Upload 26 files

text/cleaners.py

@@ -1,11 +1,19 @@
 import re
+from text.english import english_to_lazy_ipa, english_to_ipa2, english_to_lazy_ipa2
 from text.japanese import japanese_to_romaji_with_accent, japanese_to_ipa, japanese_to_ipa2, japanese_to_ipa3
 from text.mandarin import number_to_chinese, chinese_to_bopomofo, latin_to_bopomofo, chinese_to_romaji, chinese_to_lazy_ipa, chinese_to_ipa, chinese_to_ipa2
-from text.english import english_to_lazy_ipa, english_to_ipa2, english_to_lazy_ipa2
+# from text.sanskrit import devanagari_to_ipa
+# from text.english import english_to_lazy_ipa, english_to_ipa2, english_to_lazy_ipa2
+# from text.thai import num_to_thai, latin_to_thai
+# from text.shanghainese import shanghainese_to_ipa
+# from text.cantonese import cantonese_to_ipa
+# from text.ngu_dialect import ngu_dialect_to_ipa
+
 
 def japanese_cleaners(text):
     text = japanese_to_romaji_with_accent(text)
-    text = re.sub(r'([A-Za-z])$', r'\1.', text)
+    if re.match('[A-Za-z]', text[-1]):
+        text += '.'
     return text
 
 
@@ -18,7 +26,8 @@ def korean_cleaners(text):
     text = latin_to_hangul(text)
     text = number_to_hangul(text)
     text = divide_hangul(text)
-    text = re.sub(r'([\u3131-\u3163])$', r'\1.', text)
+    if re.match('[\u3131-\u3163]', text[-1]):
+        text += '.'
     return text
 
 
@@ -27,39 +36,58 @@ def chinese_cleaners(text):
     text = number_to_chinese(text)
     text = chinese_to_bopomofo(text)
     text = latin_to_bopomofo(text)
-    text = re.sub(r'([ˉˊˇˋ˙])$', r'\1。', text)
+    if re.match('[ˉˊˇˋ˙]', text[-1]):
+        text += '。'
     return text
 
 
 def zh_ja_mixture_cleaners(text):
-    text = re.sub(r'\[ZH\](.*?)\[ZH\]',
-                  lambda x: chinese_to_romaji(x.group(1))+' ', text)
-    text = re.sub(r'\[JA\](.*?)\[JA\]', lambda x: japanese_to_romaji_with_accent(
-        x.group(1)).replace('ts', 'ʦ').replace('u', 'ɯ').replace('...', '…')+' ', text)
-    text = re.sub(r'\s+$', '', text)
-    text = re.sub(r'([^\.,!\?\-…~])$', r'\1.', text)
+    chinese_texts = re.findall(r'\[ZH\].*?\[ZH\]', text)
+    japanese_texts = re.findall(r'\[JA\].*?\[JA\]', text)
+    for chinese_text in chinese_texts:
+        cleaned_text = chinese_to_romaji(chinese_text[4:-4])
+        text = text.replace(chinese_text, cleaned_text+' ', 1)
+    for japanese_text in japanese_texts:
+        cleaned_text = japanese_to_romaji_with_accent(
+            japanese_text[4:-4]).replace('ts', 'ʦ').replace('u', 'ɯ').replace('...', '…')
+        text = text.replace(japanese_text, cleaned_text+' ', 1)
+    text = text[:-1]
+    if re.match('[A-Za-zɯɹəɥ→↓↑]', text[-1]):
+        text += '.'
     return text
 
 
 def sanskrit_cleaners(text):
     text = text.replace('॥', '।').replace('ॐ', 'ओम्')
-    text = re.sub(r'([^।])$', r'\1।', text)
+    if text[-1] != '।':
+        text += ' ।'
     return text
 
