Refactor: Format Code with Ruff and Update Deprecated G2PW Link (#2255)

* ruff check --fix

* ruff format --line-length 120 --target-version py39

* Change the link for G2PW Model

* update pytorch version and colab

GPT_SoVITS/text/LangSegmenter/__init__.py

@@ -1 +1 @@
-from .langsegmenter import LangSegmenter
+from .langsegmenter import LangSegmenter

GPT_SoVITS/text/LangSegmenter/langsegmenter.py

@@ -3,38 +3,44 @@ import re
 
 # jieba静音
 import jieba
+
 jieba.setLogLevel(logging.CRITICAL)
 
 # 更改fast_langdetect大模型位置
 from pathlib import Path
 import fast_langdetect
-fast_langdetect.infer._default_detector = fast_langdetect.infer.LangDetector(fast_langdetect.infer.LangDetectConfig(cache_dir=Path(__file__).parent.parent.parent / "pretrained_models" / "fast_langdetect"))
+
+fast_langdetect.infer._default_detector = fast_langdetect.infer.LangDetector(
+    fast_langdetect.infer.LangDetectConfig(
+        cache_dir=Path(__file__).parent.parent.parent / "pretrained_models" / "fast_langdetect"
+    )
+)
 
 
 from split_lang import LangSplitter
 
 
 def full_en(text):
-    pattern = r'^[A-Za-z0-9\s\u0020-\u007E\u2000-\u206F\u3000-\u303F\uFF00-\uFFEF]+$'
+    pattern = r"^[A-Za-z0-9\s\u0020-\u007E\u2000-\u206F\u3000-\u303F\uFF00-\uFFEF]+$"
     return bool(re.match(pattern, text))
 
 
 def full_cjk(text):
     # 来自wiki
     cjk_ranges = [
-        (0x4E00, 0x9FFF),        # CJK Unified Ideographs
-        (0x3400, 0x4DB5),        # CJK Extension A
-        (0x20000, 0x2A6DD),      # CJK Extension B
-        (0x2A700, 0x2B73F),      # CJK Extension C
-        (0x2B740, 0x2B81F),      # CJK Extension D
-        (0x2B820, 0x2CEAF),      # CJK Extension E
-        (0x2CEB0, 0x2EBEF),      # CJK Extension F
-        (0x30000, 0x3134A),      # CJK Extension G
-        (0x31350, 0x323AF),      # CJK Extension H
-        (0x2EBF0, 0x2EE5D),      # CJK Extension H
+        (0x4E00, 0x9FFF),  # CJK Unified Ideographs
+        (0x3400, 0x4DB5),  # CJK Extension A
+        (0x20000, 0x2A6DD),  # CJK Extension B
+        (0x2A700, 0x2B73F),  # CJK Extension C
+        (0x2B740, 0x2B81F),  # CJK Extension D
+        (0x2B820, 0x2CEAF),  # CJK Extension E
+        (0x2CEB0, 0x2EBEF),  # CJK Extension F
+        (0x30000, 0x3134A),  # CJK Extension G
+        (0x31350, 0x323AF),  # CJK Extension H
+        (0x2EBF0, 0x2EE5D),  # CJK Extension H
     ]
 
-    pattern = r'[0-9、-〜。！？.!?… ]+$'
+    pattern = r"[0-9、-〜。！？.!?… ]+$"
 
     cjk_text = ""
     for char in text:
@@ -45,7 +51,7 @@ def full_cjk(text):
     return cjk_text
 
 
-def split_jako(tag_lang,item):
+def split_jako(tag_lang, item):
     if tag_lang == "ja":
         pattern = r"([\u3041-\u3096\u3099\u309A\u30A1-\u30FA\u30FC]+(?:[0-9、-〜。！？.!?… ]+[\u3041-\u3096\u3099\u309A\u30A1-\u30FA\u30FC]*)*)"
     else:
@@ -53,41 +59,40 @@ def split_jako(tag_lang,item):
 
     lang_list: list[dict] = []
     tag = 0
-    for match in re.finditer(pattern, item['text']):
+    for match in re.finditer(pattern, item["text"]):
         if match.start() > tag:
-            lang_list.append({'lang':item['lang'],'text':item['text'][tag:match.start()]})
+            lang_list.append({"lang": item["lang"], "text": item["text"][tag : match.start()]})
 
         tag = match.end()
-        lang_list.append({'lang':tag_lang,'text':item['text'][match.start():match.end()]})
+        lang_list.append({"lang": tag_lang, "text": item["text"][match.start() : match.end()]})
 
-    if tag < len(item['text']):
-        lang_list.append({'lang':item['lang'],'text':item['text'][tag:len(item['text'])]})
+    if tag < len(item["text"]):
+        lang_list.append({"lang": item["lang"], "text": item["text"][tag : len(item["text"])]})
 
     return lang_list
 
 
 def merge_lang(lang_list, item):
-    if lang_list and item['lang'] == lang_list[-1]['lang']:
-        lang_list[-1]['text'] += item['text']
+    if lang_list and item["lang"] == lang_list[-1]["lang"]:
+        lang_list[-1]["text"] += item["text"]
     else:
         lang_list.append(item)
     return lang_list
 
 
-class LangSegmenter():
+class LangSegmenter:
     # 默认过滤器, 基于gsv目前四种语言
     DEFAULT_LANG_MAP = {
         "zh": "zh",
         "yue": "zh",  # 粤语
         "wuu": "zh",  # 吴语
         "zh-cn": "zh",
-        "zh-tw": "x", # 繁体设置为x
+        "zh-tw": "x",  # 繁体设置为x
         "ko": "ko",
         "ja": "ja",
         "en": "en",
     }
 
-    
     def getTexts(text):
         lang_splitter = LangSplitter(lang_map=LangSegmenter.DEFAULT_LANG_MAP)
         substr = lang_splitter.split_by_lang(text=text)
@@ -95,18 +100,18 @@ class LangSegmenter():
         lang_list: list[dict] = []
 
         for _, item in enumerate(substr):
-            dict_item = {'lang':item.lang,'text':item.text}
+            dict_item = {"lang": item.lang, "text": item.text}
 
             # 处理短英文被识别为其他语言的问题
-            if full_en(dict_item['text']):  
-                dict_item['lang'] = 'en'
-                lang_list = merge_lang(lang_list,dict_item)
+            if full_en(dict_item["text"]):
+                dict_item["lang"] = "en"
+                lang_list = merge_lang(lang_list, dict_item)
                 continue
 
             # 处理非日语夹日文的问题(不包含CJK)
             ja_list: list[dict] = []
-            if dict_item['lang'] != 'ja':
-                ja_list = split_jako('ja',dict_item)
+            if dict_item["lang"] != "ja":
+                ja_list = split_jako("ja", dict_item)
 
             if not ja_list:
                 ja_list.append(dict_item)
@@ -115,8 +120,8 @@ class LangSegmenter():
             ko_list: list[dict] = []
             temp_list: list[dict] = []
             for _, ko_item in enumerate(ja_list):
-                if ko_item["lang"] != 'ko':
-                    ko_list = split_jako('ko',ko_item)
+                if ko_item["lang"] != "ko":
+                    ko_list = split_jako("ko", ko_item)
 
                 if ko_list:
                     temp_list.extend(ko_list)
@@ -126,28 +131,28 @@ class LangSegmenter():
             # 未存在非日韩文夹日韩文
             if len(temp_list) == 1:
                 # 未知语言检查是否为CJK
-                if dict_item['lang'] == 'x':
-                    cjk_text = full_cjk(dict_item['text'])
+                if dict_item["lang"] == "x":
+                    cjk_text = full_cjk(dict_item["text"])
                     if cjk_text:
-                        dict_item = {'lang':'zh','text':cjk_text}
-                        lang_list = merge_lang(lang_list,dict_item)
+                        dict_item = {"lang": "zh", "text": cjk_text}
+                        lang_list = merge_lang(lang_list, dict_item)
                     continue
                 else:
-                    lang_list = merge_lang(lang_list,dict_item)
+                    lang_list = merge_lang(lang_list, dict_item)
                     continue
 
             # 存在非日韩文夹日韩文
             for _, temp_item in enumerate(temp_list):
                 # 未知语言检查是否为CJK
-                if temp_item['lang'] == 'x':
-                    cjk_text = full_cjk(dict_item['text'])
+                if temp_item["lang"] == "x":
+                    cjk_text = full_cjk(dict_item["text"])
                     if cjk_text:
-                        dict_item = {'lang':'zh','text':cjk_text}
-                        lang_list = merge_lang(lang_list,dict_item)
+                        dict_item = {"lang": "zh", "text": cjk_text}
+                        lang_list = merge_lang(lang_list, dict_item)
                 else:
-                    lang_list = merge_lang(lang_list,temp_item)
+                    lang_list = merge_lang(lang_list, temp_item)
         return lang_list
-    
+
 
 if __name__ == "__main__":
     text = "MyGO?,你也喜欢まいご吗？"
@@ -155,4 +160,3 @@ if __name__ == "__main__":
 
     text = "ねえ、知ってる？最近、僕は天文学を勉強してるんだ。君の瞳が星空みたいにキラキラしてるからさ。"
     print(LangSegmenter.getTexts(text))
-

GPT_SoVITS/text/__init__.py

@@ -10,18 +10,19 @@ from text import symbols2 as symbols_v2
 _symbol_to_id_v1 = {s: i for i, s in enumerate(symbols_v1.symbols)}
 _symbol_to_id_v2 = {s: i for i, s in enumerate(symbols_v2.symbols)}
 
+
 def cleaned_text_to_sequence(cleaned_text, version=None):
-  '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
+    """Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
     Args:
       text: string to convert to a sequence
     Returns:
       List of integers corresponding to the symbols in the text
-  '''
-  if version is None:version=os.environ.get('version', 'v2')
-  if version == "v1":
-    phones = [_symbol_to_id_v1[symbol] for symbol in cleaned_text]
-  else:
-    phones = [_symbol_to_id_v2[symbol] for symbol in cleaned_text]
-
-  return phones
+    """
+    if version is None:
+        version = os.environ.get("version", "v2")
+    if version == "v1":
+        phones = [_symbol_to_id_v1[symbol] for symbol in cleaned_text]
+    else:
+        phones = [_symbol_to_id_v2[symbol] for symbol in cleaned_text]
 
+    return phones

GPT_SoVITS/text/cantonese.py

@@ -1,6 +1,5 @@
 # reference: https://huggingface.co/spaces/Naozumi0512/Bert-VITS2-Cantonese-Yue/blob/main/text/chinese.py
 
-import sys
 import re
 import cn2an
 import ToJyutping
@@ -99,9 +98,7 @@ def replace_punctuation(text):
 
     replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)
 
-    replaced_text = re.sub(
-        r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text
-    )
+    replaced_text = re.sub(r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text)
 
     return replaced_text
 
@@ -115,7 +112,9 @@ def text_normalize(text):
     return dest_text
 
 
-punctuation_set=set(punctuation)
+punctuation_set = set(punctuation)
+
+
 def jyuping_to_initials_finals_tones(jyuping_syllables):
     initials_finals = []
     tones = []
@@ -160,12 +159,14 @@ def jyuping_to_initials_finals_tones(jyuping_syllables):
     assert len(initials_finals) == len(tones)
 
     ###魔改为辅音+带音调的元音
-    phones=[]
-    for a,b in zip(initials_finals,tones):
-        if(b not in [-1,0]):###防止粤语和普通话重合开头加Y，如果是标点，不加。
-            todo="%s%s"%(a,b)
-        else:todo=a
-        if(todo not in punctuation_set):todo="Y%s"%todo
+    phones = []
+    for a, b in zip(initials_finals, tones):
+        if b not in [-1, 0]:  ###防止粤语和普通话重合开头加Y，如果是标点，不加。
+            todo = "%s%s" % (a, b)
+        else:
+            todo = a
+        if todo not in punctuation_set:
+            todo = "Y%s" % todo
         phones.append(todo)
 
     # return initials_finals, tones, word2ph
@@ -218,4 +219,4 @@ if __name__ == "__main__":
     # phones, tones, word2ph = g2p(text)
     phones, word2ph = g2p(text)
     # print(phones, tones, word2ph)
-    print(phones, word2ph)
+    print(phones, word2ph)

GPT_SoVITS/text/chinese.py

@@ -1,5 +1,4 @@
 import os
-import pdb
 import re
 
 import cn2an
@@ -17,7 +16,9 @@ pinyin_to_symbol_map = {
     for line in open(os.path.join(current_file_path, "opencpop-strict.txt")).readlines()
 }
 
-import jieba_fast, logging
+import jieba_fast
+import logging
+
 jieba_fast.setLogLevel(logging.CRITICAL)
 import jieba_fast.posseg as psg
 
@@ -37,7 +38,7 @@ rep_map = {
     "/": ",",
     "—": "-",
     "~": "…",
-    "～":"…",
+    "～": "…",
 }
 
 tone_modifier = ToneSandhi()
@@ -49,9 +50,7 @@ def replace_punctuation(text):
 
     replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)
 
-    replaced_text = re.sub(
-        r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text
-    )
+    replaced_text = re.sub(r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text)
 
     return replaced_text
 
@@ -62,17 +61,15 @@ def replace_punctuation_with_en(text):
 
     replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)
 
-    replaced_text = re.sub(
-        r"[^\u4e00-\u9fa5A-Za-z" + "".join(punctuation) + r"]+", "", replaced_text
-    )
+    replaced_text = re.sub(r"[^\u4e00-\u9fa5A-Za-z" + "".join(punctuation) + r"]+", "", replaced_text)
 
     return replaced_text
 
 
 def replace_consecutive_punctuation(text):
-    punctuations = ''.join(re.escape(p) for p in punctuation)
-    pattern = f'([{punctuations}])([{punctuations}])+'
-    result = re.sub(pattern, r'\1', text)
+    punctuations = "".join(re.escape(p) for p in punctuation)
+    pattern = f"([{punctuations}])([{punctuations}])+"
+    result = re.sub(pattern, r"\1", text)
     return result
 
 
@@ -87,9 +84,7 @@ def _get_initials_finals(word):
     initials = []
     finals = []
     orig_initials = lazy_pinyin(word, neutral_tone_with_five=True, style=Style.INITIALS)
-    orig_finals = lazy_pinyin(
-        word, neutral_tone_with_five=True, style=Style.FINALS_TONE3
-    )
+    orig_finals = lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3)
     for c, v in zip(orig_initials, orig_finals):
         initials.append(c)
         finals.append(v)

GPT_SoVITS/text/chinese2.py

@@ -1,10 +1,9 @@
 import os
-import pdb
 import re
 
 import cn2an
 from pypinyin import lazy_pinyin, Style
-from pypinyin.contrib.tone_convert import to_normal, to_finals_tone3, to_initials, to_finals
+from pypinyin.contrib.tone_convert import to_finals_tone3, to_initials
 
 from text.symbols import punctuation
 from text.tone_sandhi import ToneSandhi
@@ -18,18 +17,26 @@ pinyin_to_symbol_map = {
     for line in open(os.path.join(current_file_path, "opencpop-strict.txt")).readlines()
 }
 
-import jieba_fast, logging
+import jieba_fast
+import logging
+
 jieba_fast.setLogLevel(logging.CRITICAL)
 import jieba_fast.posseg as psg
 
 # is_g2pw_str = os.environ.get("is_g2pw", "True")##默认开启
 # is_g2pw = False#True if is_g2pw_str.lower() == 'true' else False
-is_g2pw = True#True if is_g2pw_str.lower() == 'true' else False
+is_g2pw = True  # True if is_g2pw_str.lower() == 'true' else False
 if is_g2pw:
     # print("当前使用g2pw进行拼音推理")
     from text.g2pw import G2PWPinyin, correct_pronunciation
+
     parent_directory = os.path.dirname(current_file_path)
-    g2pw = G2PWPinyin(model_dir="GPT_SoVITS/text/G2PWModel",model_source=os.environ.get("bert_path","GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large"),v_to_u=False, neutral_tone_with_five=True)
+    g2pw = G2PWPinyin(
+        model_dir="GPT_SoVITS/text/G2PWModel",
+        model_source=os.environ.get("bert_path", "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large"),
+        v_to_u=False,
+        neutral_tone_with_five=True,
+    )
 
 rep_map = {
     "：": ",",
@@ -46,7 +53,7 @@ rep_map = {
     "/": ",",
     "—": "-",
     "~": "…",
-    "～":"…",
+    "～": "…",
 }
 
 tone_modifier = ToneSandhi()
@@ -58,9 +65,7 @@ def replace_punctuation(text):
 
     replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)
 
-    replaced_text = re.sub(
-        r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text
-    )
+    replaced_text = re.sub(r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text)
 
     return replaced_text
 
@@ -77,9 +82,7 @@ def _get_initials_finals(word):
     finals = []
 
     orig_initials = lazy_pinyin(word, neutral_tone_with_five=True, style=Style.INITIALS)
-    orig_finals = lazy_pinyin(
-        word, neutral_tone_with_five=True, style=Style.FINALS_TONE3
-    )
+    orig_finals = lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3)
 
     for c, v in zip(orig_initials, orig_finals):
         initials.append(c)
@@ -87,31 +90,66 @@ def _get_initials_finals(word):
     return initials, finals
 
 
-must_erhua = {
-    "小院儿", "胡同儿", "范儿", "老汉儿", "撒欢儿", "寻老礼儿", "妥妥儿", "媳妇儿"
-}
+must_erhua = {"小院儿", "胡同儿", "范儿", "老汉儿", "撒欢儿", "寻老礼儿", "妥妥儿", "媳妇儿"}
 not_erhua = {
-    "虐儿", "为儿", "护儿", "瞒儿", "救儿", "替儿", "有儿", "一儿", "我儿", "俺儿", "妻儿",
-    "拐儿", "聋儿", "乞儿", "患儿", "幼儿", "孤儿", "婴儿", "婴幼儿", "连体儿", "脑瘫儿",
-    "流浪儿", "体弱儿", "混血儿", "蜜雪儿", "舫儿", "祖儿", "美儿", "应采儿", "可儿", "侄儿",
-    "孙儿", "侄孙儿", "女儿", "男儿", "红孩儿", "花儿", "虫儿", "马儿", "鸟儿", "猪儿", "猫儿",
-    "狗儿", "少儿"
+    "虐儿",
+    "为儿",
+    "护儿",
+    "瞒儿",
+    "救儿",
+    "替儿",
+    "有儿",
+    "一儿",
+    "我儿",
+    "俺儿",
+    "妻儿",
+    "拐儿",
+    "聋儿",
+    "乞儿",
+    "患儿",
+    "幼儿",
+    "孤儿",
+    "婴儿",
+    "婴幼儿",
+    "连体儿",
+    "脑瘫儿",
+    "流浪儿",
+    "体弱儿",
+    "混血儿",
+    "蜜雪儿",
+    "舫儿",
+    "祖儿",
+    "美儿",
+    "应采儿",
+    "可儿",
+    "侄儿",
+    "孙儿",
+    "侄孙儿",
+    "女儿",
+    "男儿",
+    "红孩儿",
+    "花儿",
+    "虫儿",
+    "马儿",
+    "鸟儿",
+    "猪儿",
+    "猫儿",
+    "狗儿",
+    "少儿",
 }
-def _merge_erhua(initials: list[str],
-                finals: list[str],
-                word: str,
-                pos: str) -> list[list[str]]:
+
+
+def _merge_erhua(initials: list[str], finals: list[str], word: str, pos: str) -> list[list[str]]:
     """
     Do erhub.
     """
     # fix er1
     for i, phn in enumerate(finals):
-        if i == len(finals) - 1 and word[i] == "儿" and phn == 'er1':
-            finals[i] = 'er2'
+        if i == len(finals) - 1 and word[i] == "儿" and phn == "er1":
+            finals[i] = "er2"
 
