M',
            'N',
            'O',
            'P',
            'Q',
            'R',
            'S',
            'T',
            'U',
            'V',
            'W',
            'X',
            'Y',
            'Z',
        }
    )

    # Nothing to convert, return base case
    if not word:
        if zero_pad:
            return '0' * max_length
        return '0'

    # Reverse word if computing Reverse Soundex
    if reverse:
        word = word[::-1]

    # apply the Soundex algorithm
    sdx = word.translate(_soundex_translation)

    if var == 'special':
        sdx = sdx.replace('9', '0')  # special rule for 1880-1910 census
    else:
        sdx = sdx.replace('9', '')  # rule 1
    sdx = _delete_consecutive_repeats(sdx)  # rule 3

    if word[0] in 'HW':
        sdx = word[0] + sdx
    else:
        sdx = word[0] + sdx[1:]
    sdx = sdx.replace('0', '')  # rule 1

    if zero_pad:
        sdx += '0' * max_length  # rule 4

    return sdx[:max_length]