 
 def cjks_cleaners(text):
-    text = re.sub(r'\[ZH\](.*?)\[ZH\]',
-                  lambda x: chinese_to_lazy_ipa(x.group(1))+' ', text)
-    text = re.sub(r'\[JA\](.*?)\[JA\]',
-                  lambda x: japanese_to_ipa(x.group(1))+' ', text)
-    text = re.sub(r'\[KO\](.*?)\[KO\]',
-                  lambda x: korean_to_lazy_ipa(x.group(1))+' ', text)
-    text = re.sub(r'\[SA\](.*?)\[SA\]',
-                  lambda x: devanagari_to_ipa(x.group(1))+' ', text)
-    text = re.sub(r'\[EN\](.*?)\[EN\]',
-                  lambda x: english_to_lazy_ipa(x.group(1))+' ', text)
-    text = re.sub(r'\s+$', '', text)
-    text = re.sub(r'([^\.,!\?\-…~])$', r'\1.', text)
+    chinese_texts = re.findall(r'\[ZH\].*?\[ZH\]', text)
+    japanese_texts = re.findall(r'\[JA\].*?\[JA\]', text)
+    korean_texts = re.findall(r'\[KO\].*?\[KO\]', text)
+    sanskrit_texts = re.findall(r'\[SA\].*?\[SA\]', text)
+    english_texts = re.findall(r'\[EN\].*?\[EN\]', text)
+    for chinese_text in chinese_texts:
+        cleaned_text = chinese_to_lazy_ipa(chinese_text[4:-4])
+        text = text.replace(chinese_text, cleaned_text+' ', 1)
+    for japanese_text in japanese_texts:
+        cleaned_text = japanese_to_ipa(japanese_text[4:-4])
+        text = text.replace(japanese_text, cleaned_text+' ', 1)
+    for korean_text in korean_texts:
+        cleaned_text = korean_to_lazy_ipa(korean_text[4:-4])
+        text = text.replace(korean_text, cleaned_text+' ', 1)
+    for sanskrit_text in sanskrit_texts:
+        cleaned_text = devanagari_to_ipa(sanskrit_text[4:-4])
+        text = text.replace(sanskrit_text, cleaned_text+' ', 1)
+    for english_text in english_texts:
+        cleaned_text = english_to_lazy_ipa(english_text[4:-4])
+        text = text.replace(english_text, cleaned_text+' ', 1)
+    text = text[:-1]
+    if re.match(r'[^\.,!\?\-…~]', text[-1]):
+        text += '.'
     return text
 
 
@@ -91,3 +119,58 @@ def cjke_cleaners(text):
         text += '.'
     return text
 
+
+def cjke_cleaners2(text):
+    chinese_texts = re.findall(r'\[ZH\].*?\[ZH\]', text)
+    japanese_texts = re.findall(r'\[JA\].*?\[JA\]', text)
+    korean_texts = re.findall(r'\[KO\].*?\[KO\]', text)
+    english_texts = re.findall(r'\[EN\].*?\[EN\]', text)
+    for chinese_text in chinese_texts:
+        cleaned_text = chinese_to_ipa(chinese_text[4:-4])
+        text = text.replace(chinese_text, cleaned_text+' ', 1)
+    for japanese_text in japanese_texts:
+        cleaned_text = japanese_to_ipa2(japanese_text[4:-4])
+        text = text.replace(japanese_text, cleaned_text+' ', 1)
+    for korean_text in korean_texts:
+        cleaned_text = korean_to_ipa(korean_text[4:-4])
+        text = text.replace(korean_text, cleaned_text+' ', 1)
+    for english_text in english_texts:
+        cleaned_text = english_to_ipa2(english_text[4:-4])
+        text = text.replace(english_text, cleaned_text+' ', 1)
+    text = text[:-1]
+    if re.match(r'[^\.,!\?\-…~]', text[-1]):
+        text += '.'
+    return text
+
+
+def thai_cleaners(text):
+    text = num_to_thai(text)
+    text = latin_to_thai(text)
+    return text
+
+
+def shanghainese_cleaners(text):
+    text = shanghainese_to_ipa(text)
+    if re.match(r'[^\.,!\?\-…~]', text[-1]):
+        text += '.'
+    return text
+
+
+def chinese_dialect_cleaners(text):
+    text = re.sub(r'\[MD\](.*?)\[MD\]',
+                  lambda x: chinese_to_ipa2(x.group(1))+' ', text)
+    text = re.sub(r'\[TW\](.*?)\[TW\]',
+                  lambda x: chinese_to_ipa2(x.group(1), True)+' ', text)
+    text = re.sub(r'\[JA\](.*?)\[JA\]',
+                  lambda x: japanese_to_ipa3(x.group(1)).replace('Q', 'ʔ')+' ', text)
+    text = re.sub(r'\[SH\](.*?)\[SH\]', lambda x: shanghainese_to_ipa(x.group(1)).replace('1', '˥˧').replace('5',
+                  '˧˧˦').replace('6', '˩˩˧').replace('7', '˥').replace('8', '˩˨').replace('ᴀ', 'ɐ').replace('ᴇ', 'e')+' ', text)
+    text = re.sub(r'\[GD\](.*?)\[GD\]',
+                  lambda x: cantonese_to_ipa(x.group(1))+' ', text)
+    text = re.sub(r'\[EN\](.*?)\[EN\]',
+                  lambda x: english_to_lazy_ipa2(x.group(1))+' ', text)
+    text = re.sub(r'\[([A-Z]{2})\](.*?)\[\1\]', lambda x: ngu_dialect_to_ipa(x.group(2), x.group(
+        1)).replace('ʣ', 'dz').replace('ʥ', 'dʑ').replace('ʦ', 'ts').replace('ʨ', 'tɕ')+' ', text)
+    text = re.sub(r'\s+$', '', text)
+    text = re.sub(r'([^\.,!\?\-…~])$', r'\1.', text)
+    return text

text/mandarin.py

@@ -4,7 +4,6 @@ import re
 from pypinyin import lazy_pinyin, BOPOMOFO
 import jieba
 import cn2an
-import logging
 
