|
""" from https://github.com/keithito/tacotron """ |
|
|
|
''' |
|
Cleaners are transformations that run over the input text at both training and eval time. |
|
|
|
Cleaners can be selected by passing a comma-delimited list of cleaner names as the "cleaners" |
|
hyperparameter. Some cleaners are English-specific. You'll typically want to use: |
|
1. "english_cleaners" for English text |
|
2. "transliteration_cleaners" for non-English text that can be transliterated to ASCII using |
|
the Unidecode library (https://pypi.python.org/pypi/Unidecode) |
|
3. "basic_cleaners" if you do not want to transliterate (in this case, you should also update |
|
the symbols in symbols.py to match your data). |
|
''' |
|
|
|
import re |
|
from unidecode import unidecode |
|
import jieba |
|
import pyopenjtalk |
|
from jamo import h2j, j2hcj |
|
from pypinyin import lazy_pinyin,BOPOMOFO |
|
|
|
|
|
|
|
|
|
_korean_classifiers = '๊ตฐ๋ฐ ๊ถ ๊ฐ ๊ทธ๋ฃจ ๋ข ๋ ๋ ๋ง๋ฆฌ ๋ชจ ๋ชจ๊ธ ๋ญ ๋ฐ ๋ฐ์ง ๋ฐฉ ๋ฒ ๋ฒ ๋ณด๋ฃจ ์ด ์ ์ ์ ์ ์ํผ ์ ์ง ์ฑ ์ฒ ์ฒฉ ์ถ ์ผค๋ ํจ ํต' |
|
|
|
|
|
_whitespace_re = re.compile(r'\s+') |
|
|
|
|
|
_japanese_characters = re.compile(r'[A-Za-z\d\u3005\u3040-\u30ff\u4e00-\u9fff\uff11-\uff19\uff21-\uff3a\uff41-\uff5a\uff66-\uff9d]') |
|
|
|
|
|
_japanese_marks = re.compile(r'[^A-Za-z\d\u3005\u3040-\u30ff\u4e00-\u9fff\uff11-\uff19\uff21-\uff3a\uff41-\uff5a\uff66-\uff9d]') |
|
|
|
|
|
_abbreviations = [(re.compile('\\b%s\\.' % x[0], re.IGNORECASE), x[1]) for x in [ |
|
('mrs', 'misess'), |
|
('mr', 'mister'), |
|
('dr', 'doctor'), |
|
('st', 'saint'), |
|
('co', 'company'), |
|
('jr', 'junior'), |
|
('maj', 'major'), |
|
('gen', 'general'), |
|
('drs', 'doctors'), |
|
('rev', 'reverend'), |
|
('lt', 'lieutenant'), |
|
('hon', 'honorable'), |
|
('sgt', 'sergeant'), |
|
('capt', 'captain'), |
|
('esq', 'esquire'), |
|
('ltd', 'limited'), |
|
('col', 'colonel'), |
|
('ft', 'fort'), |
|
]] |
|
|
|
|
|
_hangul_divided = [(re.compile('%s' % x[0]), x[1]) for x in [ |
|
('ใณ', 'ใฑใ
'), |
|
('ใต', 'ใดใ
'), |
|
('ใถ', 'ใดใ
'), |
|
('ใบ', 'ในใฑ'), |
|
('ใป', 'ในใ
'), |
|
('ใผ', 'ในใ
'), |
|
('ใฝ', 'ในใ
'), |
|
('ใพ', 'ในใ
'), |
|
('ใฟ', 'ในใ
'), |
|
('ใ
', 'ในใ
'), |
|
('ใ
', 'ใ
ใ
'), |
|
('ใ
', 'ใ
ใ
'), |
|
('ใ
', 'ใ
ใ
'), |
|
('ใ
', 'ใ
ใ
ฃ'), |
|
('ใ
', 'ใ
ใ
'), |
|
('ใ
', 'ใ
ใ
'), |
|
('ใ
', 'ใ
ใ
ฃ'), |
|
('ใ
ข', 'ใ
กใ
ฃ'), |
|
('ใ
', 'ใ
ฃใ
'), |
|
('ใ
', 'ใ
ฃใ
'), |
|
('ใ
', 'ใ
ฃใ
'), |
|
('ใ
', 'ใ
ฃใ
'), |
|
('ใ
', 'ใ
ฃใ
'), |
|
('ใ
', 'ใ
ฃใ
') |
|
]] |
|
|
|
|
|
_latin_to_hangul = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [ |
|
('a', '์์ด'), |
|
('b', '๋น'), |
|
('c', '์'), |
|
('d', '๋'), |
|
('e', '์ด'), |
|
('f', '์ํ'), |
|
('g', '์ง'), |
|
('h', '์์ด์น'), |
|
('i', '์์ด'), |
|
('j', '์ ์ด'), |
|
('k', '์ผ์ด'), |
|
('l', '์'), |
|
('m', '์ '), |
|
('n', '์'), |
|
('o', '์ค'), |
|
('p', 'ํผ'), |
|
('q', 'ํ'), |
|
('r', '์๋ฅด'), |
|
('s', '์์ค'), |
|
('t', 'ํฐ'), |
|
('u', '์ '), |
|
('v', '๋ธ์ด'), |
|
('w', '๋๋ธ์ '), |
|
('x', '์์ค'), |
|
('y', '์์ด'), |
|
('z', '์ ํธ') |
|
]] |
|
|
|
|
|
def expand_abbreviations(text): |
|
for regex, replacement in _abbreviations: |
|
text = re.sub(regex, replacement, text) |
|
return text |
|
|
|
|
|
def lowercase(text): |
|
return text.lower() |
|
|
|
|
|
def collapse_whitespace(text): |
|
return re.sub(_whitespace_re, ' ', text) |
|
|
|
|
|
def convert_to_ascii(text): |
|
return unidecode(text) |
|
|
|
|
|
def latin_to_hangul(text): |
|
for regex, replacement in _latin_to_hangul: |
|
text = re.sub(regex, replacement, text) |
|
return text |
|
|
|
|
|
def divide_hangul(text): |
|
for regex, replacement in _hangul_divided: |
|
text = re.sub(regex, replacement, text) |
|
return text |
|
|
|
|
|
def hangul_number(num, sino=True): |
|
'''Reference https://github.com/Kyubyong/g2pK''' |
|
num = re.sub(',', '', num) |
|
|
|
if num == '0': |
|
return '์' |
|
if not sino and num == '20': |
|
return '์ค๋ฌด' |
|
|
|
digits = '123456789' |
|
names = '์ผ์ด์ผ์ฌ์ค์ก์น ํ๊ตฌ' |
|
digit2name = {d: n for d, n in zip(digits, names)} |
|
|
|
modifiers = 'ํ ๋ ์ธ ๋ค ๋ค์ฏ ์ฌ์ฏ ์ผ๊ณฑ ์ฌ๋ ์ํ' |
|
decimals = '์ด ์ค๋ฌผ ์๋ฅธ ๋งํ ์ฐ ์์ ์ผํ ์ฌ๋ ์ํ' |
|
digit2mod = {d: mod for d, mod in zip(digits, modifiers.split())} |
|
digit2dec = {d: dec for d, dec in zip(digits, decimals.split())} |
|
|
|
spelledout = [] |
|
for i, digit in enumerate(num): |
|
i = len(num) - i - 1 |
|
if sino: |
|
if i == 0: |
|
name = digit2name.get(digit, '') |
|
elif i == 1: |
|
name = digit2name.get(digit, '') + '์ญ' |
|
name = name.replace('์ผ์ญ', '์ญ') |
|
else: |
|
if i == 0: |
|
name = digit2mod.get(digit, '') |
|
elif i == 1: |
|
name = digit2dec.get(digit, '') |
|
if digit == '0': |
|
if i % 4 == 0: |
|
last_three = spelledout[-min(3, len(spelledout)):] |
|
if ''.join(last_three) == '': |
|
spelledout.append('') |
|
continue |
|
else: |
|
spelledout.append('') |
|
continue |
|
if i == 2: |
|
name = digit2name.get(digit, '') + '๋ฐฑ' |
|
name = name.replace('์ผ๋ฐฑ', '๋ฐฑ') |
|
elif i == 3: |
|
name = digit2name.get(digit, '') + '์ฒ' |
|
name = name.replace('์ผ์ฒ', '์ฒ') |
|
elif i == 4: |
|
name = digit2name.get(digit, '') + '๋ง' |
|
name = name.replace('์ผ๋ง', '๋ง') |
|
elif i == 5: |
|
name = digit2name.get(digit, '') + '์ญ' |
|
name = name.replace('์ผ์ญ', '์ญ') |
|
elif i == 6: |
|
name = digit2name.get(digit, '') + '๋ฐฑ' |
|
name = name.replace('์ผ๋ฐฑ', '๋ฐฑ') |
|
elif i == 7: |
|
name = digit2name.get(digit, '') + '์ฒ' |
|
name = name.replace('์ผ์ฒ', '์ฒ') |
|
elif i == 8: |
|
name = digit2name.get(digit, '') + '์ต' |
|
elif i == 9: |
|
name = digit2name.get(digit, '') + '์ญ' |
|
elif i == 10: |
|
name = digit2name.get(digit, '') + '๋ฐฑ' |
|
elif i == 11: |
|
name = digit2name.get(digit, '') + '์ฒ' |
|
elif i == 12: |
|
name = digit2name.get(digit, '') + '์กฐ' |
|
elif i == 13: |
|
name = digit2name.get(digit, '') + '์ญ' |
|
elif i == 14: |
|
name = digit2name.get(digit, '') + '๋ฐฑ' |
|
elif i == 15: |
|
name = digit2name.get(digit, '') + '์ฒ' |
|
spelledout.append(name) |
|
return ''.join(elem for elem in spelledout) |
|
|
|
|
|
def number_to_hangul(text): |
|
'''Reference https://github.com/Kyubyong/g2pK''' |
|
tokens = set(re.findall(r'(\d[\d,]*)([\uac00-\ud71f]+)', text)) |
|
for token in tokens: |
|
num, classifier = token |
|
if classifier[:2] in _korean_classifiers or classifier[0] in _korean_classifiers: |
|
