pokedex/roomaji.py

   1 # encoding: utf8
   2 """Provides `romanize()` for romanizing simple Japanese text."""
   3
   4 _roomaji_kana = {
   5     u'ア': 'a',     u'イ': 'i',     u'ウ': 'u',     u'エ': 'e',     u'オ': 'o',
   6     u'カ': 'ka',    u'キ': 'ki',    u'ク': 'ku',    u'ケ': 'ke',    u'コ': 'ko',
   7     u'サ': 'sa',    u'シ': 'shi',   u'ス': 'su',    u'セ': 'se',    u'ソ': 'so',
   8     u'タ': 'ta',    u'チ': 'chi',   u'ツ': 'tsu',   u'テ': 'te',    u'ト': 'to',
   9     u'ナ': 'na',    u'ニ': 'ni',    u'ヌ': 'nu',    u'ネ': 'ne',    u'ノ': 'no',
  10     u'ハ': 'ha',    u'ヒ': 'hi',    u'フ': 'fu',    u'ヘ': 'he',    u'ホ': 'ho',
  11     u'マ': 'ma',    u'ミ': 'mi',    u'ム': 'mu',    u'メ': 'me',    u'モ': 'mo',
  12     u'ヤ': 'ya',                    u'ユ': 'yu',                    u'ヨ': 'yo',
  13     u'ラ': 'ra',    u'リ': 'ri',    u'ル': 'ru',    u'レ': 're',    u'ロ': 'ro',
  14     u'ワ': 'wa',    u'ヰ': 'wi',                    u'ヱ': 'we',    u'ヲ': 'wo',
  15                                                                     u'ン': 'n',
  16     u'ガ': 'ga',    u'ギ': 'gi',    u'グ': 'gu',    u'ゲ': 'ge',    u'ゴ': 'go',
  17     u'ザ': 'za',    u'ジ': 'ji',    u'ズ': 'zu',    u'ゼ': 'ze',    u'ゾ': 'zo',
  18     u'ダ': 'da',    u'ヂ': 'ji',    u'ヅ': 'dzu',   u'デ': 'de',    u'ド': 'do',
  19     u'バ': 'ba',    u'ビ': 'bi',    u'ブ': 'bu',    u'ベ': 'be',    u'ボ': 'bo',
  20     u'パ': 'pa',    u'ピ': 'pi',    u'プ': 'pu',    u'ペ': 'pe',    u'ポ': 'po',
  21 }
  22
  23 _roomaji_youon = {
  24     u'ャ': 'ya',                    u'ュ': 'yu',                    u'ョ': 'yo',
  25     u'ゃ': 'ya',                    u'ゅ': 'yu',                    u'ょ': 'yo',
  26 }
  27
  28 # XXX If romanize() ever handles hiragana, it will need to make sure that the
  29 # preceding character was a katakana
  30 # This does not include every small kana combination, but should include every
  31 # one used in a Pokémon name.  An exhaustive list would be..  very long
  32 _roomaji_small_kana = {
  33     u'ァ': 'a',     u'ィ': 'i',     u'ゥ': 'u',     u'ェ': 'e',     u'ォ': 'o',
  34 }
  35 _roomaji_small_kana_combos = {
  36                                                     u'チェ': 'che',
  37                                                     u'シェ': 'she',
  38     u'テァ': 'tha', u'ティ': 'ti',  u'テゥ': 'thu', u'テェ': 'tye', u'テォ': 'tho',
  39     u'デァ': 'dha', u'ディ': 'di',  u'デゥ': 'dhu', u'デェ': 'dye', u'デォ': 'dho',
  40     u'ファ': 'fa',  u'フィ': 'fi',  u'ホゥ': 'hu',  u'フェ': 'fe',  u'フォ': 'fo',
  41 }
  42
  43 def romanize(string):
  44     """Converts a string of kana to roomaji."""
  45
  46     vowels = ['a', 'e', 'i', 'o', 'u', 'y']
  47
  48     characters = []
  49     last_kana = None  # Used for ー; っ or ッ; ん or ン
  50     last_char = None  # Used for small kana combos
  51     for char in string:
  52         # Full-width Latin
  53         if ord(char) >= 0xff11 and ord(char) <= 0xff5e:
  54             if last_kana == 'sokuon':
  55                 raise ValueError("Sokuon cannot precede Latin characters.")
  56
  57             char = chr(ord(char) - 0xff11 + 0x31)
  58             characters.append(char)
  59
  60             last_kana = None
  61
  62         # Small vowel kana
  63         elif char in _roomaji_small_kana:
  64             combo = last_char + char
  65             if combo in _roomaji_small_kana_combos:
  66                 characters[-1] = _roomaji_small_kana_combos[combo]
  67
  68             else:
  69                 # If we don't know what it is...  act dumb and treat it as a
  70                 # full-size vowel.  Better than bailing, and seems to occur a
  71                 # lot, e.g. ピィ is "pii"
  72                 characters.append(_roomaji_small_kana[char])
  73
  74         # Youon
  75         elif char in _roomaji_youon:
  76             if last_kana[-1] != 'i' or last_kana == 'i':
  77                 raise ValueError("Youon must follow an -i sound.")
  78
  79             # Drop the -i and append the ya/yu/yo sound
  80             new_sound = _roomaji_youon[char]
  81             if last_kana in ['shi', 'ji']:
  82                 # Strip the y-
  83                 new_char = last_kana[:-1] + new_sound[1:]
  84             else:
  85                 new_char = last_kana[:-1] + new_sound
  86
  87             characters[-1] = new_char
  88             last_kana = new_char
  89
  90         # Sokuon
  91         elif char in (u'っ', u'ッ'):
  92             # Remember it and double the consonant next time around
  93             last_kana = 'sokuon'
  94
  95         # Extended vowel or n
  96         elif char == u'ー':
  97             if last_kana[-1] not in vowels:
  98                 raise ValueError(u"'ー' must follow by a vowel.")
  99             characters.append(last_kana[-1])
 100
 101             last_kana = None
 102
 103         # Regular ol' kana
 104         elif char in _roomaji_kana:
 105             kana = _roomaji_kana[char]
 106
 107             if last_kana == 'sokuon':
 108                 if kana[0] in vowels:
 109                     raise ValueError("Sokuon cannot precede a vowel.")
 110
 111                 characters.append(kana[0])
 112             elif last_kana == 'n' and kana[0] in vowels:
 113                 characters.append("'")
 114
 115             characters.append(kana)
 116
 117             last_kana = kana
 118
 119         # Not Japanese?
 120         else:
 121             if last_kana == 'sokuon':
 122                 raise ValueError("Sokuon must be followed by another kana.")
 123
 124             characters.append(char)
 125
 126             last_kana = None
 127
 128         last_char = char
 129
 130
 131     if last_kana == 'sokuon':
 132         raise ValueError("Sokuon cannot be the last character.")
 133
 134     return unicode(''.join(characters))