     # 发音
-    if word not in must_erhua and (word in not_erhua or
-                                        pos in {"a", "j", "nr"}):
+    if word not in must_erhua and (word in not_erhua or pos in {"a", "j", "nr"}):
         return initials, finals
 
     # "……" 等情况直接返回
@@ -124,9 +162,13 @@ def _merge_erhua(initials: list[str],
     new_initials = []
     new_finals = []
     for i, phn in enumerate(finals):
-        if i == len(finals) - 1 and word[i] == "儿" and phn in {
-                "er2", "er5"
-        } and word[-2:] not in not_erhua and new_finals:
+        if (
+            i == len(finals) - 1
+            and word[i] == "儿"
+            and phn in {"er2", "er5"}
+            and word[-2:] not in not_erhua
+            and new_finals
+        ):
             phn = "er" + new_finals[-1][-1]
 
         new_initials.append(initials[i])
@@ -160,7 +202,7 @@ def _g2p(segments):
                 # assert len(sub_initials) == len(sub_finals) == len(word)
             initials = sum(initials, [])
             finals = sum(finals, [])
-            print("pypinyin结果",initials,finals)
+            print("pypinyin结果", initials, finals)
         else:
             # g2pw采用整句推理
             pinyins = g2pw.lazy_pinyin(seg, neutral_tone_with_five=True, style=Style.TONE3)
@@ -171,19 +213,19 @@ def _g2p(segments):
                 sub_finals = []
                 now_word_length = pre_word_length + len(word)
 
-                if pos == 'eng':
+                if pos == "eng":
                     pre_word_length = now_word_length
                     continue
 
                 word_pinyins = pinyins[pre_word_length:now_word_length]
 
                 # 多音字消歧
-                word_pinyins = correct_pronunciation(word,word_pinyins)
+                word_pinyins = correct_pronunciation(word, word_pinyins)
 
                 for pinyin in word_pinyins:
                     if pinyin[0].isalpha():
                         sub_initials.append(to_initials(pinyin))
-                        sub_finals.append(to_finals_tone3(pinyin,neutral_tone_with_five=True))
+                        sub_finals.append(to_finals_tone3(pinyin, neutral_tone_with_five=True))
                     else:
                         sub_initials.append(pinyin)
                         sub_finals.append(pinyin)
@@ -259,18 +301,18 @@ def replace_punctuation_with_en(text):
 
     replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)
 
-    replaced_text = re.sub(
-        r"[^\u4e00-\u9fa5A-Za-z" + "".join(punctuation) + r"]+", "", replaced_text
-    )
+    replaced_text = re.sub(r"[^\u4e00-\u9fa5A-Za-z" + "".join(punctuation) + r"]+", "", replaced_text)
 
     return replaced_text
 
+
 def replace_consecutive_punctuation(text):
-    punctuations = ''.join(re.escape(p) for p in punctuation)
-    pattern = f'([{punctuations}])([{punctuations}])+'
-    result = re.sub(pattern, r'\1', text)
+    punctuations = "".join(re.escape(p) for p in punctuation)
+    pattern = f"([{punctuations}])([{punctuations}])+"
+    result = re.sub(pattern, r"\1", text)
     return result
 
+
 def text_normalize(text):
     # https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/paddlespeech/t2s/frontend/zh_normalization
     tx = TextNormalizer()
@@ -283,6 +325,7 @@ def text_normalize(text):
     dest_text = replace_consecutive_punctuation(dest_text)
     return dest_text
 
+
 # 不排除英文的文本格式化
 def mix_text_normalize(text):
     # https://github.com/PaddlePaddle/PaddleSpeech/tree/develop/paddlespeech/t2s/frontend/zh_normalization

GPT_SoVITS/text/cleaner.py

@@ -19,55 +19,57 @@ special = [
 
 
 def clean_text(text, language, version=None):
-    if version is None:version=os.environ.get('version', 'v2')
+    if version is None:
+        version = os.environ.get("version", "v2")
     if version == "v1":
         symbols = symbols_v1.symbols
         language_module_map = {"zh": "chinese", "ja": "japanese", "en": "english"}
     else:
         symbols = symbols_v2.symbols
-        language_module_map = {"zh": "chinese2", "ja": "japanese", "en": "english", "ko": "korean","yue":"cantonese"}
+        language_module_map = {"zh": "chinese2", "ja": "japanese", "en": "english", "ko": "korean", "yue": "cantonese"}
 
-    if(language not in language_module_map):
-        language="en"
-        text=" "
+    if language not in language_module_map:
+        language = "en"
+        text = " "
     for special_s, special_l, target_symbol in special:
         if special_s in text and language == special_l:
             return clean_special(text, language, special_s, target_symbol, version)
-    language_module = __import__("text."+language_module_map[language],fromlist=[language_module_map[language]])
-    if hasattr(language_module,"text_normalize"):
+    language_module = __import__("text." + language_module_map[language], fromlist=[language_module_map[language]])
+    if hasattr(language_module, "text_normalize"):
         norm_text = language_module.text_normalize(text)
     else:
-        norm_text=text
-    if language == "zh" or language=="yue":##########
+        norm_text = text
+    if language == "zh" or language == "yue":  ##########
         phones, word2ph = language_module.g2p(norm_text)
         assert len(phones) == sum(word2ph)
         assert len(norm_text) == len(word2ph)
     elif language == "en":
         phones = language_module.g2p(norm_text)
         if len(phones) < 4:
-            phones = [','] + phones
+            phones = [","] + phones
         word2ph = None
     else:
         phones = language_module.g2p(norm_text)
         word2ph = None
-    phones = ['UNK' if ph not in symbols else ph for ph in phones]
+    phones = ["UNK" if ph not in symbols else ph for ph in phones]
     return phones, word2ph, norm_text
 
 
 def clean_special(text, language, special_s, target_symbol, version=None):
-    if version is None:version=os.environ.get('version', 'v2')
+    if version is None:
+        version = os.environ.get("version", "v2")
     if version == "v1":
         symbols = symbols_v1.symbols
         language_module_map = {"zh": "chinese", "ja": "japanese", "en": "english"}
     else:
         symbols = symbols_v2.symbols
-        language_module_map = {"zh": "chinese2", "ja": "japanese", "en": "english", "ko": "korean","yue":"cantonese"}
+        language_module_map = {"zh": "chinese2", "ja": "japanese", "en": "english", "ko": "korean", "yue": "cantonese"}
 
     """
     特殊静音段sp符号处理
     """
     text = text.replace(special_s, ",")
-    language_module = __import__("text."+language_module_map[language],fromlist=[language_module_map[language]])
+    language_module = __import__("text." + language_module_map[language], fromlist=[language_module_map[language]])
     norm_text = language_module.text_normalize(text)
     phones = language_module.g2p(norm_text)
     new_ph = []
@@ -81,8 +83,9 @@ def clean_special(text, language, special_s, target_symbol, version=None):
 
 
 def text_to_sequence(text, language, version=None):
-    version = os.environ.get('version',version)
-    if version is None:version='v2'
+    version = os.environ.get("version", version)
+    if version is None:
+        version = "v2"
     phones = clean_text(text)
     return cleaned_text_to_sequence(phones, version)
 

GPT_SoVITS/text/en_normalization/expend.py

@@ -9,17 +9,17 @@ import unicodedata
 # 后缀计量单位替换表
 measurement_map = {
     "m": ["meter", "meters"],
-    'km': ["kilometer", "kilometers"],
+    "km": ["kilometer", "kilometers"],
     "km/h": ["kilometer per hour", "kilometers per hour"],
     "ft": ["feet", "feet"],
     "L": ["liter", "liters"],
     "tbsp": ["tablespoon", "tablespoons"],
-    'tsp': ["teaspoon", "teaspoons"],
+    "tsp": ["teaspoon", "teaspoons"],
     "h": ["hour", "hours"],
     "min": ["minute", "minutes"],
     "s": ["second", "seconds"],
     "°C": ["degree celsius", "degrees celsius"],
-    "°F": ["degree fahrenheit", "degrees fahrenheit"]
+    "°F": ["degree fahrenheit", "degrees fahrenheit"],
 }
 
 
@@ -27,41 +27,42 @@ measurement_map = {
 _inflect = inflect.engine()
 
 # 转化数字序数词
-_ordinal_number_re = re.compile(r'\b([0-9]+)\. ')
+_ordinal_number_re = re.compile(r"\b([0-9]+)\. ")
 
 # 我听说好像对于数字正则识别其实用 \d 会好一点
 
-_comma_number_re = re.compile(r'([0-9][0-9\,]+[0-9])')
+_comma_number_re = re.compile(r"([0-9][0-9\,]+[0-9])")
 
 # 时间识别
-_time_re = re.compile(r'\b([01]?[0-9]|2[0-3]):([0-5][0-9])\b')
+_time_re = re.compile(r"\b([01]?[0-9]|2[0-3]):([0-5][0-9])\b")
 
 # 后缀计量单位识别
-_measurement_re = re.compile(r'\b([0-9]+(\.[0-9]+)?(m|km|km/h|ft|L|tbsp|tsp|h|min|s|°C|°F))\b')
+_measurement_re = re.compile(r"\b([0-9]+(\.[0-9]+)?(m|km|km/h|ft|L|tbsp|tsp|h|min|s|°C|°F))\b")
 
 # 前后 £ 识别 ( 写了识别两边某一边的，但是不知道为什么失败了┭┮﹏┭┮ )
-_pounds_re_start = re.compile(r'£([0-9\.\,]*[0-9]+)')
-_pounds_re_end = re.compile(r'([0-9\.\,]*[0-9]+)£')
+_pounds_re_start = re.compile(r"£([0-9\.\,]*[0-9]+)")
+_pounds_re_end = re.compile(r"([0-9\.\,]*[0-9]+)£")
 
 # 前后 $ 识别
-_dollars_re_start = re.compile(r'\$([0-9\.\,]*[0-9]+)')
-_dollars_re_end = re.compile(r'([(0-9\.\,]*[0-9]+)\$')
+_dollars_re_start = re.compile(r"\$([0-9\.\,]*[0-9]+)")
+_dollars_re_end = re.compile(r"([(0-9\.\,]*[0-9]+)\$")
 
 # 小数的识别
-_decimal_number_re = re.compile(r'([0-9]+\.\s*[0-9]+)')
+_decimal_number_re = re.compile(r"([0-9]+\.\s*[0-9]+)")
 
 # 分数识别 (形式 "3/4" )
-_fraction_re = re.compile(r'([0-9]+/[0-9]+)')
+_fraction_re = re.compile(r"([0-9]+/[0-9]+)")
 
 # 序数词识别
-_ordinal_re = re.compile(r'[0-9]+(st|nd|rd|th)')
+_ordinal_re = re.compile(r"[0-9]+(st|nd|rd|th)")
 
 # 数字处理
-_number_re = re.compile(r'[0-9]+')
+_number_re = re.compile(r"[0-9]+")
+
 
 def _convert_ordinal(m):
     """
-    标准化序数词, 例如: 1. 2. 3. 4. 5. 6. 
+    标准化序数词, 例如: 1. 2. 3. 4. 5. 6.
     Examples:
         input: "1. "
         output: "1st"
@@ -70,24 +71,26 @@ def _convert_ordinal(m):
     ordinal = _inflect.ordinal(m.group(1))
     return ordinal + ", "
 
+
 def _remove_commas(m):
-    return m.group(1).replace(',', '')
+    return m.group(1).replace(",", "")
+
 
 def _expand_time(m):
     """
     将 24 小时制的时间转换为 12 小时制的时间表示方式。
-    
+
     Examples:
         input: "13:00 / 4:00 / 13:30"
         output: "one o'clock p.m. / four o'clock am. / one thirty p.m."
     """
     hours, minutes = map(int, m.group(1, 2))
-    period = 'a.m.' if hours < 12 else 'p.m.'
+    period = "a.m." if hours < 12 else "p.m."
     if hours > 12:
         hours -= 12
 
     hour_word = _inflect.number_to_words(hours)
-    minute_word = _inflect.number_to_words(minutes) if minutes != 0 else ''
+    minute_word = _inflect.number_to_words(minutes) if minutes != 0 else ""
 
     if minutes == 0:
         return f"{hour_word} o'clock {period}"
@@ -103,7 +106,7 @@ def _expand_measurement(m):
     sign = m.group(3)
     ptr = 1
     # 想不到怎么方便的取数字，又懒得改正则，诶，1.2 反正也是复数读法，干脆直接去掉 "."
-    num = int(m.group(1).replace(sign, '').replace(".",''))
+    num = int(m.group(1).replace(sign, "").replace(".", ""))
     decimal_part = m.group(2)
     # 上面判断的漏洞，比如 0.1 的情况，在这里排除了
     if decimal_part == None and num == 1:
@@ -116,23 +119,24 @@ def _expand_pounds(m):
     没找到特别规范的说明，和美元的处理一样，其实可以把两个合并在一起
     """
     match = m.group(1)
-    parts = match.split('.')
+    parts = match.split(".")
     if len(parts) > 2:
-        return match + ' pounds'    # Unexpected format
+        return match + " pounds"  # Unexpected format
     pounds = int(parts[0]) if parts[0] else 0
-    pence = int(parts[1].ljust(2, '0')) if len(parts) > 1 and parts[1] else 0
+    pence = int(parts[1].ljust(2, "0")) if len(parts) > 1 and parts[1] else 0
     if pounds and pence:
-        pound_unit = 'pound' if pounds == 1 else 'pounds'
-        penny_unit = 'penny' if pence == 1 else 'pence'
-        return '%s %s and %s %s' % (pounds, pound_unit, pence, penny_unit)
+        pound_unit = "pound" if pounds == 1 else "pounds"
+        penny_unit = "penny" if pence == 1 else "pence"
+        return "%s %s and %s %s" % (pounds, pound_unit, pence, penny_unit)
     elif pounds:
-        pound_unit = 'pound' if pounds == 1 else 'pounds'
-        return '%s %s' % (pounds, pound_unit)
+        pound_unit = "pound" if pounds == 1 else "pounds"
+        return "%s %s" % (pounds, pound_unit)
     elif pence:
-        penny_unit = 'penny' if pence == 1 else 'pence'
-        return '%s %s' % (pence, penny_unit)
+        penny_unit = "penny" if pence == 1 else "pence"
+        return "%s %s" % (pence, penny_unit)
     else:
-        return 'zero pounds'
+        return "zero pounds"
+
 
 def _expand_dollars(m):
     """
@@ -142,23 +146,24 @@ def _expand_dollars(m):
         output: "thirty-two dollars and thirty cents" / "six dollars and twenty-four cents"
     """
     match = m.group(1)
-    parts = match.split('.')
+    parts = match.split(".")
     if len(parts) > 2:
-        return match + ' dollars'    # Unexpected format
+        return match + " dollars"  # Unexpected format
     dollars = int(parts[0]) if parts[0] else 0
-    cents = int(parts[1].ljust(2, '0')) if len(parts) > 1 and parts[1] else 0
+    cents = int(parts[1].ljust(2, "0")) if len(parts) > 1 and parts[1] else 0
     if dollars and cents:
-        dollar_unit = 'dollar' if dollars == 1 else 'dollars'
-        cent_unit = 'cent' if cents == 1 else 'cents'
-        return '%s %s and %s %s' % (dollars, dollar_unit, cents, cent_unit)
+        dollar_unit = "dollar" if dollars == 1 else "dollars"
+        cent_unit = "cent" if cents == 1 else "cents"
+        return "%s %s and %s %s" % (dollars, dollar_unit, cents, cent_unit)
     elif dollars:
-        dollar_unit = 'dollar' if dollars == 1 else 'dollars'
-        return '%s %s' % (dollars, dollar_unit)
+        dollar_unit = "dollar" if dollars == 1 else "dollars"
+        return "%s %s" % (dollars, dollar_unit)
     elif cents:
-        cent_unit = 'cent' if cents == 1 else 'cents'
-        return '%s %s' % (cents, cent_unit)
+        cent_unit = "cent" if cents == 1 else "cents"
+        return "%s %s" % (cents, cent_unit)
     else:
-        return 'zero dollars'
+        return "zero dollars"
+
 
 # 小数的处理
 def _expand_decimal_number(m):
@@ -168,11 +173,11 @@ def _expand_decimal_number(m):
         output: "thirteen point two three four"
     """
     match = m.group(1)
-    parts = match.split('.')
+    parts = match.split(".")
     words = []
     # 遍历字符串中的每个字符
     for char in parts[1]:
-        if char == '.':
+        if char == ".":
             words.append("point")
         else:
             words.append(char)
@@ -186,7 +191,7 @@ def _expend_fraction(m):
     规则2: 如果分子大于 1, 在读分母的时候使用序数词复数读法.
     规则3: 当分母为2的时候, 分母读做 half, 并且当分子大于 1 的时候, half 也要用复数读法, 读为 halves.
     Examples:
-    
+
     | Written |	Said |
     |:---:|:---:|
     | 1/3 | one third |
@@ -196,39 +201,41 @@ def _expend_fraction(m):
     | 3/2 | three halves |
     """
     match = m.group(0)
-    numerator, denominator = map(int, match.split('/'))
+    numerator, denominator = map(int, match.split("/"))
 
     numerator_part = _inflect.number_to_words(numerator)
     if denominator == 2:
         if numerator == 1:
-            denominator_part = 'half'
+            denominator_part = "half"
         else:
-            denominator_part = 'halves'
+            denominator_part = "halves"
     elif denominator == 1:
-        return f'{numerator_part}'
+        return f"{numerator_part}"
     else:
         denominator_part = _inflect.ordinal(_inflect.number_to_words(denominator))
         if numerator > 1:
-            denominator_part += 's'
+            denominator_part += "s"
+
+    return f"{numerator_part} {denominator_part}"
 
-    return f'{numerator_part} {denominator_part}'
 
 def _expand_ordinal(m):
     return _inflect.number_to_words(m.group(0))
 