 
 # List of (Latin alphabet, bopomofo) pairs:
@@ -240,7 +239,7 @@ def number_to_chinese(text):
     return text
 
 
-def chinese_to_bopomofo(text):
+def chinese_to_bopomofo(text, taiwanese=False):
     text = text.replace('、', '，').replace('；', '，').replace('：', '，')
     words = jieba.lcut(text, cut_all=False)
     text = ''
@@ -253,7 +252,10 @@ def chinese_to_bopomofo(text):
             bopomofos[i] = re.sub(r'([\u3105-\u3129])$', r'\1ˉ', bopomofos[i])
         if text != '':
             text += ' '
-        text += ''.join(bopomofos)
+        if taiwanese:
+            text += '#'+'#'.join(bopomofos)
+        else:
+            text += ''.join(bopomofos)
     return text
 
 
@@ -314,9 +316,9 @@ def chinese_to_ipa(text):
     return text
 
 
-def chinese_to_ipa2(text):
+def chinese_to_ipa2(text, taiwanese=False):
     text = number_to_chinese(text)
-    text = chinese_to_bopomofo(text)
+    text = chinese_to_bopomofo(text, taiwanese)
     text = latin_to_bopomofo(text)
     text = bopomofo_to_ipa2(text)
     text = re.sub(r'i([aoe])', r'j\1', text)

text/symbols.py

@@ -1,18 +1,15 @@
 '''
 Defines the set of symbols used in text input to the model.
 '''
-
-'''# japanese_cleaners
 _pad        = '_'
-_punctuation = ',.!?-'
-_letters = 'AEINOQUabdefghijkmnoprstuvwyzʃʧ↓↑ '
+_punctuation = ',.!?-~…'
+_letters = 'AEINOQUabdefghijklmnoprstuvwyzʃʧʦɯɹəɥ⁼ʰ`→↓↑ '
 '''
-
 # japanese_cleaners2
 _pad        = '_'
 _punctuation = ',.!?-~…'
 _letters = 'AEINOQUabdefghijkmnoprstuvwyzʃʧʦ↓↑ '
-
+'''
 
 '''# korean_cleaners
 _pad        = '_'
@@ -26,11 +23,6 @@ _punctuation = '，。！？—…'
 _letters = 'ㄅㄆㄇㄈㄉㄊㄋㄌㄍㄎㄏㄐㄑㄒㄓㄔㄕㄖㄗㄘㄙㄚㄛㄜㄝㄞㄟㄠㄡㄢㄣㄤㄥㄦㄧㄨㄩˉˊˇˋ˙ '
 '''
 
-'''# zh_ja_mixture_cleaners
-_pad        = '_'
-_punctuation = ',.!?-~…'
-_letters = 'AEINOQUabdefghijklmnoprstuvwyzʃʧʦɯɹəɥ⁼ʰ`→↓↑ '
-'''
 
 '''# sanskrit_cleaners
 _pad        = '_'
@@ -65,7 +57,7 @@ _letters = 'abdfghiklmnopstuvyzøŋȵɑɔɕəɤɦɪɿʑʔʰ̩̃ᴀᴇ15678 '
 '''# chinese_dialect_cleaners
 _pad        = '_'
 _punctuation = ',.!?~…─'
-_letters = '#Nabdefghijklmnoprstuvwxyzæçøŋœȵɐɑɒɓɔɕɗɘəɚɛɜɣɤɦɪɭɯɵɷɸɻɾɿʂʅʊʋʌʏʑʔʦʮʰʷˀː˥˦˧˨˩̥̩̃̚ᴀᴇ↑↓∅ⱼ '
+_letters = '#Nabdefghijklmnoprstuvwxyzæçøŋœȵɐɑɒɓɔɕɗɘəɚɛɜɣɤɦɪɭɯɵɷɸɻɾɿʂʅʊʋʌʏʑʔʦʮʰʷˀː˥˦˧˨˩̥̩̃̚αᴀᴇ↑↓∅ⱼ '
 '''
 
 # Export all symbols: