fighter-programmer/voicegen

1import os2import sys3import re4from pypinyin import lazy_pinyin, BOPOMOFO5import jieba6import cn2an7import logging8 9logging.getLogger('jieba').setLevel(logging.WARNING)10jieba.initialize()11 12 13# List of (Latin alphabet, bopomofo) pairs:14_latin_to_bopomofo = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [15    ('a', 'ㄟˉ'),16    ('b', 'ㄅㄧˋ'),17    ('c', 'ㄙㄧˉ'),18    ('d', 'ㄉㄧˋ'),19    ('e', 'ㄧˋ'),20    ('f', 'ㄝˊㄈㄨˋ'),21    ('g', 'ㄐㄧˋ'),22    ('h', 'ㄝˇㄑㄩˋ'),23    ('i', 'ㄞˋ'),24    ('j', 'ㄐㄟˋ'),25    ('k', 'ㄎㄟˋ'),26    ('l', 'ㄝˊㄛˋ'),27    ('m', 'ㄝˊㄇㄨˋ'),28    ('n', 'ㄣˉ'),29    ('o', 'ㄡˉ'),30    ('p', 'ㄆㄧˉ'),31    ('q', 'ㄎㄧㄡˉ'),32    ('r', 'ㄚˋ'),33    ('s', 'ㄝˊㄙˋ'),34    ('t', 'ㄊㄧˋ'),35    ('u', 'ㄧㄡˉ'),36    ('v', 'ㄨㄧˉ'),37    ('w', 'ㄉㄚˋㄅㄨˋㄌㄧㄡˋ'),38    ('x', 'ㄝˉㄎㄨˋㄙˋ'),39    ('y', 'ㄨㄞˋ'),40    ('z', 'ㄗㄟˋ')41]]42 43# List of (bopomofo, romaji) pairs:44_bopomofo_to_romaji = [(re.compile('%s' % x[0]), x[1]) for x in [45    ('ㄅㄛ', 'p⁼wo'),46    ('ㄆㄛ', 'pʰwo'),47    ('ㄇㄛ', 'mwo'),48    ('ㄈㄛ', 'fwo'),49    ('ㄅ', 'p⁼'),50    ('ㄆ', 'pʰ'),51    ('ㄇ', 'm'),52    ('ㄈ', 'f'),53    ('ㄉ', 't⁼'),54    ('ㄊ', 'tʰ'),55    ('ㄋ', 'n'),56    ('ㄌ', 'l'),57    ('ㄍ', 'k⁼'),58    ('ㄎ', 'kʰ'),59    ('ㄏ', 'h'),60    ('ㄐ', 'ʧ⁼'),61    ('ㄑ', 'ʧʰ'),62    ('ㄒ', 'ʃ'),63    ('ㄓ', 'ʦ`⁼'),64    ('ㄔ', 'ʦ`ʰ'),65    ('ㄕ', 's`'),66    ('ㄖ', 'ɹ`'),67    ('ㄗ', 'ʦ⁼'),68    ('ㄘ', 'ʦʰ'),69    ('ㄙ', 's'),70    ('ㄚ', 'a'),71    ('ㄛ', 'o'),72    ('ㄜ', 'ə'),73    ('ㄝ', 'e'),74    ('ㄞ', 'ai'),75    ('ㄟ', 'ei'),76    ('ㄠ', 'au'),77    ('ㄡ', 'ou'),78    ('ㄧㄢ', 'yeNN'),79    ('ㄢ', 'aNN'),80    ('ㄧㄣ', 'iNN'),81    ('ㄣ', 'əNN'),82    ('ㄤ', 'aNg'),83    ('ㄧㄥ', 'iNg'),84    ('ㄨㄥ', 'uNg'),85    ('ㄩㄥ', 'yuNg'),86    ('ㄥ', 'əNg'),87    ('ㄦ', 'əɻ'),88    ('ㄧ', 'i'),89    ('ㄨ', 'u'),90    ('ㄩ', 'ɥ'),91    ('ˉ', '→'),92    ('ˊ', '↑'),93    ('ˇ', '↓↑'),94    ('ˋ', '↓'),95    ('˙', ''),96    ('，', ','),97    ('。', '.'),98    ('！', '!'),99    ('？', '?'),100    ('—', '-')101]]102 103# List of (romaji, ipa) pairs:104_romaji_to_ipa = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [105    ('ʃy', 'ʃ'),106    ('ʧʰy', 'ʧʰ'),107    ('ʧ⁼y', 'ʧ⁼'),108    ('NN', 'n'),109    ('Ng', 'ŋ'),110    ('y', 'j'),111    ('h', 'x')112]]113 114# List of (bopomofo, ipa) pairs:115_bopomofo_to_ipa = [(re.compile('%s' % x[0]), x[1]) for x in [116    ('ㄅㄛ', 'p⁼wo'),117    ('ㄆㄛ', 'pʰwo'),118    ('ㄇㄛ', 'mwo'),119    ('ㄈㄛ', 'fwo'),120    ('ㄅ', 'p⁼'),121    ('ㄆ', 'pʰ'),122    ('ㄇ', 'm'),123    ('ㄈ', 'f'),124    ('ㄉ', 't⁼'),125    ('ㄊ', 'tʰ'),126    ('ㄋ', 'n'),127    ('ㄌ', 'l'),128    ('ㄍ', 'k⁼'),129    ('ㄎ', 'kʰ'),130    ('ㄏ', 'x'),131    ('ㄐ', 'tʃ⁼'),132    ('ㄑ', 'tʃʰ'),133    ('ㄒ', 'ʃ'),134    ('ㄓ', 'ts`⁼'),135    ('ㄔ', 'ts`ʰ'),136    ('ㄕ', 's`'),137    ('ㄖ', 'ɹ`'),138    ('ㄗ', 'ts⁼'),139    ('ㄘ', 'tsʰ'),140    ('ㄙ', 's'),141    ('ㄚ', 'a'),142    ('ㄛ', 'o'),143    ('ㄜ', 'ə'),144    ('ㄝ', 'ɛ'),145    ('ㄞ', 'aɪ'),146    ('ㄟ', 'eɪ'),147    ('ㄠ', 'ɑʊ'),148    ('ㄡ', 'oʊ'),149    ('ㄧㄢ', 'jɛn'),150    ('ㄩㄢ', 'ɥæn'),151    ('ㄢ', 'an'),152    ('ㄧㄣ', 'in'),153    ('ㄩㄣ', 'ɥn'),154    ('ㄣ', 'ən'),155    ('ㄤ', 'ɑŋ'),156    ('ㄧㄥ', 'iŋ'),157    ('ㄨㄥ', 'ʊŋ'),158    ('ㄩㄥ', 'jʊŋ'),159    ('ㄥ', 'əŋ'),160    ('ㄦ', 'əɻ'),161    ('ㄧ', 'i'),162    ('ㄨ', 'u'),163    ('ㄩ', 'ɥ'),164    ('ˉ', '→'),165    ('ˊ', '↑'),166    ('ˇ', '↓↑'),167    ('ˋ', '↓'),168    ('˙', ''),169    ('，', ','),170    ('。', '.'),171    ('！', '!'),172    ('？', '?'),173    ('—', '-')174]]175 176# List of (bopomofo, ipa2) pairs:177_bopomofo_to_ipa2 = [(re.compile('%s' % x[0]), x[1]) for x in [178    ('ㄅㄛ', 'pwo'),179    ('ㄆㄛ', 'pʰwo'),180    ('ㄇㄛ', 'mwo'),181    ('ㄈㄛ', 'fwo'),182    ('ㄅ', 'p'),183    ('ㄆ', 'pʰ'),184    ('ㄇ', 'm'),185    ('ㄈ', 'f'),186    ('ㄉ', 't'),187    ('ㄊ', 'tʰ'),188    ('ㄋ', 'n'),189    ('ㄌ', 'l'),190    ('ㄍ', 'k'),191    ('ㄎ', 'kʰ'),192    ('ㄏ', 'h'),193    ('ㄐ', 'tɕ'),194    ('ㄑ', 'tɕʰ'),195    ('ㄒ', 'ɕ'),196    ('ㄓ', 'tʂ'),197    ('ㄔ', 'tʂʰ'),198    ('ㄕ', 'ʂ'),199    ('ㄖ', 'ɻ'),200    ('ㄗ', 'ts'),201    ('ㄘ', 'tsʰ'),202    ('ㄙ', 's'),203    ('ㄚ', 'a'),204    ('ㄛ', 'o'),205    ('ㄜ', 'ɤ'),206    ('ㄝ', 'ɛ'),207    ('ㄞ', 'aɪ'),208    ('ㄟ', 'eɪ'),209    ('ㄠ', 'ɑʊ'),210    ('ㄡ', 'oʊ'),211    ('ㄧㄢ', 'jɛn'),212    ('ㄩㄢ', 'yæn'),213    ('ㄢ', 'an'),214    ('ㄧㄣ', 'in'),215    ('ㄩㄣ', 'yn'),216    ('ㄣ', 'ən'),217    ('ㄤ', 'ɑŋ'),218    ('ㄧㄥ', 'iŋ'),219    ('ㄨㄥ', 'ʊŋ'),220    ('ㄩㄥ', 'jʊŋ'),221    ('ㄥ', 'ɤŋ'),222    ('ㄦ', 'əɻ'),223    ('ㄧ', 'i'),224    ('ㄨ', 'u'),225    ('ㄩ', 'y'),226    ('ˉ', '˥'),227    ('ˊ', '˧˥'),228    ('ˇ', '˨˩˦'),229    ('ˋ', '˥˩'),230    ('˙', ''),231    ('，', ','),232    ('。', '.'),233    ('！', '!'),234    ('？', '?'),235    ('—', '-')236]]237 238 239def number_to_chinese(text):240    numbers = re.findall(r'\d+(?:\.?