+
 def _expand_number(m):
     num = int(m.group(0))
     if num > 1000 and num < 3000:
         if num == 2000:
-            return 'two thousand'
+            return "two thousand"
         elif num > 2000 and num < 2010:
-            return 'two thousand ' + _inflect.number_to_words(num % 100)
+            return "two thousand " + _inflect.number_to_words(num % 100)
         elif num % 100 == 0:
-            return _inflect.number_to_words(num // 100) + ' hundred'
+            return _inflect.number_to_words(num // 100) + " hundred"
         else:
-            return _inflect.number_to_words(num, andword='', zero='oh', group=2).replace(', ', ' ')
+            return _inflect.number_to_words(num, andword="", zero="oh", group=2).replace(", ", " ")
     else:
-        return _inflect.number_to_words(num, andword='')
+        return _inflect.number_to_words(num, andword="")
 
 
 def normalize(text):
@@ -238,7 +245,7 @@ def normalize(text):
     """
 
     text = re.sub(_ordinal_number_re, _convert_ordinal, text)
-    text = re.sub(r'(?<!\d)-|-(?!\d)', ' minus ', text)
+    text = re.sub(r"(?<!\d)-|-(?!\d)", " minus ", text)
     text = re.sub(_comma_number_re, _remove_commas, text)
     text = re.sub(_time_re, _expand_time, text)
     text = re.sub(_measurement_re, _expand_measurement, text)
@@ -251,19 +258,20 @@ def normalize(text):
     text = re.sub(_ordinal_re, _expand_ordinal, text)
     text = re.sub(_number_re, _expand_number, text)
 
-    text = ''.join(char for char in unicodedata.normalize('NFD', text)
-                    if unicodedata.category(char) != 'Mn')  # Strip accents
+    text = "".join(
+        char for char in unicodedata.normalize("NFD", text) if unicodedata.category(char) != "Mn"
+    )  # Strip accents
 
     text = re.sub("%", " percent", text)
     text = re.sub("[^ A-Za-z'.,?!\-]", "", text)
     text = re.sub(r"(?i)i\.e\.", "that is", text)
     text = re.sub(r"(?i)e\.g\.", "for example", text)
     # 增加纯大写单词拆分
-    text = re.sub(r'(?<!^)(?<![\s])([A-Z])', r' \1', text)
+    text = re.sub(r"(?<!^)(?<![\s])([A-Z])", r" \1", text)
     return text
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     # 我觉得其实可以把切分结果展示出来（只读，或者修改不影响传给TTS的实际text）
     # 然后让用户确认后再输入给 TTS，可以让用户检查自己有没有不标准的输入
     print(normalize("1. test ordinal number 1st"))
@@ -272,4 +280,4 @@ if __name__ == '__main__':
     print(normalize("1st, 22nd"))
     print(normalize("a test 20h, 1.2s, 1L, 0.1km"))
     print(normalize("a test of time 4:00, 13:00, 13:30"))
-    print(normalize("a test of temperature 4°F, 23°C, -19°C"))
+    print(normalize("a test of temperature 4°F, 23°C, -19°C"))

GPT_SoVITS/text/english.py

@@ -8,10 +8,10 @@ from text.symbols import punctuation
 
 from text.symbols2 import symbols
 
-import unicodedata
 from builtins import str as unicode
 from text.en_normalization.expend import normalize
 from nltk.tokenize import TweetTokenizer
+
 word_tokenize = TweetTokenizer().tokenize
 from nltk import pos_tag
 
@@ -122,9 +122,9 @@ def replace_phs(phs):
 
 
 def replace_consecutive_punctuation(text):
-    punctuations = ''.join(re.escape(p) for p in punctuation)
-    pattern = f'([{punctuations}\s])([{punctuations}])+'
-    result = re.sub(pattern, r'\1', text)
+    punctuations = "".join(re.escape(p) for p in punctuation)
+    pattern = f"([{punctuations}\s])([{punctuations}])+"
+    result = re.sub(pattern, r"\1", text)
     return result
 
 
@@ -183,6 +183,7 @@ def read_dict_new():
 
     return g2p_dict
 
+
 def hot_reload_hot(g2p_dict):
     with open(CMU_DICT_HOT_PATH) as f:
         line = f.readline()
@@ -259,9 +260,12 @@ class en_G2p(G2p):
             del self.cmu[word.lower()]
 
         # 修正多音字
-        self.homograph2features["read"] = (['R', 'IY1', 'D'], ['R', 'EH1', 'D'], 'VBP')
-        self.homograph2features["complex"] = (['K', 'AH0', 'M', 'P', 'L', 'EH1', 'K', 'S'], ['K', 'AA1', 'M', 'P', 'L', 'EH0', 'K', 'S'], 'JJ')
-
+        self.homograph2features["read"] = (["R", "IY1", "D"], ["R", "EH1", "D"], "VBP")
+        self.homograph2features["complex"] = (
+            ["K", "AH0", "M", "P", "L", "EH1", "K", "S"],
+            ["K", "AA1", "M", "P", "L", "EH0", "K", "S"],
+            "JJ",
+        )
 
     def __call__(self, text):
         # tokenization
@@ -280,7 +284,7 @@ class en_G2p(G2p):
             elif len(word) == 1:
                 # 单读 A 发音修正, 这里需要原格式 o_word 判断大写
                 if o_word == "A":
-                    pron = ['EY1']
+                    pron = ["EY1"]
                 else:
                     pron = self.cmu[word][0]
             # g2p_en 原版多音字处理
@@ -289,7 +293,7 @@ class en_G2p(G2p):
                 if pos.startswith(pos1):
                     pron = pron1
                 # pos1比pos长仅出现在read
-                elif len(pos) < len(pos1) and pos == pos1[:len(pos)]:
+                elif len(pos) < len(pos1) and pos == pos1[: len(pos)]:
                     pron = pron1
                 else:
                     pron = pron2
@@ -302,7 +306,6 @@ class en_G2p(G2p):
 
         return prons[:-1]
 
-
     def qryword(self, o_word):
         word = o_word.lower()
 
@@ -320,7 +323,7 @@ class en_G2p(G2p):
             for w in word:
                 # 单读 A 发音修正, 此处不存在大写的情况
                 if w == "a":
-                    phones.extend(['EY1'])
+                    phones.extend(["EY1"])
                 elif not w.isalpha():
                     phones.extend([w])
                 else:
@@ -331,23 +334,23 @@ class en_G2p(G2p):
         if re.match(r"^([a-z]+)('s)$", word):
             phones = self.qryword(word[:-2])[:]
             # P T K F TH HH 无声辅音结尾 's 发 ['S']
-            if phones[-1] in ['P', 'T', 'K', 'F', 'TH', 'HH']:
-                phones.extend(['S'])
+            if phones[-1] in ["P", "T", "K", "F", "TH", "HH"]:
+                phones.extend(["S"])
             # S Z SH ZH CH JH 擦声结尾 's 发 ['IH1', 'Z'] 或 ['AH0', 'Z']
-            elif phones[-1] in ['S', 'Z', 'SH', 'ZH', 'CH', 'JH']:
-                phones.extend(['AH0', 'Z'])
+            elif phones[-1] in ["S", "Z", "SH", "ZH", "CH", "JH"]:
+                phones.extend(["AH0", "Z"])
             # B D G DH V M N NG L R W Y 有声辅音结尾 's 发 ['Z']
             # AH0 AH1 AH2 EY0 EY1 EY2 AE0 AE1 AE2 EH0 EH1 EH2 OW0 OW1 OW2 UH0 UH1 UH2 IY0 IY1 IY2 AA0 AA1 AA2 AO0 AO1 AO2
             # ER ER0 ER1 ER2 UW0 UW1 UW2 AY0 AY1 AY2 AW0 AW1 AW2 OY0 OY1 OY2 IH IH0 IH1 IH2 元音结尾 's 发 ['Z']
             else:
-                phones.extend(['Z'])
+                phones.extend(["Z"])
             return phones
 
         # 尝试进行分词，应对复合词
         comps = wordsegment.segment(word.lower())
 
         # 无法分词的送回去预测
-        if len(comps)==1:
+        if len(comps) == 1:
             return self.predict(word)
 
         # 可以分词的递归处理

GPT_SoVITS/text/g2pw/__init__.py

@@ -1 +1 @@
-from text.g2pw.g2pw import *
+from text.g2pw.g2pw import *

GPT_SoVITS/text/g2pw/dataset.py

@@ -15,6 +15,7 @@
 Credits
     This code is modified from https://github.com/GitYCC/g2pW
 """
+
 from typing import Dict
 from typing import List
 from typing import Tuple
@@ -23,21 +24,24 @@ import numpy as np
 
 from .utils import tokenize_and_map
 
-ANCHOR_CHAR = '▁'
+ANCHOR_CHAR = "▁"
 
 
-def prepare_onnx_input(tokenizer,
-                       labels: List[str],
-                       char2phonemes: Dict[str, List[int]],
-                       chars: List[str],
-                       texts: List[str],
-                       query_ids: List[int],
-                       use_mask: bool=False,
-                       window_size: int=None,
-                       max_len: int=512) -> Dict[str, np.array]:
+def prepare_onnx_input(
+    tokenizer,
+    labels: List[str],
+    char2phonemes: Dict[str, List[int]],
+    chars: List[str],
+    texts: List[str],
+    query_ids: List[int],
+    use_mask: bool = False,
+    window_size: int = None,
+    max_len: int = 512,
+) -> Dict[str, np.array]:
     if window_size is not None:
         truncated_texts, truncated_query_ids = _truncate_texts(
-            window_size=window_size, texts=texts, query_ids=query_ids)
+            window_size=window_size, texts=texts, query_ids=query_ids
+        )
     input_ids = []
     token_type_ids = []
     attention_masks = []
@@ -50,33 +54,27 @@ def prepare_onnx_input(tokenizer,
         query_id = (truncated_query_ids if window_size else query_ids)[idx]
 
         try:
-            tokens, text2token, token2text = tokenize_and_map(
-                tokenizer=tokenizer, text=text)
+            tokens, text2token, token2text = tokenize_and_map(tokenizer=tokenizer, text=text)
         except Exception:
             print(f'warning: text "{text}" is invalid')
             return {}
 
         text, query_id, tokens, text2token, token2text = _truncate(
-            max_len=max_len,
-            text=text,
-            query_id=query_id,
-            tokens=tokens,
-            text2token=text2token,
-            token2text=token2text)
+            max_len=max_len, text=text, query_id=query_id, tokens=tokens, text2token=text2token, token2text=token2text
+        )
 
-        processed_tokens = ['[CLS]'] + tokens + ['[SEP]']
+        processed_tokens = ["[CLS]"] + tokens + ["[SEP]"]
 
-        input_id = list(
-            np.array(tokenizer.convert_tokens_to_ids(processed_tokens)))
-        token_type_id = list(np.zeros((len(processed_tokens), ), dtype=int))
-        attention_mask = list(np.ones((len(processed_tokens), ), dtype=int))
+        input_id = list(np.array(tokenizer.convert_tokens_to_ids(processed_tokens)))
+        token_type_id = list(np.zeros((len(processed_tokens),), dtype=int))
+        attention_mask = list(np.ones((len(processed_tokens),), dtype=int))
 
         query_char = text[query_id]
-        phoneme_mask = [1 if i in char2phonemes[query_char] else 0 for i in range(len(labels))] \
-            if use_mask else [1] * len(labels)
+        phoneme_mask = (
+            [1 if i in char2phonemes[query_char] else 0 for i in range(len(labels))] if use_mask else [1] * len(labels)
+        )
         char_id = chars.index(query_char)
-        position_id = text2token[
-            query_id] + 1  # [CLS] token locate at first place
+        position_id = text2token[query_id] + 1  # [CLS] token locate at first place
 
         input_ids.append(input_id)
         token_type_ids.append(token_type_id)
@@ -86,18 +84,17 @@ def prepare_onnx_input(tokenizer,
         position_ids.append(position_id)
 
     outputs = {
-        'input_ids': np.array(input_ids).astype(np.int64),
-        'token_type_ids': np.array(token_type_ids).astype(np.int64),
-        'attention_masks': np.array(attention_masks).astype(np.int64),
-        'phoneme_masks': np.array(phoneme_masks).astype(np.float32),
-        'char_ids': np.array(char_ids).astype(np.int64),
-        'position_ids': np.array(position_ids).astype(np.int64),
+        "input_ids": np.array(input_ids).astype(np.int64),
+        "token_type_ids": np.array(token_type_ids).astype(np.int64),
+        "attention_masks": np.array(attention_masks).astype(np.int64),
+        "phoneme_masks": np.array(phoneme_masks).astype(np.float32),
+        "char_ids": np.array(char_ids).astype(np.int64),
+        "position_ids": np.array(position_ids).astype(np.int64),
     }
     return outputs
 
 
-def _truncate_texts(window_size: int, texts: List[str],
-                    query_ids: List[int]) -> Tuple[List[str], List[int]]:
+def _truncate_texts(window_size: int, texts: List[str], query_ids: List[int]) -> Tuple[List[str], List[int]]:
     truncated_texts = []
     truncated_query_ids = []
     for text, query_id in zip(texts, query_ids):
@@ -111,12 +108,9 @@ def _truncate_texts(window_size: int, texts: List[str],
     return truncated_texts, truncated_query_ids
 
 
-def _truncate(max_len: int,
-              text: str,
-              query_id: int,
-              tokens: List[str],
-              text2token: List[int],
-              token2text: List[Tuple[int]]):
+def _truncate(
+    max_len: int, text: str, query_id: int, tokens: List[str], text2token: List[int], token2text: List[Tuple[int]]
+):
     truncate_len = max_len - 2
     if len(tokens) <= truncate_len:
         return (text, query_id, tokens, text2token, token2text)
@@ -137,14 +131,16 @@ def _truncate(max_len: int,
     start = token2text[token_start][0]
     end = token2text[token_end - 1][1]
 
-    return (text[start:end], query_id - start, tokens[token_start:token_end], [
-        i - token_start if i is not None else None
-        for i in text2token[start:end]
-    ], [(s - start, e - start) for s, e in token2text[token_start:token_end]])
+    return (
+        text[start:end],
+        query_id - start,
+        tokens[token_start:token_end],
+        [i - token_start if i is not None else None for i in text2token[start:end]],
+        [(s - start, e - start) for s, e in token2text[token_start:token_end]],
+    )
 
 
-def get_phoneme_labels(polyphonic_chars: List[List[str]]
-                       ) -> Tuple[List[str], Dict[str, List[int]]]:
+def get_phoneme_labels(polyphonic_chars: List[List[str]]) -> Tuple[List[str], Dict[str, List[int]]]:
     labels = sorted(list(set([phoneme for char, phoneme in polyphonic_chars])))
     char2phonemes = {}
     for char, phoneme in polyphonic_chars:
@@ -154,13 +150,11 @@ def get_phoneme_labels(polyphonic_chars: List[List[str]]
     return labels, char2phonemes
 
 
-def get_char_phoneme_labels(polyphonic_chars: List[List[str]]
-                            ) -> Tuple[List[str], Dict[str, List[int]]]:
-    labels = sorted(
-        list(set([f'{char} {phoneme}' for char, phoneme in polyphonic_chars])))
+def get_char_phoneme_labels(polyphonic_chars: List[List[str]]) -> Tuple[List[str], Dict[str, List[int]]]:
+    labels = sorted(list(set([f"{char} {phoneme}" for char, phoneme in polyphonic_chars])))
     char2phonemes = {}
     for char, phoneme in polyphonic_chars:
         if char not in char2phonemes:
             char2phonemes[char] = []
-        char2phonemes[char].append(labels.index(f'{char} {phoneme}'))
+        char2phonemes[char].append(labels.index(f"{char} {phoneme}"))
     return labels, char2phonemes

GPT_SoVITS/text/g2pw/g2pw.py

@@ -17,17 +17,25 @@ PP_FIX_DICT_PATH = os.path.join(current_file_path, "polyphonic-fix.rep")
 
 
 class G2PWPinyin(Pinyin):
-    def __init__(self, model_dir='G2PWModel/', model_source=None,
-                 enable_non_tradional_chinese=True,
-                 v_to_u=False, neutral_tone_with_five=False, tone_sandhi=False, **kwargs):
+    def __init__(
+        self,
+        model_dir="G2PWModel/",
+        model_source=None,
+        enable_non_tradional_chinese=True,
+        v_to_u=False,
+        neutral_tone_with_five=False,
+        tone_sandhi=False,
+        **kwargs,
+    ):
         self._g2pw = G2PWOnnxConverter(
             model_dir=model_dir,
-            style='pinyin',
+            style="pinyin",
             model_source=model_source,
             enable_non_tradional_chinese=enable_non_tradional_chinese,
         )
         self._converter = Converter(
-            self._g2pw, v_to_u=v_to_u,
+            self._g2pw,
+            v_to_u=v_to_u,
             neutral_tone_with_five=neutral_tone_with_five,
             tone_sandhi=tone_sandhi,
         )
@@ -37,31 +45,25 @@ class G2PWPinyin(Pinyin):
 
 
 class Converter(UltimateConverter):
-    def __init__(self, g2pw_instance, v_to_u=False,
-                 neutral_tone_with_five=False,
-                 tone_sandhi=False, **kwargs):
+    def __init__(self, g2pw_instance, v_to_u=False, neutral_tone_with_five=False, tone_sandhi=False, **kwargs):
         super(Converter, self).__init__(
-            v_to_u=v_to_u,
-            neutral_tone_with_five=neutral_tone_with_five,
-            tone_sandhi=tone_sandhi, **kwargs)
+            v_to_u=v_to_u, neutral_tone_with_five=neutral_tone_with_five, tone_sandhi=tone_sandhi, **kwargs
+        )
 
         self._g2pw = g2pw_instance
 
     def convert(self, words, style, heteronym, errors, strict, **kwargs):
         pys = []
         if RE_HANS.match(words):
-            pys = self._to_pinyin(words, style=style, heteronym=heteronym,
-                                  errors=errors, strict=strict)
+            pys = self._to_pinyin(words, style=style, heteronym=heteronym, errors=errors, strict=strict)
             post_data = self.post_pinyin(words, heteronym, pys)
             if post_data is not None:
                 pys = post_data
 
-            pys = self.convert_styles(
-                pys, words, style, heteronym, errors, strict)
+            pys = self.convert_styles(pys, words, style, heteronym, errors, strict)
 
         else:
-            py = self.handle_nopinyin(words, style=style, errors=errors,
-                                      heteronym=heteronym, strict=strict)
+            py = self.handle_nopinyin(words, style=style, errors=errors, heteronym=heteronym, strict=strict)
             if py:
                 pys.extend(py)
 
@@ -73,13 +75,11 @@ class Converter(UltimateConverter):
         g2pw_pinyin = self._g2pw(han)
 
         if not g2pw_pinyin:  # g2pw 不支持的汉字改为使用 pypinyin 原有逻辑
-            return super(Converter, self).convert(
-                         han, Style.TONE, heteronym, errors, strict, **kwargs)
+            return super(Converter, self).convert(han, Style.TONE, heteronym, errors, strict, **kwargs)
 
         for i, item in enumerate(g2pw_pinyin[0]):
             if item is None:  # g2pw 不支持的汉字改为使用 pypinyin 原有逻辑
-                py = super(Converter, self).convert(
-                           han[i], Style.TONE, heteronym, errors, strict, **kwargs)
+                py = super(Converter, self).convert(han[i], Style.TONE, heteronym, errors, strict, **kwargs)
                 pinyins.extend(py)
             else:
                 pinyins.append([to_tone(item)])
@@ -104,7 +104,7 @@ def _remove_dup_and_empty(lst_list):
         if lst:
             new_lst_list.append(lst)
         else:
-            new_lst_list.append([''])
+            new_lst_list.append([""])
 
     return new_lst_list
 
@@ -127,17 +127,17 @@ def get_dict():
 
 def read_dict():
     polyphonic_dict = {}
-    with open(PP_DICT_PATH,encoding="utf-8") as f:
+    with open(PP_DICT_PATH, encoding="utf-8") as f:
         line = f.readline()
         while line:
-            key, value_str = line.split(':')
+            key, value_str = line.split(":")
             value = eval(value_str.strip())
             polyphonic_dict[key.strip()] = value
             line = f.readline()
-    with open(PP_FIX_DICT_PATH,encoding="utf-8") as f:
+    with open(PP_FIX_DICT_PATH, encoding="utf-8") as f:
         line = f.readline()
         while line:
-            key, value_str = line.split(':')
+            key, value_str = line.split(":")
             value = eval(value_str.strip())
             polyphonic_dict[key.strip()] = value
             line = f.readline()

GPT_SoVITS/text/g2pw/onnx_api.py

@@ -2,44 +2,43 @@
 # This code is modified from https://github.com/GitYCC/g2pW
 
 import warnings
+
 warnings.filterwarnings("ignore")
 import json
 import os
-import zipfile,requests
-from typing import Any
-from typing import Dict
-from typing import List
-from typing import Tuple
+import zipfile
+from typing import Any, Dict, List, Tuple
 
 import numpy as np
 import onnxruntime
+import requests
+
 onnxruntime.set_default_logger_severity(3)
 from opencc import OpenCC
+from pypinyin import Style, pinyin
 from transformers import AutoTokenizer
-from pypinyin import pinyin
-from pypinyin import Style
 
-from .dataset import get_char_phoneme_labels
-from .dataset import get_phoneme_labels
-from .dataset import prepare_onnx_input
-from .utils import load_config
 from ..zh_normalization.char_convert import tranditional_to_simplified
+from .dataset import get_char_phoneme_labels, get_phoneme_labels, prepare_onnx_input
+from .utils import load_config
 
-model_version = '1.1'
+model_version = "1.1"
 
 
-def predict(session, onnx_input: Dict[str, Any],
-            labels: List[str]) -> Tuple[List[str], List[float]]:
+def predict(session, onnx_input: Dict[str, Any], labels: List[str]) -> Tuple[List[str], List[float]]:
     all_preds = []
     all_confidences = []
-    probs = session.run([], {
-        "input_ids": onnx_input['input_ids'],
-        "token_type_ids": onnx_input['token_type_ids'],
-        "attention_mask": onnx_input['attention_masks'],
-        "phoneme_mask": onnx_input['phoneme_masks'],
-        "char_ids": onnx_input['char_ids'],
-        "position_ids": onnx_input['position_ids']
-    })[0]
+    probs = session.run(
+        [],
+        {
+            "input_ids": onnx_input["input_ids"],
+            "token_type_ids": onnx_input["token_type_ids"],
+            "attention_mask": onnx_input["attention_masks"],
+            "phoneme_mask": onnx_input["phoneme_masks"],
+            "char_ids": onnx_input["char_ids"],
+            "position_ids": onnx_input["position_ids"],
+        },
+    )[0]
 
     preds = np.argmax(probs, axis=1).tolist()
     max_probs = []
@@ -51,17 +50,17 @@ def predict(session, onnx_input: Dict[str, Any],
     return all_preds, all_confidences
 
 
-def download_and_decompress(model_dir: str='G2PWModel/'):
+def download_and_decompress(model_dir: str = "G2PWModel/"):
     if not os.path.exists(model_dir):
         parent_directory = os.path.dirname(model_dir)
-        zip_dir = os.path.join(parent_directory,"G2PWModel_1.1.zip")
-        extract_dir = os.path.join(parent_directory,"G2PWModel_1.1")
-        extract_dir_new = os.path.join(parent_directory,"G2PWModel")
+        zip_dir = os.path.join(parent_directory, "G2PWModel_1.1.zip")
+        extract_dir = os.path.join(parent_directory, "G2PWModel_1.1")
+        extract_dir_new = os.path.join(parent_directory, "G2PWModel")
         print("Downloading g2pw model...")
-        modelscope_url = "https://www.modelscope.cn/models/kamiorinn/g2pw/resolve/master/G2PWModel_1.1.zip"#"https://paddlespeech.cdn.bcebos.com/Parakeet/released_models/g2p/G2PWModel_1.1.zip"
+        modelscope_url = "https://www.modelscope.cn/models/kamiorinn/g2pw/resolve/master/G2PWModel_1.1.zip"  # "https://paddlespeech.cdn.bcebos.com/Parakeet/released_models/g2p/G2PWModel_1.1.zip"
         with requests.get(modelscope_url, stream=True) as r:
             r.raise_for_status()
-            with open(zip_dir, 'wb') as f:
+            with open(zip_dir, "wb") as f:
                 for chunk in r.iter_content(chunk_size=8192):
                     if chunk:
                         f.write(chunk)
@@ -69,17 +68,20 @@ def download_and_decompress(model_dir: str='G2PWModel/'):
         print("Extracting g2pw model...")
         with zipfile.ZipFile(zip_dir, "r") as zip_ref:
             zip_ref.extractall(parent_directory)
-        
+
         os.rename(extract_dir, extract_dir_new)
 
     return model_dir
 
+
 class G2PWOnnxConverter:
-    def __init__(self,
-                 model_dir: str='G2PWModel/',
-                 style: str='bopomofo',
-                 model_source: str=None,
-                 enable_non_tradional_chinese: bool=False):
+    def __init__(
+        self,
+        model_dir: str = "G2PWModel/",
+        style: str = "bopomofo",
+        model_source: str = None,
+        enable_non_tradional_chinese: bool = False,
+    ):
         uncompress_path = download_and_decompress(model_dir)
 
         sess_options = onnxruntime.SessionOptions()
@@ -87,41 +89,59 @@ class G2PWOnnxConverter:
         sess_options.execution_mode = onnxruntime.ExecutionMode.ORT_SEQUENTIAL
         sess_options.intra_op_num_threads = 2
         try:
-            self.session_g2pW = onnxruntime.InferenceSession(os.path.join(uncompress_path, 'g2pW.onnx'),sess_options=sess_options, providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
+            self.session_g2pW = onnxruntime.InferenceSession(
+                os.path.join(uncompress_path, "g2pW.onnx"),
+                sess_options=sess_options,
+                providers=["CUDAExecutionProvider", "CPUExecutionProvider"],
+            )
         except:
-            self.session_g2pW = onnxruntime.InferenceSession(os.path.join(uncompress_path, 'g2pW.onnx'),sess_options=sess_options, providers=['CPUExecutionProvider'])
-        self.config = load_config(
-            config_path=os.path.join(uncompress_path, 'config.py'),
-            use_default=True)
+            self.session_g2pW = onnxruntime.InferenceSession(
+                os.path.join(uncompress_path, "g2pW.onnx"),
+                sess_options=sess_options,
+                providers=["CPUExecutionProvider"],
+            )
+        self.config = load_config(config_path=os.path.join(uncompress_path, "config.py"), use_default=True)
 
         self.model_source = model_source if model_source else self.config.model_source
         self.enable_opencc = enable_non_tradional_chinese
 
         self.tokenizer = AutoTokenizer.from_pretrained(self.model_source)
 
-        polyphonic_chars_path = os.path.join(uncompress_path,
-                                             'POLYPHONIC_CHARS.txt')
-        monophonic_chars_path = os.path.join(uncompress_path,
-                                             'MONOPHONIC_CHARS.txt')
+        polyphonic_chars_path = os.path.join(uncompress_path, "POLYPHONIC_CHARS.txt")
+        monophonic_chars_path = os.path.join(uncompress_path, "MONOPHONIC_CHARS.txt")
         self.polyphonic_chars = [
-            line.split('\t')
-            for line in open(polyphonic_chars_path, encoding='utf-8').read()
-            .strip().split('\n')
+            line.split("\t") for line in open(polyphonic_chars_path, encoding="utf-8").read().strip().split("\n")
         ]
         self.non_polyphonic = {
-            '一', '不', '和', '咋', '嗲', '剖', '差', '攢', '倒', '難', '奔', '勁', '拗',
-            '肖', '瘙', '誒', '泊', '听', '噢'
+            "一",
+            "不",
+            "和",
+            "咋",
+            "嗲",
+            "剖",
+            "差",
+            "攢",
+            "倒",
+            "難",
+            "奔",
+            "勁",
+            "拗",
+            "肖",
+            "瘙",
+            "誒",
+            "泊",
+            "听",
+            "噢",
         }
-        self.non_monophonic = {'似', '攢'}
+        self.non_monophonic = {"似", "攢"}
         self.monophonic_chars = [
-            line.split('\t')
-            for line in open(monophonic_chars_path, encoding='utf-8').read()
-            .strip().split('\n')
+            line.split("\t") for line in open(monophonic_chars_path, encoding="utf-8").read().strip().split("\n")
         ]
-        self.labels, self.char2phonemes = get_char_phoneme_labels(
-            polyphonic_chars=self.polyphonic_chars
-        ) if self.config.use_char_phoneme else get_phoneme_labels(
-            polyphonic_chars=self.polyphonic_chars)
+        self.labels, self.char2phonemes = (
+            get_char_phoneme_labels(polyphonic_chars=self.polyphonic_chars)
+            if self.config.use_char_phoneme
+            else get_phoneme_labels(polyphonic_chars=self.polyphonic_chars)
+        )
 
         self.chars = sorted(list(self.char2phonemes.keys()))
 
@@ -130,41 +150,29 @@ class G2PWOnnxConverter:
             if char in self.polyphonic_chars_new:
                 self.polyphonic_chars_new.remove(char)
 
-        self.monophonic_chars_dict = {
-            char: phoneme
-            for char, phoneme in self.monophonic_chars
-        }
+        self.monophonic_chars_dict = {char: phoneme for char, phoneme in self.monophonic_chars}
         for char in self.non_monophonic:
             if char in self.monophonic_chars_dict:
                 self.monophonic_chars_dict.pop(char)
 
-        self.pos_tags = [
-            'UNK', 'A', 'C', 'D', 'I', 'N', 'P', 'T', 'V', 'DE', 'SHI'
-        ]
+        self.pos_tags = ["UNK", "A", "C", "D", "I", "N", "P", "T", "V", "DE", "SHI"]
 
-        with open(
-                os.path.join(uncompress_path,
-                             'bopomofo_to_pinyin_wo_tune_dict.json'),
-                'r',
-                encoding='utf-8') as fr:
+        with open(os.path.join(uncompress_path, "bopomofo_to_pinyin_wo_tune_dict.json"), "r", encoding="utf-8") as fr:
             self.bopomofo_convert_dict = json.load(fr)
         self.style_convert_func = {
-            'bopomofo': lambda x: x,
-            'pinyin': self._convert_bopomofo_to_pinyin,
+            "bopomofo": lambda x: x,
+            "pinyin": self._convert_bopomofo_to_pinyin,
         }[style]
 
-        with open(
-                os.path.join(uncompress_path, 'char_bopomofo_dict.json'),
-                'r',
-                encoding='utf-8') as fr:
+        with open(os.path.join(uncompress_path, "char_bopomofo_dict.json"), "r", encoding="utf-8") as fr:
             self.char_bopomofo_dict = json.load(fr)
 
         if self.enable_opencc:
-            self.cc = OpenCC('s2tw')
+            self.cc = OpenCC("s2tw")
 
     def _convert_bopomofo_to_pinyin(self, bopomofo: str) -> str:
         tone = bopomofo[-1]
-        assert tone in '12345'
+        assert tone in "12345"
         component = self.bopomofo_convert_dict.get(bopomofo[:-1])
         if component:
             return component + tone
@@ -184,8 +192,7 @@ class G2PWOnnxConverter:
                 translated_sentences.append(translated_sent)
             sentences = translated_sentences
 
-        texts, query_ids, sent_ids, partial_results = self._prepare_data(
-            sentences=sentences)
+        texts, query_ids, sent_ids, partial_results = self._prepare_data(sentences=sentences)
         if len(texts) == 0:
             # sentences no polyphonic words
             return partial_results
@@ -198,14 +205,12 @@ class G2PWOnnxConverter:
             texts=texts,
             query_ids=query_ids,
             use_mask=self.config.use_mask,
-            window_size=None)
+            window_size=None,
+        )
 
-        preds, confidences = predict(
-            session=self.session_g2pW,
-            onnx_input=onnx_input,
-            labels=self.labels)
+        preds, confidences = predict(session=self.session_g2pW, onnx_input=onnx_input, labels=self.labels)
         if self.config.use_char_phoneme:
-            preds = [pred.split(' ')[1] for pred in preds]
+            preds = [pred.split(" ")[1] for pred in preds]
 
         results = partial_results
         for sent_id, query_id, pred in zip(sent_ids, query_ids, preds):
@@ -213,15 +218,12 @@ class G2PWOnnxConverter:
 
         return results
 
-    def _prepare_data(
-            self, sentences: List[str]
-    ) -> Tuple[List[str], List[int], List[int], List[List[str]]]:
+    def _prepare_data(self, sentences: List[str]) -> Tuple[List[str], List[int], List[int], List[List[str]]]:
         texts, query_ids, sent_ids, partial_results = [], [], [], []
         for sent_id, sent in enumerate(sentences):
             # pypinyin works well for Simplified Chinese than Traditional Chinese
             sent_s = tranditional_to_simplified(sent)
-            pypinyin_result = pinyin(
-                sent_s, neutral_tone_with_five=True, style=Style.TONE3)
+            pypinyin_result = pinyin(sent_s, neutral_tone_with_five=True, style=Style.TONE3)
             partial_result = [None] * len(sent)
             for i, char in enumerate(sent):
                 if char in self.polyphonic_chars_new:
@@ -229,8 +231,7 @@ class G2PWOnnxConverter:
                     query_ids.append(i)
                     sent_ids.append(sent_id)
                 elif char in self.monophonic_chars_dict:
-                    partial_result[i] = self.style_convert_func(
-                        self.monophonic_chars_dict[char])
+                    partial_result[i] = self.style_convert_func(self.monophonic_chars_dict[char])
                 elif char in self.char_bopomofo_dict:
                     partial_result[i] = pypinyin_result[i][0]
                     # partial_result[i] =  self.style_convert_func(self.char_bopomofo_dict[char][0])

GPT_SoVITS/text/g2pw/utils.py

@@ -15,6 +15,7 @@
 Credits
     This code is modified from https://github.com/GitYCC/g2pW
 """
+
 import os
 import re
 
@@ -24,14 +25,14 @@ def wordize_and_map(text: str):
     index_map_from_text_to_word = []
     index_map_from_word_to_text = []
     while len(text) > 0:
-        match_space = re.match(r'^ +', text)
+        match_space = re.match(r"^ +", text)
         if match_space:
             space_str = match_space.group(0)
             index_map_from_text_to_word += [None] * len(space_str)
-            text = text[len(space_str):]
+            text = text[len(space_str) :]
             continue
 
-        match_en = re.match(r'^[a-zA-Z0-9]+', text)
+        match_en = re.match(r"^[a-zA-Z0-9]+", text)
         if match_en:
             en_word = match_en.group(0)
 
@@ -42,7 +43,7 @@ def wordize_and_map(text: str):
             index_map_from_text_to_word += [len(words)] * len(en_word)
 
             words.append(en_word)
-            text = text[len(en_word):]
+            text = text[len(en_word) :]
         else:
             word_start_pos = len(index_map_from_text_to_word)
             word_end_pos = word_start_pos + 1
@@ -63,15 +64,14 @@ def tokenize_and_map(tokenizer, text: str):
     for word, (word_start, word_end) in zip(words, word2text):
         word_tokens = tokenizer.tokenize(word)
 
-        if len(word_tokens) == 0 or word_tokens == ['[UNK]']:
+        if len(word_tokens) == 0 or word_tokens == ["[UNK]"]:
             index_map_from_token_to_text.append((word_start, word_end))
-            tokens.append('[UNK]')
+            tokens.append("[UNK]")
         else:
             current_word_start = word_start
             for word_token in word_tokens:
-                word_token_len = len(re.sub(r'^##', '', word_token))
-                index_map_from_token_to_text.append(
-                    (current_word_start, current_word_start + word_token_len))
+                word_token_len = len(re.sub(r"^##", "", word_token))
+                index_map_from_token_to_text.append((current_word_start, current_word_start + word_token_len))
                 current_word_start = current_word_start + word_token_len
                 tokens.append(word_token)
 
@@ -85,53 +85,51 @@ def tokenize_and_map(tokenizer, text: str):
 
 def _load_config(config_path: os.PathLike):
     import importlib.util
-    spec = importlib.util.spec_from_file_location('__init__', config_path)
+
+    spec = importlib.util.spec_from_file_location("__init__", config_path)
     config = importlib.util.module_from_spec(spec)
     spec.loader.exec_module(config)
     return config
 
 
 default_config_dict = {
-    'manual_seed': 1313,
-    'model_source': 'bert-base-chinese',
-    'window_size': 32,
-    'num_workers': 2,
-    'use_mask': True,
-    'use_char_phoneme': False,
-    'use_conditional': True,
-    'param_conditional': {
-        'affect_location': 'softmax',
-        'bias': True,
-        'char-linear': True,
-        'pos-linear': False,
-        'char+pos-second': True,
-        'char+pos-second_lowrank': False,
-        'lowrank_size': 0,
-        'char+pos-second_fm': False,
-        'fm_size': 0,
-        'fix_mode': None,
-        'count_json': 'train.count.json'
+    "manual_seed": 1313,
+    "model_source": "bert-base-chinese",
+    "window_size": 32,
+    "num_workers": 2,
+    "use_mask": True,
+    "use_char_phoneme": False,
+    "use_conditional": True,
+    "param_conditional": {
+        "affect_location": "softmax",
+        "bias": True,
+        "char-linear": True,
+        "pos-linear": False,
+        "char+pos-second": True,
+        "char+pos-second_lowrank": False,
+        "lowrank_size": 0,
+        "char+pos-second_fm": False,
+        "fm_size": 0,
+        "fix_mode": None,
+        "count_json": "train.count.json",
     },
-    'lr': 5e-5,
-    'val_interval': 200,
-    'num_iter': 10000,
-    'use_focal': False,
-    'param_focal': {
-        'alpha': 0.0,
-        'gamma': 0.7
+    "lr": 5e-5,
+    "val_interval": 200,
+    "num_iter": 10000,
+    "use_focal": False,
+    "param_focal": {"alpha": 0.0, "gamma": 0.7},
+    "use_pos": True,
+    "param_pos ": {
+        "weight": 0.1,
+        "pos_joint_training": True,
+        "train_pos_path": "train.pos",
+        "valid_pos_path": "dev.pos",
+        "test_pos_path": "test.pos",
     },
-    'use_pos': True,
-    'param_pos ': {
-        'weight': 0.1,
-        'pos_joint_training': True,
-        'train_pos_path': 'train.pos',
-        'valid_pos_path': 'dev.pos',
-        'test_pos_path': 'test.pos'
-    }
 }
 
 
-def load_config(config_path: os.PathLike, use_default: bool=False):
+def load_config(config_path: os.PathLike, use_default: bool = False):
     config = _load_config(config_path)
     if use_default:
         for attr, val in default_config_dict.items():

GPT_SoVITS/text/japanese.py

@@ -2,43 +2,51 @@
 import re
 import os
 import hashlib
+
 try:
     import pyopenjtalk
+
     current_file_path = os.path.dirname(__file__)
 
     # 防止win下无法读取模型
-    if os.name == 'nt':
+    if os.name == "nt":
         python_dir = os.getcwd()
         OPEN_JTALK_DICT_DIR = pyopenjtalk.OPEN_JTALK_DICT_DIR.decode("utf-8")
-        if not (re.match(r'^[A-Za-z0-9_/\\:.\-]*$', OPEN_JTALK_DICT_DIR)):
-            if (OPEN_JTALK_DICT_DIR[:len(python_dir)].upper() == python_dir.upper()):
-                OPEN_JTALK_DICT_DIR = os.path.join(os.path.relpath(OPEN_JTALK_DICT_DIR,python_dir))
+        if not (re.match(r"^[A-Za-z0-9_/\\:.\-]*$", OPEN_JTALK_DICT_DIR)):
+            if OPEN_JTALK_DICT_DIR[: len(python_dir)].upper() == python_dir.upper():
+                OPEN_JTALK_DICT_DIR = os.path.join(os.path.relpath(OPEN_JTALK_DICT_DIR, python_dir))
             else:
                 import shutil
-                if not os.path.exists('TEMP'):
-                    os.mkdir('TEMP')
+
+                if not os.path.exists("TEMP"):
+                    os.mkdir("TEMP")
                 if not os.path.exists(os.path.join("TEMP", "ja")):
                     os.mkdir(os.path.join("TEMP", "ja"))
                 if os.path.exists(os.path.join("TEMP", "ja", "open_jtalk_dic")):
                     shutil.rmtree(os.path.join("TEMP", "ja", "open_jtalk_dic"))
-                shutil.copytree(pyopenjtalk.OPEN_JTALK_DICT_DIR.decode("utf-8"), os.path.join("TEMP", "ja", "open_jtalk_dic"), )
+                shutil.copytree(
+                    pyopenjtalk.OPEN_JTALK_DICT_DIR.decode("utf-8"),
+                    os.path.join("TEMP", "ja", "open_jtalk_dic"),
+                )
                 OPEN_JTALK_DICT_DIR = os.path.join("TEMP", "ja", "open_jtalk_dic")
             pyopenjtalk.OPEN_JTALK_DICT_DIR = OPEN_JTALK_DICT_DIR.encode("utf-8")
 
-        if not (re.match(r'^[A-Za-z0-9_/\\:.\-]*$', current_file_path)):
-            if (current_file_path[:len(python_dir)].upper() == python_dir.upper()):
-                current_file_path = os.path.join(os.path.relpath(current_file_path,python_dir))
+        if not (re.match(r"^[A-Za-z0-9_/\\:.\-]*$", current_file_path)):
+            if current_file_path[: len(python_dir)].upper() == python_dir.upper():
+                current_file_path = os.path.join(os.path.relpath(current_file_path, python_dir))
             else:
-                if not os.path.exists('TEMP'):
-                    os.mkdir('TEMP')
+                if not os.path.exists("TEMP"):
+                    os.mkdir("TEMP")
                 if not os.path.exists(os.path.join("TEMP", "ja")):
                     os.mkdir(os.path.join("TEMP", "ja"))
                 if not os.path.exists(os.path.join("TEMP", "ja", "ja_userdic")):
                     os.mkdir(os.path.join("TEMP", "ja", "ja_userdic"))
-                    shutil.copyfile(os.path.join(current_file_path, "ja_userdic", "userdict.csv"),os.path.join("TEMP", "ja", "ja_userdic", "userdict.csv"))
+                    shutil.copyfile(
+                        os.path.join(current_file_path, "ja_userdic", "userdict.csv"),
+                        os.path.join("TEMP", "ja", "ja_userdic", "userdict.csv"),
+                    )
                 current_file_path = os.path.join("TEMP", "ja")
 
-
     def get_hash(fp: str) -> str:
         hash_md5 = hashlib.md5()
         with open(fp, "rb") as f:
@@ -51,21 +59,26 @@ try:
     USERDIC_HASH_PATH = os.path.join(current_file_path, "ja_userdic", "userdict.md5")
     # 如果没有用户词典，就生成一个；如果有，就检查md5，如果不一样，就重新生成
     if os.path.exists(USERDIC_CSV_PATH):
-        if not os.path.exists(USERDIC_BIN_PATH) or get_hash(USERDIC_CSV_PATH) != open(USERDIC_HASH_PATH, "r",encoding='utf-8').read():
+        if (
+            not os.path.exists(USERDIC_BIN_PATH)
+            or get_hash(USERDIC_CSV_PATH) != open(USERDIC_HASH_PATH, "r", encoding="utf-8").read()
+        ):
             pyopenjtalk.mecab_dict_index(USERDIC_CSV_PATH, USERDIC_BIN_PATH)
-            with open(USERDIC_HASH_PATH, "w", encoding='utf-8') as f:
+            with open(USERDIC_HASH_PATH, "w", encoding="utf-8") as f:
                 f.write(get_hash(USERDIC_CSV_PATH))
 
     if os.path.exists(USERDIC_BIN_PATH):
-        pyopenjtalk.update_global_jtalk_with_user_dict(USERDIC_BIN_PATH)   
-except Exception as e:
+        pyopenjtalk.update_global_jtalk_with_user_dict(USERDIC_BIN_PATH)
+except Exception:
     # print(e)
     import pyopenjtalk
+
     # failed to load user dictionary, ignore.
     pass
 
 
 from text.symbols import punctuation
+
 # 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]"
@@ -123,9 +136,9 @@ def post_replace_ph(ph):
 
 
 def replace_consecutive_punctuation(text):
-    punctuations = ''.join(re.escape(p) for p in punctuation)
-    pattern = f'([{punctuations}])([{punctuations}])+'
-    result = re.sub(pattern, r'\1', text)
+    punctuations = "".join(re.escape(p) for p in punctuation)
+    pattern = f"([{punctuations}])([{punctuations}])+"
+    result = re.sub(pattern, r"\1", text)
     return result
 
 
@@ -152,7 +165,7 @@ def preprocess_jap(text, with_prosody=False):
                 text += p.split(" ")
 
         if i < len(marks):
-            if marks[i] == " ":# 防止意外的UNK
+            if marks[i] == " ":  # 防止意外的UNK
                 continue
             text += [marks[i].replace(" ", "")]
     return text
@@ -165,6 +178,7 @@ def text_normalize(text):
     text = replace_consecutive_punctuation(text)
     return text
 
+
 # Copied from espnet https://github.com/espnet/espnet/blob/master/espnet2/text/phoneme_tokenizer.py
 def pyopenjtalk_g2p_prosody(text, drop_unvoiced_vowels=True):
     """Extract phoneme + prosoody symbol sequence from input full-context labels.
@@ -241,6 +255,7 @@ def pyopenjtalk_g2p_prosody(text, drop_unvoiced_vowels=True):
 
     return phones
 
+
 # Copied from espnet https://github.com/espnet/espnet/blob/master/espnet2/text/phoneme_tokenizer.py
 def _numeric_feature_by_regex(regex, s):
     match = re.search(regex, s)
@@ -248,6 +263,7 @@ def _numeric_feature_by_regex(regex, s):
         return -50
     return int(match.group(1))
 
+
 def g2p(norm_text, with_prosody=True):
     phones = preprocess_jap(norm_text, with_prosody)
     phones = [post_replace_ph(i) for i in phones]

GPT_SoVITS/text/korean.py

@@ -9,39 +9,43 @@ import importlib
 import os
 
 # 防止win下无法读取模型
-if os.name == 'nt':
+if os.name == "nt":
+
     class win_G2p(G2p):
         def check_mecab(self):
             super().check_mecab()
             spam_spec = importlib.util.find_spec("eunjeon")
             non_found = spam_spec is None
             if non_found:
-                print(f'you have to install eunjeon. install it...')
+                print("you have to install eunjeon. install it...")
             else:
                 installpath = spam_spec.submodule_search_locations[0]
-                if not (re.match(r'^[A-Za-z0-9_/\\:.\-]*$', installpath)):
-
+                if not (re.match(r"^[A-Za-z0-9_/\\:.\-]*$", installpath)):
                     import sys
                     from eunjeon import Mecab as _Mecab
+
                     class Mecab(_Mecab):
                         def get_dicpath(installpath):
-                            if not (re.match(r'^[A-Za-z0-9_/\\:.\-]*$', installpath)):
+                            if not (re.match(r"^[A-Za-z0-9_/\\:.\-]*$", installpath)):
                                 import shutil
-                                python_dir = os.getcwd()
-                                if (installpath[:len(python_dir)].upper() == python_dir.upper()):
-                                    dicpath = os.path.join(os.path.relpath(installpath,python_dir),'data','mecabrc')
-                                else:
-                                    if not os.path.exists('TEMP'):
-                                        os.mkdir('TEMP')
-                                    if not os.path.exists(os.path.join('TEMP', 'ko')):
-                                        os.mkdir(os.path.join('TEMP', 'ko'))
-                                    if os.path.exists(os.path.join('TEMP', 'ko', 'ko_dict')):
-                                        shutil.rmtree(os.path.join('TEMP', 'ko', 'ko_dict'))
 
-                                    shutil.copytree(os.path.join(installpath, 'data'), os.path.join('TEMP', 'ko', 'ko_dict'))
-                                    dicpath = os.path.join('TEMP', 'ko', 'ko_dict', 'mecabrc')
+                                python_dir = os.getcwd()
+                                if installpath[: len(python_dir)].upper() == python_dir.upper():
+                                    dicpath = os.path.join(os.path.relpath(installpath, python_dir), "data", "mecabrc")
+                                else:
+                                    if not os.path.exists("TEMP"):
+                                        os.mkdir("TEMP")
+                                    if not os.path.exists(os.path.join("TEMP", "ko")):
+                                        os.mkdir(os.path.join("TEMP", "ko"))
+                                    if os.path.exists(os.path.join("TEMP", "ko", "ko_dict")):
+                                        shutil.rmtree(os.path.join("TEMP", "ko", "ko_dict"))
+
+                                    shutil.copytree(
+                                        os.path.join(installpath, "data"), os.path.join("TEMP", "ko", "ko_dict")
+                                    )
+                                    dicpath = os.path.join("TEMP", "ko", "ko_dict", "mecabrc")
                             else:
-                                dicpath=os.path.abspath(os.path.join(installpath, 'data/mecabrc'))
+                                dicpath = os.path.abspath(os.path.join(installpath, "data/mecabrc"))
                             return dicpath
 
                         def __init__(self, dicpath=get_dicpath(installpath)):
@@ -52,97 +56,108 @@ if os.name == 'nt':
     G2p = win_G2p
 
 
-from text.symbols2 import symbols 
+from text.symbols2 import symbols
 
 # This is a list of Korean classifiers preceded by pure Korean numerals.
-_korean_classifiers = '군데 권 개 그루 닢 대 두 마리 모 모금 뭇 발 발짝 방 번 벌 보루 살 수 술 시 쌈 움큼 정 짝 채 척 첩 축 켤레 톨 통'
+_korean_classifiers = (
+    "군데 권 개 그루 닢 대 두 마리 모 모금 뭇 발 발짝 방 번 벌 보루 살 수 술 시 쌈 움큼 정 짝 채 척 첩 축 켤레 톨 통"
+)
 
 # List of (hangul, hangul divided) pairs:
-_hangul_divided = [(re.compile('%s' % x[0]), x[1]) for x in [
-    # ('ㄳ', 'ㄱㅅ'),   # g2pk2, A Syllable-ending Rule
-    # ('ㄵ', 'ㄴㅈ'),
-    # ('ㄶ', 'ㄴㅎ'),
-    # ('ㄺ', 'ㄹㄱ'),
-    # ('ㄻ', 'ㄹㅁ'),
-    # ('ㄼ', 'ㄹㅂ'),
-    # ('ㄽ', 'ㄹㅅ'),
-    # ('ㄾ', 'ㄹㅌ'),
-    # ('ㄿ', 'ㄹㅍ'),
-    # ('ㅀ', 'ㄹㅎ'),
-    # ('ㅄ', 'ㅂㅅ'),
-    ('ㅘ', 'ㅗㅏ'),
-    ('ㅙ', 'ㅗㅐ'),
-    ('ㅚ', 'ㅗㅣ'),
-    ('ㅝ', 'ㅜㅓ'),
-    ('ㅞ', 'ㅜㅔ'),
-    ('ㅟ', 'ㅜㅣ'),
-    ('ㅢ', 'ㅡㅣ'),
-    ('ㅑ', 'ㅣㅏ'),
-    ('ㅒ', 'ㅣㅐ'),
-    ('ㅕ', 'ㅣㅓ'),
-    ('ㅖ', 'ㅣㅔ'),
-    ('ㅛ', 'ㅣㅗ'),
-    ('ㅠ', 'ㅣㅜ')
-]]
+_hangul_divided = [
+    (re.compile("%s" % x[0]), x[1])
+    for x in [
+        # ('ㄳ', 'ㄱㅅ'),   # g2pk2, A Syllable-ending Rule
+        # ('ㄵ', 'ㄴㅈ'),
+        # ('ㄶ', 'ㄴㅎ'),
+        # ('ㄺ', 'ㄹㄱ'),
+        # ('ㄻ', 'ㄹㅁ'),
+        # ('ㄼ', 'ㄹㅂ'),
+        # ('ㄽ', 'ㄹㅅ'),
+        # ('ㄾ', 'ㄹㅌ'),
+        # ('ㄿ', 'ㄹㅍ'),
+        # ('ㅀ', 'ㄹㅎ'),
+        # ('ㅄ', 'ㅂㅅ'),
+        ("ㅘ", "ㅗㅏ"),
+        ("ㅙ", "ㅗㅐ"),
+        ("ㅚ", "ㅗㅣ"),
+        ("ㅝ", "ㅜㅓ"),
+        ("ㅞ", "ㅜㅔ"),
+        ("ㅟ", "ㅜㅣ"),
+        ("ㅢ", "ㅡㅣ"),
+        ("ㅑ", "ㅣㅏ"),
+        ("ㅒ", "ㅣㅐ"),
+        ("ㅕ", "ㅣㅓ"),
+        ("ㅖ", "ㅣㅔ"),
+        ("ㅛ", "ㅣㅗ"),
+        ("ㅠ", "ㅣㅜ"),
+    ]
+]
 
 # List of (Latin alphabet, hangul) pairs:
-_latin_to_hangul = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [
-    ('a', '에이'),
-    ('b', '비'),
-    ('c', '시'),
-    ('d', '디'),
-    ('e', '이'),
-    ('f', '에프'),
-    ('g', '지'),
-    ('h', '에이치'),
-    ('i', '아이'),
-    ('j', '제이'),
-    ('k', '케이'),
-    ('l', '엘'),
-    ('m', '엠'),
-    ('n', '엔'),
-    ('o', '오'),
-    ('p', '피'),
-    ('q', '큐'),
-    ('r', '아르'),
-    ('s', '에스'),
-    ('t', '티'),
-    ('u', '유'),
-    ('v', '브이'),
-    ('w', '더블유'),
-    ('x', '엑스'),
-    ('y', '와이'),
-    ('z', '제트')
-]]
+_latin_to_hangul = [
+    (re.compile("%s" % x[0], re.IGNORECASE), x[1])
+    for x in [
+        ("a", "에이"),
+        ("b", "비"),
+        ("c", "시"),
+        ("d", "디"),
+        ("e", "이"),
+        ("f", "에프"),
+        ("g", "지"),
+        ("h", "에이치"),
+        ("i", "아이"),
+        ("j", "제이"),
+        ("k", "케이"),
+        ("l", "엘"),
+        ("m", "엠"),
+        ("n", "엔"),
+        ("o", "오"),
+        ("p", "피"),
+        ("q", "큐"),
+        ("r", "아르"),
+        ("s", "에스"),
+        ("t", "티"),
+        ("u", "유"),
+        ("v", "브이"),
+        ("w", "더블유"),
+        ("x", "엑스"),
+        ("y", "와이"),
+        ("z", "제트"),
+    ]
+]
 
 # List of (ipa, lazy ipa) pairs:
-_ipa_to_lazy_ipa = [(re.compile('%s' % x[0], re.IGNORECASE), x[1]) for x in [
-    ('t͡ɕ','ʧ'),
-    ('d͡ʑ','ʥ'),
-    ('ɲ','n^'),
-    ('ɕ','ʃ'),
-    ('ʷ','w'),
-    ('ɭ','l`'),
-    ('ʎ','ɾ'),
-    ('ɣ','ŋ'),
-    ('ɰ','ɯ'),
-    ('ʝ','j'),
-    ('ʌ','ə'),
-    ('ɡ','g'),
-    ('\u031a','#'),
-    ('\u0348','='),
-    ('\u031e',''),
-    ('\u0320',''),
-    ('\u0339','')
-]]
+_ipa_to_lazy_ipa = [
+    (re.compile("%s" % x[0], re.IGNORECASE), x[1])
+    for x in [
+        ("t͡ɕ", "ʧ"),
+        ("d͡ʑ", "ʥ"),
+        ("ɲ", "n^"),
+        ("ɕ", "ʃ"),
+        ("ʷ", "w"),
+        ("ɭ", "l`"),
+        ("ʎ", "ɾ"),
+        ("ɣ", "ŋ"),
+        ("ɰ", "ɯ"),
+        ("ʝ", "j"),
+        ("ʌ", "ə"),
+        ("ɡ", "g"),
+        ("\u031a", "#"),
+        ("\u0348", "="),
+        ("\u031e", ""),
+        ("\u0320", ""),
+        ("\u0339", ""),
+    ]
+]
 
 
 def fix_g2pk2_error(text):
     new_text = ""
     i = 0
     while i < len(text) - 4:
-        if (text[i:i+3] == 'ㅇㅡㄹ' or text[i:i+3] == 'ㄹㅡㄹ') and text[i+3] == ' ' and text[i+4] == 'ㄹ':
-            new_text += text[i:i+3] + ' ' + 'ㄴ'
+        if (text[i : i + 3] == "ㅇㅡㄹ" or text[i : i + 3] == "ㄹㅡㄹ") and text[i + 3] == " " and text[i + 4] == "ㄹ":
+            new_text += text[i : i + 3] + " " + "ㄴ"
             i += 5
         else:
             new_text += text[i]
@@ -166,20 +181,20 @@ def divide_hangul(text):
 
 
 def hangul_number(num, sino=True):
-    '''Reference https://github.com/Kyubyong/g2pK'''
-    num = re.sub(',', '', num)
+    """Reference https://github.com/Kyubyong/g2pK"""
+    num = re.sub(",", "", num)
 
-    if num == '0':
-        return '영'
-    if not sino and num == '20':
-        return '스무'
+    if num == "0":
+        return "영"
+    if not sino and num == "20":
+        return "스무"
 
-    digits = '123456789'
-    names = '일이삼사오육칠팔구'
+    digits = "123456789"
+    names = "일이삼사오육칠팔구"
     digit2name = {d: n for d, n in zip(digits, names)}
 
-    modifiers = '한 두 세 네 다섯 여섯 일곱 여덟 아홉'
-    decimals = '열 스물 서른 마흔 쉰 예순 일흔 여든 아흔'
+    modifiers = "한 두 세 네 다섯 여섯 일곱 여덟 아홉"
+    decimals = "열 스물 서른 마흔 쉰 예순 일흔 여든 아흔"
     digit2mod = {d: mod for d, mod in zip(digits, modifiers.split())}
     digit2dec = {d: dec for d, dec in zip(digits, decimals.split())}
 
@@ -188,75 +203,75 @@ def hangul_number(num, sino=True):
         i = len(num) - i - 1
         if sino:
             if i == 0:
-                name = digit2name.get(digit, '')
+                name = digit2name.get(digit, "")
             elif i == 1:
-                name = digit2name.get(digit, '') + '십'
-                name = name.replace('일십', '십')
+                name = digit2name.get(digit, "") + "십"
+                name = name.replace("일십", "십")
         else:
             if i == 0:
-                name = digit2mod.get(digit, '')
+                name = digit2mod.get(digit, "")
             elif i == 1:
-                name = digit2dec.get(digit, '')
-        if digit == '0':
+                name = digit2dec.get(digit, "")
+        if digit == "0":
             if i % 4 == 0:
-                last_three = spelledout[-min(3, len(spelledout)):]
-                if ''.join(last_three) == '':
-                    spelledout.append('')
+                last_three = spelledout[-min(3, len(spelledout)) :]
+                if "".join(last_three) == "":
+                    spelledout.append("")
                     continue
             else:
-                spelledout.append('')
+                spelledout.append("")
                 continue
         if i == 2:
-            name = digit2name.get(digit, '') + '백'
-            name = name.replace('일백', '백')
+            name = digit2name.get(digit, "") + "백"
+            name = name.replace("일백", "백")
         elif i == 3:
-            name = digit2name.get(digit, '') + '천'
-            name = name.replace('일천', '천')
+            name = digit2name.get(digit, "") + "천"
+            name = name.replace("일천", "천")
         elif i == 4:
-            name = digit2name.get(digit, '') + '만'
-            name = name.replace('일만', '만')
+            name = digit2name.get(digit, "") + "만"
+            name = name.replace("일만", "만")
         elif i == 5:
-            name = digit2name.get(digit, '') + '십'
-            name = name.replace('일십', '십')
+            name = digit2name.get(digit, "") + "십"
+            name = name.replace("일십", "십")
         elif i == 6:
-            name = digit2name.get(digit, '') + '백'
-            name = name.replace('일백', '백')
+            name = digit2name.get(digit, "") + "백"
+            name = name.replace("일백", "백")
         elif i == 7:
-            name = digit2name.get(digit, '') + '천'
-            name = name.replace('일천', '천')
+            name = digit2name.get(digit, "") + "천"
+            name = name.replace("일천", "천")
         elif i == 8:
-            name = digit2name.get(digit, '') + '억'
+            name = digit2name.get(digit, "") + "억"
         elif i == 9:
-            name = digit2name.get(digit, '') + '십'
+            name = digit2name.get(digit, "") + "십"
         elif i == 10:
-            name = digit2name.get(digit, '') + '백'
+            name = digit2name.get(digit, "") + "백"
         elif i == 11:
-            name = digit2name.get(digit, '') + '천'
+            name = digit2name.get(digit, "") + "천"
         elif i == 12:
-            name = digit2name.get(digit, '') + '조'
+            name = digit2name.get(digit, "") + "조"
         elif i == 13:
-            name = digit2name.get(digit, '') + '십'
+            name = digit2name.get(digit, "") + "십"
         elif i == 14:
-            name = digit2name.get(digit, '') + '백'
+            name = digit2name.get(digit, "") + "백"
         elif i == 15:
-            name = digit2name.get(digit, '') + '천'
+            name = digit2name.get(digit, "") + "천"
         spelledout.append(name)
-    return ''.join(elem for elem in spelledout)
+    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))
+    """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}')
+        text = text.replace(f"{num}{classifier}", f"{spelledout}{classifier}")
     # digit by digit for remaining digits
-    digits = '0123456789'
-    names = '영일이삼사오육칠팔구'
+    digits = "0123456789"
+    names = "영일이삼사오육칠팔구"
     for d, n in zip(digits, names):
         text = text.replace(d, n)
     return text
@@ -265,19 +280,23 @@ def number_to_hangul(text):
 def korean_to_lazy_ipa(text):
     text = latin_to_hangul(text)
     text = number_to_hangul(text)
-    text=re.sub('[\uac00-\ud7af]+',lambda x:ko_pron.romanise(x.group(0),'ipa').split('] ~ [')[0],text)
+    text = re.sub("[\uac00-\ud7af]+", lambda x: ko_pron.romanise(x.group(0), "ipa").split("] ~ [")[0], text)
     for regex, replacement in _ipa_to_lazy_ipa:
         text = re.sub(regex, replacement, text)
     return text
 
-_g2p=G2p()
+
+_g2p = G2p()
+
+
 def korean_to_ipa(text):
     text = latin_to_hangul(text)
     text = number_to_hangul(text)
     text = _g2p(text)
     text = fix_g2pk2_error(text)
     text = korean_to_lazy_ipa(text)
-    return text.replace('ʧ','tʃ').replace('ʥ','dʑ')
+    return text.replace("ʧ", "tʃ").replace("ʥ", "dʑ")
+
 
 def post_replace_ph(ph):
     rep_map = {
@@ -301,12 +320,13 @@ def post_replace_ph(ph):
         ph = "停"
     return ph
 
+
 def g2p(text):
     text = latin_to_hangul(text)
     text = _g2p(text)
     text = divide_hangul(text)
     text = fix_g2pk2_error(text)
-    text = re.sub(r'([\u3131-\u3163])$', r'\1.', text)
+    text = re.sub(r"([\u3131-\u3163])$", r"\1.", text)
     # text = "".join([post_replace_ph(i) for i in text])
     text = [post_replace_ph(i) for i in text]
     return text
@@ -314,4 +334,4 @@ def g2p(text):
 
 if __name__ == "__main__":
     text = "안녕하세요"
-    print(g2p(text))
+    print(g2p(text))

GPT_SoVITS/text/symbols.py

@@ -1,5 +1,3 @@
-import os
-
 # punctuation = ['!', '?', '…', ",", ".","@"]#@是SP停顿
 punctuation = ["!", "?", "…", ",", "."]  # @是SP停顿
 punctuation.append("-")

GPT_SoVITS/text/symbols2.py

@@ -1,5 +1,3 @@
-import os
-
 # punctuation = ['!', '?', '…', ",", ".","@"]#@是SP停顿
 punctuation = ["!", "?", "…", ",", "."]  # @是SP停顿
 punctuation.append("-")
@@ -396,24 +394,404 @@ arpa = {
     "SH",
 }
 
-ko_symbols='ㄱㄴㄷㄹㅁㅂㅅㅇㅈㅊㅋㅌㅍㅎㄲㄸㅃㅆㅉㅏㅓㅗㅜㅡㅣㅐㅔ空停'
+ko_symbols = "ㄱㄴㄷㄹㅁㅂㅅㅇㅈㅊㅋㅌㅍㅎㄲㄸㅃㅆㅉㅏㅓㅗㅜㅡㅣㅐㅔ空停"
 # ko_symbols='ㄱㄴㄷㄹㅁㅂㅅㅇㅈㅊㅋㅌㅍㅎㄲㄸㅃㅆㅉㅏㅓㅗㅜㅡㅣㅐㅔ '
 
-yue_symbols={'Yeot3', 'Yip1', 'Yyu3', 'Yeng4', 'Yut5', 'Yaan5', 'Ym5', 'Yaan6', 'Yang1', 'Yun4', 'Yon2', 'Yui5', 'Yun2', 'Yat3', 'Ye', 'Yeot1', 'Yoeng5', 'Yoek2', 'Yam2', 'Yeon6', 'Yu6', 'Yiu3', 'Yaang6', 'Yp5', 'Yai4', 'Yoek4', 'Yit6', 'Yam5', 'Yoeng6', 'Yg1', 'Yk3', 'Yoe4', 'Yam3', 'Yc', 'Yyu4', 'Yyut1', 'Yiu4', 'Ying3', 'Yip3', 'Yaap3', 'Yau3', 'Yan4', 'Yau1', 'Yap4', 'Yk6', 'Yok3', 'Yai1', 'Yeot6', 'Yan2', 'Yoek6', 'Yt1', 'Yoi1', 'Yit5', 'Yn4', 'Yaau3', 'Yau4', 'Yuk6', 'Ys', 'Yuk', 'Yin6', 'Yung6', 'Ya', 'You', 'Yaai5', 'Yau5', 'Yoi3', 'Yaak3', 'Yaat3', 'Ying2', 'Yok5', 'Yeng2', 'Yyut3', 'Yam1', 'Yip5', 'You1', 'Yam6', 'Yaa5', 'Yi6', 'Yek4', 'Yyu2', 'Yuk5', 'Yaam1', 'Yang2', 'Yai', 'Yiu6', 'Yin4', 'Yok4', 'Yot3', 'Yui2', 'Yeoi5', 'Yyun6', 'Yyu5', 'Yoi5', 'Yeot2', 'Yim4', 'Yeoi2', 'Yaan1', 'Yang6', 'Yong1', 'Yaang4', 'Yung5', 'Yeon1', 'Yin2', 'Ya3', 'Yaang3', 'Yg', 'Yk2', 'Yaau5', 'Yut1', 'Yt5', 'Yip4', 'Yung4', 'Yj', 'Yong3', 'Ya1', 'Yg6', 'Yaau6', 'Yit3', 'Yun3', 'Ying1', 'Yn2', 'Yg4', 'Yl', 'Yp3', 'Yn3', 'Yak1', 'Yang5', 'Yoe6', 'You2', 'Yap2', 'Yak2', 'Yt3', 'Yot5', 'Yim2', 'Yi1', 'Yn6', 'Yaat5', 'Yaam3', 'Yoek5', 'Ye3', 'Yeon4', 'Yaa2', 'Yu3', 'Yim6', 'Ym', 'Yoe3', 'Yaai2', 'Ym2', 'Ya6', 'Yeng6', 'Yik4', 'Yot4', 'Yaai4', 'Yyun3', 'Yu1', 'Yoeng1', 'Yaap2', 'Yuk3', 'Yoek3', 'Yeng5', 'Yeoi1', 'Yiu2', 'Yok1', 'Yo1', 'Yoek1', 'Yoeng2', 'Yeon5', 'Yiu1', 'Yoeng4', 'Yuk2', 'Yat4', 'Yg5', 'Yut4', 'Yan6', 'Yin3', 'Yaa6', 'Yap1', 'Yg2', 'Yoe5', 'Yt4', 'Ya5', 'Yo4', 'Yyu1', 'Yak3', 'Yeon2', 'Yong4', 'Ym1', 'Ye2', 'Yaang5', 'Yoi2', 'Yeng3', 'Yn', 'Yyut4', 'Yau', 'Yaak2', 'Yaan4', 'Yek2', 'Yin1', 'Yi5', 'Yoe2', 'Yei5', 'Yaat6', 'Yak5', 'Yp6', 'Yok6', 'Yei2', 'Yaap1', 'Yyut5', 'Yi4', 'Yim1', 'Yk5', 'Ye4', 'Yok2', 'Yaam6', 'Yat2', 'Yon6', 'Yei3', 'Yyu6', 'Yeot5', 'Yk4', 'Yai6', 'Yd', 'Yg3', 'Yei6', 'Yau2', 'Yok', 'Yau6', 'Yung3', 'Yim5', 'Yut6', 'Yit1', 'Yon3', 'Yat1', 'Yaam2', 'Yyut2', 'Yui6', 'Yt2', 'Yek6', 'Yt', 'Ye6', 'Yang3', 'Ying6', 'Yaau1', 'Yeon3', 'Yng', 'Yh', 'Yang4', 'Ying5', 'Yaap6', 'Yoeng3', 'Yyun4', 'You3', 'Yan5', 'Yat5', 'Yot1', 'Yun1', 'Yi3', 'Yaa1', 'Yaap4', 'You6', 'Yaang2', 'Yaap5', 'Yaa3', 'Yaak6', 'Yeng1', 'Yaak1', 'Yo5', 'Yoi4', 'Yam4', 'Yik1', 'Ye1', 'Yai5', 'Yung1', 'Yp2', 'Yui4', 'Yaak4', 'Yung2', 'Yak4', 'Yaat4', 'Yeoi4', 'Yut2', 'Yin5', 'Yaau4', 'Yap6', 'Yb', 'Yaam4', 'Yw', 'Yut3', 'Yong2', 'Yt6', 'Yaai6', 'Yap5', 'Yik5', 'Yun6', 'Yaam5', 'Yun5', 'Yik3', 'Ya2', 'Yyut6', 'Yon4', 'Yk1', 'Yit4', 'Yak6', 'Yaan2', 'Yuk1', 'Yai2', 'Yik2', 'Yaat2', 'Yo3', 'Ykw', 'Yn5', 'Yaa', 'Ye5', 'Yu4', 'Yei1', 'Yai3', 'Yyun5', 'Yip2', 'Yaau2', 'Yiu5', 'Ym4', 'Yeoi6', 'Yk', 'Ym6', 'Yoe1', 'Yeoi3', 'Yon', 'Yuk4', 'Yaai3', 'Yaa4', 'Yot6', 'Yaang1', 'Yei4', 'Yek1', 'Yo', 'Yp', 'Yo6', 'Yp4', 'Yan3', 'Yoi', 'Yap3', 'Yek3', 'Yim3', 'Yz', 'Yot2', 'Yoi6', 'Yit2', 'Yu5', 'Yaan3', 'Yan1', 'Yon5', 'Yp1', 'Yong5', 'Ygw', 'Yak', 'Yat6', 'Ying4', 'Yu2', 'Yf', 'Ya4', 'Yon1', 'You4', 'Yik6', 'Yui1', 'Yaat1', 'Yeot4', 'Yi2', 'Yaai1', 'Yek5', 'Ym3', 'Yong6', 'You5', 'Yyun1', 'Yn1', 'Yo2', 'Yip6', 'Yui3', 'Yaak5', 'Yyun2'}
+yue_symbols = {
+    "Yeot3",
+    "Yip1",
+    "Yyu3",
+    "Yeng4",
+    "Yut5",
+    "Yaan5",
+    "Ym5",
+    "Yaan6",
+    "Yang1",
+    "Yun4",
+    "Yon2",
+    "Yui5",
+    "Yun2",
+    "Yat3",
+    "Ye",
+    "Yeot1",
+    "Yoeng5",
+    "Yoek2",
+    "Yam2",
+    "Yeon6",
+    "Yu6",
+    "Yiu3",
+    "Yaang6",
+    "Yp5",
+    "Yai4",
+    "Yoek4",
+    "Yit6",
+    "Yam5",
+    "Yoeng6",
+    "Yg1",
+    "Yk3",
+    "Yoe4",
+    "Yam3",
+    "Yc",
+    "Yyu4",
+    "Yyut1",
+    "Yiu4",
+    "Ying3",
+    "Yip3",
+    "Yaap3",
+    "Yau3",
+    "Yan4",
+    "Yau1",
+    "Yap4",
+    "Yk6",
+    "Yok3",
+    "Yai1",
+    "Yeot6",
+    "Yan2",
+    "Yoek6",
+    "Yt1",
+    "Yoi1",
+    "Yit5",
+    "Yn4",
+    "Yaau3",
+    "Yau4",
+    "Yuk6",
+    "Ys",
+    "Yuk",
+    "Yin6",
+    "Yung6",
+    "Ya",
+    "You",
+    "Yaai5",
+    "Yau5",
+    "Yoi3",
+    "Yaak3",
+    "Yaat3",
+    "Ying2",
+    "Yok5",
+    "Yeng2",
+    "Yyut3",
+    "Yam1",
+    "Yip5",
+    "You1",
+    "Yam6",
+    "Yaa5",
+    "Yi6",
+    "Yek4",
+    "Yyu2",
+    "Yuk5",
+    "Yaam1",
+    "Yang2",
+    "Yai",
+    "Yiu6",
+    "Yin4",
+    "Yok4",
+    "Yot3",
+    "Yui2",
+    "Yeoi5",
+    "Yyun6",
+    "Yyu5",
+    "Yoi5",
+    "Yeot2",
+    "Yim4",
+    "Yeoi2",
+    "Yaan1",
+    "Yang6",
+    "Yong1",
+    "Yaang4",
+    "Yung5",
+    "Yeon1",
+    "Yin2",
+    "Ya3",
+    "Yaang3",
+    "Yg",
+    "Yk2",
+    "Yaau5",
+    "Yut1",
+    "Yt5",
+    "Yip4",
+    "Yung4",
+    "Yj",
+    "Yong3",
+    "Ya1",
+    "Yg6",
+    "Yaau6",
+    "Yit3",
+    "Yun3",
+    "Ying1",
+    "Yn2",
+    "Yg4",
+    "Yl",
+    "Yp3",
+    "Yn3",
+    "Yak1",
+    "Yang5",
+    "Yoe6",
+    "You2",
+    "Yap2",
+    "Yak2",
+    "Yt3",
+    "Yot5",
+    "Yim2",
+    "Yi1",
+    "Yn6",
+    "Yaat5",
+    "Yaam3",
+    "Yoek5",
+    "Ye3",
+    "Yeon4",
+    "Yaa2",
+    "Yu3",
+    "Yim6",
+    "Ym",
+    "Yoe3",
+    "Yaai2",
+    "Ym2",
+    "Ya6",
+    "Yeng6",
+    "Yik4",
+    "Yot4",
+    "Yaai4",
+    "Yyun3",
+    "Yu1",
+    "Yoeng1",
+    "Yaap2",
+    "Yuk3",
+    "Yoek3",
+    "Yeng5",
+    "Yeoi1",
+    "Yiu2",
+    "Yok1",
+    "Yo1",
+    "Yoek1",
+    "Yoeng2",
+    "Yeon5",
+    "Yiu1",
+    "Yoeng4",
+    "Yuk2",
+    "Yat4",
+    "Yg5",
+    "Yut4",
+    "Yan6",
+    "Yin3",
+    "Yaa6",
+    "Yap1",
+    "Yg2",
+    "Yoe5",
+    "Yt4",
+    "Ya5",
+    "Yo4",
+    "Yyu1",
+    "Yak3",
+    "Yeon2",
+    "Yong4",
+    "Ym1",
+    "Ye2",
+    "Yaang5",
+    "Yoi2",
+    "Yeng3",
+    "Yn",
+    "Yyut4",
+    "Yau",
+    "Yaak2",
+    "Yaan4",
+    "Yek2",
+    "Yin1",
+    "Yi5",
+    "Yoe2",
+    "Yei5",
+    "Yaat6",
+    "Yak5",
+    "Yp6",
+    "Yok6",
+    "Yei2",
+    "Yaap1",
+    "Yyut5",
+    "Yi4",
+    "Yim1",
+    "Yk5",
+    "Ye4",
+    "Yok2",
+    "Yaam6",
+    "Yat2",
+    "Yon6",
+    "Yei3",
+    "Yyu6",
+    "Yeot5",
+    "Yk4",
+    "Yai6",
+    "Yd",
+    "Yg3",
+    "Yei6",
+    "Yau2",
+    "Yok",
+    "Yau6",
+    "Yung3",
+    "Yim5",
+    "Yut6",
+    "Yit1",
+    "Yon3",
+    "Yat1",
+    "Yaam2",
+    "Yyut2",
+    "Yui6",
+    "Yt2",
+    "Yek6",
+    "Yt",
+    "Ye6",
+    "Yang3",
+    "Ying6",
+    "Yaau1",
+    "Yeon3",
+    "Yng",
+    "Yh",
+    "Yang4",
+    "Ying5",
+    "Yaap6",
+    "Yoeng3",
+    "Yyun4",
+    "You3",
+    "Yan5",
+    "Yat5",
+    "Yot1",
+    "Yun1",
+    "Yi3",
+    "Yaa1",
+    "Yaap4",
+    "You6",
+    "Yaang2",
+    "Yaap5",
+    "Yaa3",
+    "Yaak6",
+    "Yeng1",
+    "Yaak1",
+    "Yo5",
+    "Yoi4",
+    "Yam4",
+    "Yik1",
+    "Ye1",
+    "Yai5",
+    "Yung1",
+    "Yp2",
+    "Yui4",
+    "Yaak4",
+    "Yung2",
+    "Yak4",
+    "Yaat4",
+    "Yeoi4",
+    "Yut2",
+    "Yin5",
+    "Yaau4",
+    "Yap6",
+    "Yb",
+    "Yaam4",
+    "Yw",
+    "Yut3",
+    "Yong2",
+    "Yt6",
+    "Yaai6",
+    "Yap5",
+    "Yik5",
+    "Yun6",
+    "Yaam5",
+    "Yun5",
+    "Yik3",
+    "Ya2",
+    "Yyut6",
+    "Yon4",
+    "Yk1",
+    "Yit4",
+    "Yak6",
+    "Yaan2",
+    "Yuk1",
+    "Yai2",
+    "Yik2",
+    "Yaat2",
+    "Yo3",
+    "Ykw",
+    "Yn5",
+    "Yaa",
+    "Ye5",
+    "Yu4",
+    "Yei1",
+    "Yai3",
+    "Yyun5",
+    "Yip2",
+    "Yaau2",
+    "Yiu5",
+    "Ym4",
+    "Yeoi6",
+    "Yk",
+    "Ym6",
+    "Yoe1",
+    "Yeoi3",
+    "Yon",
+    "Yuk4",
+    "Yaai3",
+    "Yaa4",
+    "Yot6",
+    "Yaang1",
+    "Yei4",
+    "Yek1",
+    "Yo",
+    "Yp",
+    "Yo6",
+    "Yp4",
+    "Yan3",
+    "Yoi",
+    "Yap3",
+    "Yek3",
+    "Yim3",
+    "Yz",
+    "Yot2",
+    "Yoi6",
+    "Yit2",
+    "Yu5",
+    "Yaan3",
+    "Yan1",
+    "Yon5",
+    "Yp1",
+    "Yong5",
+    "Ygw",
+    "Yak",
+    "Yat6",
+    "Ying4",
+    "Yu2",
+    "Yf",
+    "Ya4",
+    "Yon1",
+    "You4",
+    "Yik6",
+    "Yui1",
+    "Yaat1",
+    "Yeot4",
+    "Yi2",
+    "Yaai1",
+    "Yek5",
+    "Ym3",
+    "Yong6",
+    "You5",
+    "Yyun1",
+    "Yn1",
+    "Yo2",
+    "Yip6",
+    "Yui3",
+    "Yaak5",
+    "Yyun2",
+}
 
 # symbols = [pad] + c + v + ja_symbols + pu_symbols + list(arpa)+list(ko_symbols)#+list(yue_symbols)###直接这么加yue顺序乱了
 symbols = [pad] + c + v + ja_symbols + pu_symbols + list(arpa)
 symbols = sorted(set(symbols))
 # print(len(symbols))
-symbols+=["[","]"]##日文新增上升下降调型
-symbols+=sorted(list(ko_symbols))
-symbols+=sorted(list(yue_symbols))##新加的yue统一摆在后头#已查过开头加Y后没有重复，韩文显然不会重复
+symbols += ["[", "]"]  ##日文新增上升下降调型
+symbols += sorted(list(ko_symbols))
+symbols += sorted(list(yue_symbols))  ##新加的yue统一摆在后头#已查过开头加Y后没有重复，韩文显然不会重复
 # print(len(symbols))
 if __name__ == "__main__":
     print(len(symbols))
-'''
+"""
 粤语：
     732-353=379
 韩文+粤语：
     732-322=410
-'''
+"""

GPT_SoVITS/text/tone_sandhi.py

@@ -510,12 +510,7 @@ class ToneSandhi:
         # e.g. 走了, 看着, 去过
         elif len(word) == 1 and word in "了着过" and pos in {"ul", "uz", "ug"}:
             finals[-1] = finals[-1][:-1] + "5"
-        elif (
-            len(word) > 1
-            and word[-1] in "们子"
-            and pos in {"r", "n"}
-            and word not in self.must_not_neural_tone_words
-        ):
+        elif len(word) > 1 and word[-1] in "们子" and pos in {"r", "n"} and word not in self.must_not_neural_tone_words:
             finals[-1] = finals[-1][:-1] + "5"
         # e.g. 桌上, 地下, 家里
         elif len(word) > 1 and word[-1] in "上下里" and pos in {"s", "l", "f"}:
@@ -525,25 +520,18 @@ class ToneSandhi:
             finals[-1] = finals[-1][:-1] + "5"
         # 个做量词
         elif (
-            ge_idx >= 1
-            and (word[ge_idx - 1].isnumeric() or word[ge_idx - 1] in "几有两半多各整每做是")
+            ge_idx >= 1 and (word[ge_idx - 1].isnumeric() or word[ge_idx - 1] in "几有两半多各整每做是")
         ) or word == "个":
             finals[ge_idx] = finals[ge_idx][:-1] + "5"
         else:
-            if (
-                word in self.must_neural_tone_words
-                or word[-2:] in self.must_neural_tone_words
-            ):
+            if word in self.must_neural_tone_words or word[-2:] in self.must_neural_tone_words:
                 finals[-1] = finals[-1][:-1] + "5"
 
         word_list = self._split_word(word)
         finals_list = [finals[: len(word_list[0])], finals[len(word_list[0]) :]]
         for i, word in enumerate(word_list):
             # conventional neural in Chinese
-            if (
-                word in self.must_neural_tone_words
-                or word[-2:] in self.must_neural_tone_words
-            ):
+            if word in self.must_neural_tone_words or word[-2:] in self.must_neural_tone_words:
                 finals_list[i][-1] = finals_list[i][-1][:-1] + "5"
         finals = sum(finals_list, [])
         return finals
@@ -561,9 +549,7 @@ class ToneSandhi:
 
     def _yi_sandhi(self, word: str, finals: List[str]) -> List[str]:
         # "一" in number sequences, e.g. 一零零, 二一零
-        if word.find("一") != -1 and all(
-            [item.isnumeric() for item in word if item != "一"]
-        ):
+        if word.find("一") != -1 and all([item.isnumeric() for item in word if item != "一"]):
             return finals
         # "一" between reduplication words shold be yi5, e.g. 看一看
         elif len(word) == 3 and word[1] == "一" and word[0] == word[-1]:
@@ -697,13 +683,10 @@ class ToneSandhi:
         return new_seg
 
     # the first and the second words are all_tone_three
-    def _merge_continuous_three_tones(
-        self, seg: List[Tuple[str, str]]
-    ) -> List[Tuple[str, str]]:
+    def _merge_continuous_three_tones(self, seg: List[Tuple[str, str]]) -> List[Tuple[str, str]]:
         new_seg = []
         sub_finals_list = [
-            lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3)
-            for (word, pos) in seg
+            lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3) for (word, pos) in seg
         ]
         assert len(sub_finals_list) == len(seg)
         merge_last = [False] * len(seg)
@@ -715,10 +698,7 @@ class ToneSandhi:
                 and not merge_last[i - 1]
             ):
                 # if the last word is reduplication, not merge, because reduplication need to be _neural_sandhi
-                if (
-                    not self._is_reduplication(seg[i - 1][0])
-                    and len(seg[i - 1][0]) + len(seg[i][0]) <= 3
-                ):
+                if not self._is_reduplication(seg[i - 1][0]) and len(seg[i - 1][0]) + len(seg[i][0]) <= 3:
                     new_seg[-1][0] = new_seg[-1][0] + seg[i][0]
                     merge_last[i] = True
                 else:
@@ -732,13 +712,10 @@ class ToneSandhi:
         return len(word) == 2 and word[0] == word[1]
 
     # the last char of first word and the first char of second word is tone_three
-    def _merge_continuous_three_tones_2(
-        self, seg: List[Tuple[str, str]]
-    ) -> List[Tuple[str, str]]:
+    def _merge_continuous_three_tones_2(self, seg: List[Tuple[str, str]]) -> List[Tuple[str, str]]:
         new_seg = []
         sub_finals_list = [
-            lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3)
-            for (word, pos) in seg
+            lazy_pinyin(word, neutral_tone_with_five=True, style=Style.FINALS_TONE3) for (word, pos) in seg
         ]
         assert len(sub_finals_list) == len(seg)
         merge_last = [False] * len(seg)
@@ -750,10 +727,7 @@ class ToneSandhi:
                 and not merge_last[i - 1]
             ):
                 # if the last word is reduplication, not merge, because reduplication need to be _neural_sandhi
-                if (
-                    not self._is_reduplication(seg[i - 1][0])
-                    and len(seg[i - 1][0]) + len(seg[i][0]) <= 3
-                ):
+                if not self._is_reduplication(seg[i - 1][0]) and len(seg[i - 1][0]) + len(seg[i][0]) <= 3:
                     new_seg[-1][0] = new_seg[-1][0] + seg[i][0]
                     merge_last[i] = True
                 else:

GPT_SoVITS/text/zh_normalization/chronology.py

@@ -21,25 +21,29 @@ from .num import verbalize_digit
 
 def _time_num2str(num_string: str) -> str:
     """A special case for verbalizing number in time."""
-    result = num2str(num_string.lstrip('0'))
-    if num_string.startswith('0'):
-        result = DIGITS['0'] + result
+    result = num2str(num_string.lstrip("0"))
+    if num_string.startswith("0"):
+        result = DIGITS["0"] + result
     return result
 
 
 # 时刻表达式
-RE_TIME = re.compile(r'([0-1]?[0-9]|2[0-3])'
-                     r':([0-5][0-9])'
-                     r'(:([0-5][0-9]))?')
+RE_TIME = re.compile(
+    r"([0-1]?[0-9]|2[0-3])"
+    r":([0-5][0-9])"
+    r"(:([0-5][0-9]))?"
+)
 
 # 时间范围，如8:30-12:30
-RE_TIME_RANGE = re.compile(r'([0-1]?[0-9]|2[0-3])'
-                           r':([0-5][0-9])'
-                           r'(:([0-5][0-9]))?'
-                           r'(~|-)'
-                           r'([0-1]?[0-9]|2[0-3])'
-                           r':([0-5][0-9])'
-                           r'(:([0-5][0-9]))?')
+RE_TIME_RANGE = re.compile(
+    r"([0-1]?[0-9]|2[0-3])"
+    r":([0-5][0-9])"
+    r"(:([0-5][0-9]))?"
+    r"(~|-)"
+    r"([0-1]?[0-9]|2[0-3])"
+    r":([0-5][0-9])"
+    r"(:([0-5][0-9]))?"
+)
 
 
 def replace_time(match) -> str:
@@ -62,31 +66,33 @@ def replace_time(match) -> str:
         second_2 = match.group(9)
 
     result = f"{num2str(hour)}点"
-    if minute.lstrip('0'):
+    if minute.lstrip("0"):
         if int(minute) == 30:
             result += "半"
         else:
             result += f"{_time_num2str(minute)}分"
-    if second and second.lstrip('0'):
+    if second and second.lstrip("0"):
         result += f"{_time_num2str(second)}秒"
 
     if is_range:
         result += "至"
         result += f"{num2str(hour_2)}点"
-        if minute_2.lstrip('0'):
+        if minute_2.lstrip("0"):
             if int(minute) == 30:
                 result += "半"
             else:
                 result += f"{_time_num2str(minute_2)}分"
-        if second_2 and second_2.lstrip('0'):
+        if second_2 and second_2.lstrip("0"):
             result += f"{_time_num2str(second_2)}秒"
 
     return result
 
 
-RE_DATE = re.compile(r'(\d{4}|\d{2})年'
-                     r'((0?[1-9]|1[0-2])月)?'
-                     r'(((0?[1-9])|((1|2)[0-9])|30|31)([日号]))?')
+RE_DATE = re.compile(
+    r"(\d{4}|\d{2})年"
+    r"((0?[1-9]|1[0-2])月)?"
+    r"(((0?[1-9])|((1|2)[0-9])|30|31)([日号]))?"
+)
 
 
 def replace_date(match) -> str:
@@ -110,8 +116,7 @@ def replace_date(match) -> str:
 
 
 # 用 / 或者 - 分隔的 YY/MM/DD 或者 YY-MM-DD 日期
-RE_DATE2 = re.compile(
-    r'(\d{4})([- /.])(0[1-9]|1[012])\2(0[1-9]|[12][0-9]|3[01])')
+RE_DATE2 = re.compile(r"(\d{4})([- /.])(0[1-9]|1[012])\2(0[1-9]|[12][0-9]|3[01])")
 
 
 def replace_date2(match) -> str:

GPT_SoVITS/text/zh_normalization/constants.py

@@ -18,10 +18,7 @@ from pypinyin.constants import SUPPORT_UCS4
 
 # 全角半角转换
 # 英文字符全角 -> 半角映射表 (num: 52)
-F2H_ASCII_LETTERS = {
-    ord(char) + 65248: ord(char)
-    for char in string.ascii_letters
-}
+F2H_ASCII_LETTERS = {ord(char) + 65248: ord(char) for char in string.ascii_letters}
 
 # 英文字符半角 -> 全角映射表
 H2F_ASCII_LETTERS = {value: key for key, value in F2H_ASCII_LETTERS.items()}
@@ -37,26 +34,29 @@ F2H_PUNCTUATIONS = {ord(char) + 65248: ord(char) for char in string.punctuation}
 H2F_PUNCTUATIONS = {value: key for key, value in F2H_PUNCTUATIONS.items()}
 
 # 空格 (num: 1)
-F2H_SPACE = {'\u3000': ' '}
-H2F_SPACE = {' ': '\u3000'}
+F2H_SPACE = {"\u3000": " "}
+H2F_SPACE = {" ": "\u3000"}
 
 # 非"有拼音的汉字"的字符串，可用于NSW提取
 if SUPPORT_UCS4:
-    RE_NSW = re.compile(r'(?:[^'
-                        r'\u3007'  # 〇
-                        r'\u3400-\u4dbf'  # CJK扩展A:[3400-4DBF]
-                        r'\u4e00-\u9fff'  # CJK基本:[4E00-9FFF]
-                        r'\uf900-\ufaff'  # CJK兼容:[F900-FAFF]
-                        r'\U00020000-\U0002A6DF'  # CJK扩展B:[20000-2A6DF]
-                        r'\U0002A703-\U0002B73F'  # CJK扩展C:[2A700-2B73F]
-                        r'\U0002B740-\U0002B81D'  # CJK扩展D:[2B740-2B81D]
-                        r'\U0002F80A-\U0002FA1F'  # CJK兼容扩展:[2F800-2FA1F]
-                        r'])+')
+    RE_NSW = re.compile(
+        r"(?:[^"
+        r"\u3007"  # 〇
+        r"\u3400-\u4dbf"  # CJK扩展A:[3400-4DBF]
+        r"\u4e00-\u9fff"  # CJK基本:[4E00-9FFF]
+        r"\uf900-\ufaff"  # CJK兼容:[F900-FAFF]
+        r"\U00020000-\U0002A6DF"  # CJK扩展B:[20000-2A6DF]
+        r"\U0002A703-\U0002B73F"  # CJK扩展C:[2A700-2B73F]
+        r"\U0002B740-\U0002B81D"  # CJK扩展D:[2B740-2B81D]
+        r"\U0002F80A-\U0002FA1F"  # CJK兼容扩展:[2F800-2FA1F]
+        r"])+"
+    )
 else:
     RE_NSW = re.compile(  # pragma: no cover
-        r'(?:[^'
-        r'\u3007'  # 〇
-        r'\u3400-\u4dbf'  # CJK扩展A:[3400-4DBF]
-        r'\u4e00-\u9fff'  # CJK基本:[4E00-9FFF]
-        r'\uf900-\ufaff'  # CJK兼容:[F900-FAFF]
-        r'])+')
+        r"(?:[^"
+        r"\u3007"  # 〇
+        r"\u3400-\u4dbf"  # CJK扩展A:[3400-4DBF]
+        r"\u4e00-\u9fff"  # CJK基本:[4E00-9FFF]
+        r"\uf900-\ufaff"  # CJK兼容:[F900-FAFF]
+        r"])+"
+    )