spelledout = hangul_number(num, sino=False) |
|
else: |
|
spelledout = hangul_number(num, sino=True) |
|
text = text.replace(f'{num}{classifier}', f'{spelledout}{classifier}') |
|
|
|
digits = '0123456789' |
|
names = '์์ผ์ด์ผ์ฌ์ค์ก์น ํ๊ตฌ' |
|
for d, n in zip(digits, names): |
|
text = text.replace(d, n) |
|
return text |
|
|
|
|
|
def basic_cleaners(text): |
|
'''Basic pipeline that lowercases and collapses whitespace without transliteration.''' |
|
text = lowercase(text) |
|
text = collapse_whitespace(text) |
|
return text |
|
|
|
|
|
def transliteration_cleaners(text): |
|
'''Pipeline for non-English text that transliterates to ASCII.''' |
|
text = convert_to_ascii(text) |
|
text = lowercase(text) |
|
text = collapse_whitespace(text) |
|
return text |
|
|
|
|
|
def japanese_cleaners(text): |
|
'''Pipeline for notating accent in Japanese text. |
|
Reference https://r9y9.github.io/ttslearn/latest/notebooks/ch10_Recipe-Tacotron.html''' |
|
sentences = re.split(_japanese_marks, text) |
|
marks = re.findall(_japanese_marks, text) |
|
text = '' |
|
for i, sentence in enumerate(sentences): |
|
if re.match(_japanese_characters, sentence): |
|
if text!='': |
|
text+=' ' |
|
labels = pyopenjtalk.extract_fullcontext(sentence) |
|
for n, label in enumerate(labels): |
|
phoneme = re.search(r'\-([^\+]*)\+', label).group(1) |
|
if phoneme not in ['sil','pau']: |
|
text += phoneme.replace('ch','สง').replace('sh','ส').replace('cl','Q') |
|
else: |
|
continue |
|
n_moras = int(re.search(r'/F:(\d+)_', label).group(1)) |
|
a1 = int(re.search(r"/A:(\-?[0-9]+)\+", label).group(1)) |
|
a2 = int(re.search(r"\+(\d+)\+", label).group(1)) |
|
a3 = int(re.search(r"\+(\d+)/", label).group(1)) |
|
if re.search(r'\-([^\+]*)\+', labels[n + 1]).group(1) in ['sil','pau']: |
|
a2_next=-1 |
|
else: |
|
a2_next = int(re.search(r"\+(\d+)\+", labels[n + 1]).group(1)) |
|
|
|
if a3 == 1 and a2_next == 1: |
|
text += ' ' |
|
|
|
elif a1 == 0 and a2_next == a2 + 1 and a2 != n_moras: |
|
text += 'โ' |
|
|
|
elif a2 == 1 and a2_next == 2: |
|
text += 'โ' |
|
if i<len(marks): |
|
text += unidecode(marks[i]).replace(' ','') |
|
if re.match('[A-Za-z]',text[-1]): |
|
text += '.' |
|
return text |
|
|
|
|
|
def japanese_cleaners2(text): |
|
return japanese_cleaners(text).replace('ts','สฆ').replace('...','โฆ') |
|
|
|
|
|
def korean_cleaners(text): |
|
'''Pipeline for Korean text''' |
|
text = latin_to_hangul(text) |
|
text = number_to_hangul(text) |
|
text = j2hcj(h2j(text)) |
|
text = divide_hangul(text) |
|
if re.match('[\u3131-\u3163]',text[-1]): |
|
text += '.' |
|
return text |
|
|
|
|
|
def chinese_cleaners(text): |
|
'''Pipeline for Chinese text''' |
|
text=text.replace('ใ','๏ผ').replace('๏ผ','๏ผ').replace('๏ผ','๏ผ') |
|
words=jieba.lcut(text,cut_all=False) |
|
text='' |
|
for word in words: |
|
bopomofos=lazy_pinyin(word,BOPOMOFO) |
|
if not re.search('[\u4e00-\u9fff]',word): |
|
text+=word |
|
continue |
|
for i in range(len(bopomofos)): |
|
if re.match('[\u3105-\u3129]',bopomofos[i][-1]): |
|
bopomofos[i]+='ห' |
|
if text!='': |
|
text+=' ' |
|
text+=''.join(bopomofos) |
|
if re.match('[หหหหห]',text[-1]): |
|
text += 'ใ' |
|
return text |
|
|