\d+)?', text)241    for number in numbers:242        text = text.replace(number, cn2an.an2cn(number), 1)243    return text244 245 246def chinese_to_bopomofo(text):247    text = text.replace('、', '，').replace('；', '，').replace('：', '，')248    words = jieba.lcut(text, cut_all=False)249    text = ''250    for word in words:251        bopomofos = lazy_pinyin(word, BOPOMOFO)252        if not re.search('[\u4e00-\u9fff]', word):253            text += word254            continue255        for i in range(len(bopomofos)):256            bopomofos[i] = re.sub(r'([\u3105-\u3129])$', r'\1ˉ', bopomofos[i])257        if text != '':258            text += ' '259        text += ''.join(bopomofos)260    return text261 262 263def latin_to_bopomofo(text):264    for regex, replacement in _latin_to_bopomofo:265        text = re.sub(regex, replacement, text)266    return text267 268 269def bopomofo_to_romaji(text):270    for regex, replacement in _bopomofo_to_romaji:271        text = re.sub(regex, replacement, text)272    return text273 274 275def bopomofo_to_ipa(text):276    for regex, replacement in _bopomofo_to_ipa:277        text = re.sub(regex, replacement, text)278    return text279 280 281def bopomofo_to_ipa2(text):282    for regex, replacement in _bopomofo_to_ipa2:283        text = re.sub(regex, replacement, text)284    return text285 286 287def chinese_to_romaji(text):288    text = number_to_chinese(text)289    text = chinese_to_bopomofo(text)290    text = latin_to_bopomofo(text)291    text = bopomofo_to_romaji(text)292    text = re.sub('i([aoe])', r'y\1', text)293    text = re.sub('u([aoəe])', r'w\1', text)294    text = re.sub('([ʦsɹ]`[⁼ʰ]?)([→↓↑ ]+|$)',295                  r'\1ɹ`\2', text).replace('ɻ', 'ɹ`')296    text = re.sub('([ʦs][⁼ʰ]?)([→↓↑ ]+|$)', r'\1ɹ\2', text)297    return text298 299 300def chinese_to_lazy_ipa(text):301    text = chinese_to_romaji(text)302    for regex, replacement in _romaji_to_ipa:303        text = re.sub(regex, replacement, text)304    return text305 306 307def chinese_to_ipa(text):308    text = number_to_chinese(text)309    text = chinese_to_bopomofo(text)310    text = latin_to_bopomofo(text)311    text = bopomofo_to_ipa(text)312    text = re.sub('i([aoe])', r'j\1', text)313    text = re.sub('u([aoəe])', r'w\1', text)314    text = re.sub('([sɹ]`[⁼ʰ]?)([→↓↑ ]+|$)',315                  r'\1ɹ`\2', text).replace('ɻ', 'ɹ`')316    text = re.sub('([s][⁼ʰ]?)([→↓↑ ]+|$)', r'\1ɹ\2', text)317    return text318 319 320def chinese_to_ipa2(text):321    text = number_to_chinese(text)322    text = chinese_to_bopomofo(text)323    text = latin_to_bopomofo(text)324    text = bopomofo_to_ipa2(text)325    text = re.sub(r'i([aoe])', r'j\1', text)326    text = re.sub(r'u([aoəe])', r'w\1', text)327    text = re.sub(r'([ʂɹ]ʰ?)([˩˨˧˦˥ ]+|$)', r'\1ʅ\2', text)328    text = re.sub(r'(sʰ?)([˩˨˧˦˥ ]+|$)', r'\1ɿ\2', text)329    return text