File size: 6,344 Bytes
a91158d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
""" 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 pyopenjtalk
from janome.tokenizer import Tokenizer


# Regular expression matching whitespace:
_whitespace_re = re.compile(r'\s+')

# List of (regular expression, replacement) pairs for abbreviations:
_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'),
]]

# Regular expression matching Japanese without punctuation marks:
_japanese_characters = re.compile(r'[A-Za-z\d\u3005\u3040-\u30ff\u4e00-\u9fff\uff11-\uff19\uff21-\uff3a\uff41-\uff5a\uff66-\uff9d]')

# Regular expression matching non-Japanese characters or punctuation marks:
_japanese_marks = re.compile(r'[^A-Za-z\d\u3005\u3040-\u30ff\u4e00-\u9fff\uff11-\uff19\uff21-\uff3a\uff41-\uff5a\uff66-\uff9d]')


# Tokenizer for Japanese
tokenizer = Tokenizer()


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 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 Japanese text.'''
  sentences = re.split(_japanese_marks, text)
  marks = re.findall(_japanese_marks, text)
  text = ''
  for i, mark in enumerate(marks):
    if re.match(_japanese_characters, sentences[i]):
      text += pyopenjtalk.g2p(sentences[i], kana=False).replace('pau','').replace(' ','')
    text += unidecode(mark).replace(' ','')
  if re.match(_japanese_characters, sentences[-1]):
      text += pyopenjtalk.g2p(sentences[-1], kana=False).replace('pau','').replace(' ','')
  if re.match('[A-Za-z]',text[-1]):
    text += '.'
  return text


def japanese_tokenization_cleaners(text):
  '''Pipeline for tokenizing Japanese text.'''
  words = []
  for token in tokenizer.tokenize(text):
    if token.phonetic!='*':
      words.append(token.phonetic)
    else:
      words.append(token.surface)
  text = ''
  for word in words:
    if re.match(_japanese_characters, word):
      if word[0] == '\u30fc':
        continue
      if len(text)>0:
        text += ' '
      text += pyopenjtalk.g2p(word, kana=False).replace(' ','')
    else:
      text += unidecode(word).replace(' ','')
  if re.match('[A-Za-z]',text[-1]):
    text += '.'
  return text


def japanese_accent_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):
      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
        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))
        # Accent phrase boundary
        if a3 == 1 and a2_next == 1:
          text += ' '
        # Falling
        elif a1 == 0 and a2_next == a2 + 1 and a2 != n_moras:
          text += ')'
        # Rising
        elif a2 == 1 and a2_next == 2:
          text += '('
    if i<len(marks):
      text += unidecode(marks[i]).replace(' ','')
  if re.match('[A-Za-z]',text[-1]):
    text += '.'
  return text


def japanese_phrase_cleaners(text):
  '''Pipeline for dividing Japanese text into phrases.'''
  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
        else:
          continue
        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))
        # Accent phrase boundary
        if a3 == 1 and a2_next == 1:
          text += ' '
    if i<len(marks):
      text += unidecode(marks[i]).replace(' ','')
  if re.match('[A-Za-z]',text[-1]):
    text += '.'
  return text