GPT_SoVITS/text/zh_normalization/num.py

@@ -15,23 +15,26 @@
 Rules to verbalize numbers into Chinese characters.
 https://zh.wikipedia.org/wiki/中文数字#現代中文
 """
+
 import re
 from collections import OrderedDict
 from typing import List
 
-DIGITS = {str(i): tran for i, tran in enumerate('零一二三四五六七八九')}
-UNITS = OrderedDict({
-    1: '十',
-    2: '百',
-    3: '千',
-    4: '万',
-    8: '亿',
-})
+DIGITS = {str(i): tran for i, tran in enumerate("零一二三四五六七八九")}
+UNITS = OrderedDict(
+    {
+        1: "十",
+        2: "百",
+        3: "千",
+        4: "万",
+        8: "亿",
+    }
+)
 
-COM_QUANTIFIERS = '(处|台|架|枚|趟|幅|平|方|堵|间|床|株|批|项|例|列|篇|栋|注|亩|封|艘|把|目|套|段|人|所|朵|匹|张|座|回|场|尾|条|个|首|阙|阵|网|炮|顶|丘|棵|只|支|袭|辆|挑|担|颗|壳|窠|曲|墙|群|腔|砣|座|客|贯|扎|捆|刀|令|打|手|罗|坡|山|岭|江|溪|钟|队|单|双|对|出|口|头|脚|板|跳|枝|件|贴|针|线|管|名|位|身|堂|课|本|页|家|户|层|丝|毫|厘|分|钱|两|斤|担|铢|石|钧|锱|忽|(千|毫|微)克|毫|厘|(公)分|分|寸|尺|丈|里|寻|常|铺|程|(千|分|厘|毫|微)米|米|撮|勺|合|升|斗|石|盘|碗|碟|叠|桶|笼|盆|盒|杯|钟|斛|锅|簋|篮|盘|桶|罐|瓶|壶|卮|盏|箩|箱|煲|啖|袋|钵|年|月|日|季|刻|时|周|天|秒|分|小时|旬|纪|岁|世|更|夜|春|夏|秋|冬|代|伏|辈|丸|泡|粒|颗|幢|堆|条|根|支|道|面|片|张|颗|块|元|(亿|千万|百万|万|千|百)|(亿|千万|百万|万|千|百|美|)元|(亿|千万|百万|万|千|百|十|)吨|(亿|千万|百万|万|千|百|)块|角|毛|分)'
+COM_QUANTIFIERS = "(处|台|架|枚|趟|幅|平|方|堵|间|床|株|批|项|例|列|篇|栋|注|亩|封|艘|把|目|套|段|人|所|朵|匹|张|座|回|场|尾|条|个|首|阙|阵|网|炮|顶|丘|棵|只|支|袭|辆|挑|担|颗|壳|窠|曲|墙|群|腔|砣|座|客|贯|扎|捆|刀|令|打|手|罗|坡|山|岭|江|溪|钟|队|单|双|对|出|口|头|脚|板|跳|枝|件|贴|针|线|管|名|位|身|堂|课|本|页|家|户|层|丝|毫|厘|分|钱|两|斤|担|铢|石|钧|锱|忽|(千|毫|微)克|毫|厘|(公)分|分|寸|尺|丈|里|寻|常|铺|程|(千|分|厘|毫|微)米|米|撮|勺|合|升|斗|石|盘|碗|碟|叠|桶|笼|盆|盒|杯|钟|斛|锅|簋|篮|盘|桶|罐|瓶|壶|卮|盏|箩|箱|煲|啖|袋|钵|年|月|日|季|刻|时|周|天|秒|分|小时|旬|纪|岁|世|更|夜|春|夏|秋|冬|代|伏|辈|丸|泡|粒|颗|幢|堆|条|根|支|道|面|片|张|颗|块|元|(亿|千万|百万|万|千|百)|(亿|千万|百万|万|千|百|美|)元|(亿|千万|百万|万|千|百|十|)吨|(亿|千万|百万|万|千|百|)块|角|毛|分)"
 
 # 分数表达式
-RE_FRAC = re.compile(r'(-?)(\d+)/(\d+)')
+RE_FRAC = re.compile(r"(-?)(\d+)/(\d+)")
 
 
 def replace_frac(match) -> str:
@@ -52,7 +55,7 @@ def replace_frac(match) -> str:
 
 
 # 百分数表达式
-RE_PERCENTAGE = re.compile(r'(-?)(\d+(\.\d+)?)%')
+RE_PERCENTAGE = re.compile(r"(-?)(\d+(\.\d+)?)%")
 
 
 def replace_percentage(match) -> str:
@@ -72,7 +75,7 @@ def replace_percentage(match) -> str:
 
 # 整数表达式
 # 带负号的整数 -10
-RE_INTEGER = re.compile(r'(-)' r'(\d+)')
+RE_INTEGER = re.compile(r"(-)" r"(\d+)")
 
 
 def replace_negative_num(match) -> str:
@@ -92,7 +95,7 @@ def replace_negative_num(match) -> str:
 
 # 编号-无符号整形
 # 00078
-RE_DEFAULT_NUM = re.compile(r'\d{3}\d*')
+RE_DEFAULT_NUM = re.compile(r"\d{3}\d*")
 
 
 def replace_default_num(match):
@@ -110,15 +113,11 @@ def replace_default_num(match):
 # RE_ASMD = re.compile(
 #     r'((-?)((\d+)(\.\d+)?)|(\.(\d+)))([\+\-\×÷=])((-?)((\d+)(\.\d+)?)|(\.(\d+)))')
 RE_ASMD = re.compile(
-    r'((-?)((\d+)(\.\d+)?[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|(\.\d+[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|([A-Za-z][⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*))([\+\-\×÷=])((-?)((\d+)(\.\d+)?[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|(\.\d+[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|([A-Za-z][⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*))')
+    r"((-?)((\d+)(\.\d+)?[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|(\.\d+[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|([A-Za-z][⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*))([\+\-\×÷=])((-?)((\d+)(\.\d+)?[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|(\.\d+[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*)|([A-Za-z][⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]*))"
+)
+
+asmd_map = {"+": "加", "-": "减", "×": "乘", "÷": "除", "=": "等于"}
 
-asmd_map = {
-    '+': '加',
-    '-': '减',
-    '×': '乘',
-    '÷': '除',
-    '=': '等于'
-}
 
 def replace_asmd(match) -> str:
     """
@@ -132,24 +131,25 @@ def replace_asmd(match) -> str:
 
 
 # 次方专项
-RE_POWER = re.compile(r'[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]+')
+RE_POWER = re.compile(r"[⁰¹²³⁴⁵⁶⁷⁸⁹ˣʸⁿ]+")
 
 power_map = {
-    '⁰': '0',
-    '¹': '1',
-    '²': '2',
-    '³': '3',
-    '⁴': '4',
-    '⁵': '5',
-    '⁶': '6',
-    '⁷': '7',
-    '⁸': '8',
-    '⁹': '9',
-    'ˣ': 'x',
-    'ʸ': 'y',
-    'ⁿ': 'n'
+    "⁰": "0",
+    "¹": "1",
+    "²": "2",
+    "³": "3",
+    "⁴": "4",
+    "⁵": "5",
+    "⁶": "6",
+    "⁷": "7",
+    "⁸": "8",
+    "⁹": "9",
+    "ˣ": "x",
+    "ʸ": "y",
+    "ⁿ": "n",
 }
 
+
 def replace_power(match) -> str:
     """
     Args:
@@ -166,10 +166,10 @@ def replace_power(match) -> str:
 
 # 数字表达式
 # 纯小数
-RE_DECIMAL_NUM = re.compile(r'(-?)((\d+)(\.\d+))' r'|(\.(\d+))')
+RE_DECIMAL_NUM = re.compile(r"(-?)((\d+)(\.\d+))" r"|(\.(\d+))")
 # 正整数 + 量词
 RE_POSITIVE_QUANTIFIERS = re.compile(r"(\d+)([多余几\+])?" + COM_QUANTIFIERS)
-RE_NUMBER = re.compile(r'(-?)((\d+)(\.\d+)?)' r'|(\.(\d+))')
+RE_NUMBER = re.compile(r"(-?)((\d+)(\.\d+)?)" r"|(\.(\d+))")
 
 
 def replace_positive_quantifier(match) -> str:
@@ -220,7 +220,9 @@ RE_RANGE = re.compile(
     [-~]                   # 匹配范围分隔符
     ((-?)((\d+)(\.\d+)?))  # 匹配范围结束的负数或正数（整数或小数）
     (?![\d\+\-\×÷=])       # 使用正向前瞻以确保数字范围之后没有其他数字和操作符
-    """, re.VERBOSE)
+    """,
+    re.VERBOSE,
+)
 
 
 def replace_range(match) -> str:
@@ -239,7 +241,9 @@ def replace_range(match) -> str:
 
 # ~至表达式
 RE_TO_RANGE = re.compile(
-    r'((-?)((\d+)(\.\d+)?)|(\.(\d+)))(%|°C|℃|度|摄氏度|cm2|cm²|cm3|cm³|cm|db|ds|kg|km|m2|m²|m³|m3|ml|m|mm|s)[~]((-?)((\d+)(\.\d+)?)|(\.(\d+)))(%|°C|℃|度|摄氏度|cm2|cm²|cm3|cm³|cm|db|ds|kg|km|m2|m²|m³|m3|ml|m|mm|s)')
+    r"((-?)((\d+)(\.\d+)?)|(\.(\d+)))(%|°C|℃|度|摄氏度|cm2|cm²|cm3|cm³|cm|db|ds|kg|km|m2|m²|m³|m3|ml|m|mm|s)[~]((-?)((\d+)(\.\d+)?)|(\.(\d+)))(%|°C|℃|度|摄氏度|cm2|cm²|cm3|cm³|cm|db|ds|kg|km|m2|m²|m³|m3|ml|m|mm|s)"
+)
+
 
 def replace_to_range(match) -> str:
     """
@@ -248,71 +252,66 @@ def replace_to_range(match) -> str:
     Returns:
         str
     """
-    result = match.group(0).replace('~', '至')
+    result = match.group(0).replace("~", "至")
     return result
 
 
-def _get_value(value_string: str, use_zero: bool=True) -> List[str]:
-    stripped = value_string.lstrip('0')
+def _get_value(value_string: str, use_zero: bool = True) -> List[str]:
+    stripped = value_string.lstrip("0")
     if len(stripped) == 0:
         return []
     elif len(stripped) == 1:
         if use_zero and len(stripped) < len(value_string):
-            return [DIGITS['0'], DIGITS[stripped]]
+            return [DIGITS["0"], DIGITS[stripped]]
         else:
             return [DIGITS[stripped]]
     else:
-        largest_unit = next(
-            power for power in reversed(UNITS.keys()) if power < len(stripped))
+        largest_unit = next(power for power in reversed(UNITS.keys()) if power < len(stripped))
         first_part = value_string[:-largest_unit]
         second_part = value_string[-largest_unit:]
-        return _get_value(first_part) + [UNITS[largest_unit]] + _get_value(
-            second_part)
+        return _get_value(first_part) + [UNITS[largest_unit]] + _get_value(second_part)
 
 
 def verbalize_cardinal(value_string: str) -> str:
     if not value_string:
-        return ''
+        return ""
 
     # 000 -> '零' , 0 -> '零'
-    value_string = value_string.lstrip('0')
+    value_string = value_string.lstrip("0")
     if len(value_string) == 0:
-        return DIGITS['0']
+        return DIGITS["0"]
 
     result_symbols = _get_value(value_string)
     # verbalized number starting with '一十*' is abbreviated as `十*`
-    if len(result_symbols) >= 2 and result_symbols[0] == DIGITS[
-            '1'] and result_symbols[1] == UNITS[1]:
+    if len(result_symbols) >= 2 and result_symbols[0] == DIGITS["1"] and result_symbols[1] == UNITS[1]:
         result_symbols = result_symbols[1:]
-    return ''.join(result_symbols)
+    return "".join(result_symbols)
 
 
 def verbalize_digit(value_string: str, alt_one=False) -> str:
     result_symbols = [DIGITS[digit] for digit in value_string]
-    result = ''.join(result_symbols)
+    result = "".join(result_symbols)
     if alt_one:
         result = result.replace("一", "幺")
     return result
 
 
 def num2str(value_string: str) -> str:
-    integer_decimal = value_string.split('.')
+    integer_decimal = value_string.split(".")
     if len(integer_decimal) == 1:
         integer = integer_decimal[0]
-        decimal = ''
+        decimal = ""
     elif len(integer_decimal) == 2:
         integer, decimal = integer_decimal
     else:
-        raise ValueError(
-            f"The value string: '${value_string}' has more than one point in it."
-        )
+        raise ValueError(f"The value string: '${value_string}' has more than one point in it.")
 
     result = verbalize_cardinal(integer)
 
-    decimal = decimal.rstrip('0')
+    decimal = decimal.rstrip("0")
     if decimal:
         # '.22' is verbalized as '零点二二'
         # '3.20' is verbalized as '三点二
         result = result if result else "零"
-        result += '点' + verbalize_digit(decimal)
+        result += "点" + verbalize_digit(decimal)
     return result

GPT_SoVITS/text/zh_normalization/phonecode.py

@@ -21,10 +21,8 @@ from .num import verbalize_digit
 # 移动：139、138、137、136、135、134、159、158、157、150、151、152、188、187、182、183、184、178、198
 # 联通：130、131、132、156、155、186、185、176
 # 电信：133、153、189、180、181、177
-RE_MOBILE_PHONE = re.compile(
-    r"(?<!\d)((\+?86 ?)?1([38]\d|5[0-35-9]|7[678]|9[89])\d{8})(?!\d)")
-RE_TELEPHONE = re.compile(
-    r"(?<!\d)((0(10|2[1-3]|[3-9]\d{2})-?)?[1-9]\d{6,7})(?!\d)")
+RE_MOBILE_PHONE = re.compile(r"(?<!\d)((\+?86 ?)?1([38]\d|5[0-35-9]|7[678]|9[89])\d{8})(?!\d)")
+RE_TELEPHONE = re.compile(r"(?<!\d)((0(10|2[1-3]|[3-9]\d{2})-?)?[1-9]\d{6,7})(?!\d)")
 
 # 全国统一的号码400开头
 RE_NATIONAL_UNIFORM_NUMBER = re.compile(r"(400)(-)?\d{3}(-)?\d{4}")
@@ -32,14 +30,12 @@ RE_NATIONAL_UNIFORM_NUMBER = re.compile(r"(400)(-)?\d{3}(-)?\d{4}")
 
 def phone2str(phone_string: str, mobile=True) -> str:
     if mobile:
-        sp_parts = phone_string.strip('+').split()
-        result = '，'.join(
-            [verbalize_digit(part, alt_one=True) for part in sp_parts])
+        sp_parts = phone_string.strip("+").split()
+        result = "，".join([verbalize_digit(part, alt_one=True) for part in sp_parts])
         return result
     else:
-        sil_parts = phone_string.split('-')
-        result = '，'.join(
-            [verbalize_digit(part, alt_one=True) for part in sil_parts])
+        sil_parts = phone_string.split("-")
+        result = "，".join([verbalize_digit(part, alt_one=True) for part in sil_parts])
         return result
 
 

GPT_SoVITS/text/zh_normalization/quantifier.py

@@ -17,7 +17,7 @@ from .num import num2str
 
 # 温度表达式，温度会影响负号的读法
 # -3°C 零下三度
-RE_TEMPERATURE = re.compile(r'(-?)(\d+(\.\d+)?)(°C|℃|度|摄氏度)')
+RE_TEMPERATURE = re.compile(r"(-?)(\d+(\.\d+)?)(°C|℃|度|摄氏度)")
 measure_dict = {
     "cm2": "平方厘米",
     "cm²": "平方厘米",
@@ -35,7 +35,7 @@ measure_dict = {
     "ml": "毫升",
     "m": "米",
     "mm": "毫米",
-    "s": "秒"
+    "s": "秒",
 }
 
 

GPT_SoVITS/text/zh_normalization/text_normlization.py

@@ -56,9 +56,9 @@ from .quantifier import replace_measure
 from .quantifier import replace_temperature
 
 
-class TextNormalizer():
+class TextNormalizer:
     def __init__(self):
-        self.SENTENCE_SPLITOR = re.compile(r'([：、，；。？！,;?!][”’]?)')
+        self.SENTENCE_SPLITOR = re.compile(r"([：、，；。？！,;?!][”’]?)")
 
     def _split(self, text: str, lang="zh") -> List[str]:
         """Split long text into sentences with sentence-splitting punctuations.
@@ -71,66 +71,64 @@ class TextNormalizer():
         if lang == "zh":
             text = text.replace(" ", "")
             # 过滤掉特殊字符
-            text = re.sub(r'[——《》【】<>{}()（）#&@“”^_|\\]', '', text)
-        text = self.SENTENCE_SPLITOR.sub(r'\1\n', text)
+            text = re.sub(r"[——《》【】<>{}()（）#&@“”^_|\\]", "", text)
+        text = self.SENTENCE_SPLITOR.sub(r"\1\n", text)
         text = text.strip()
-        sentences = [sentence.strip() for sentence in re.split(r'\n+', text)]
+        sentences = [sentence.strip() for sentence in re.split(r"\n+", text)]
         return sentences
 
     def _post_replace(self, sentence: str) -> str:
-        sentence = sentence.replace('/', '每')
+        sentence = sentence.replace("/", "每")
         # sentence = sentence.replace('~', '至')
         # sentence = sentence.replace('～', '至')
-        sentence = sentence.replace('①', '一')
-        sentence = sentence.replace('②', '二')
-        sentence = sentence.replace('③', '三')
-        sentence = sentence.replace('④', '四')
-        sentence = sentence.replace('⑤', '五')
-        sentence = sentence.replace('⑥', '六')
-        sentence = sentence.replace('⑦', '七')
-        sentence = sentence.replace('⑧', '八')
-        sentence = sentence.replace('⑨', '九')
-        sentence = sentence.replace('⑩', '十')
-        sentence = sentence.replace('α', '阿尔法')
-        sentence = sentence.replace('β', '贝塔')
-        sentence = sentence.replace('γ', '伽玛').replace('Γ', '伽玛')
-        sentence = sentence.replace('δ', '德尔塔').replace('Δ', '德尔塔')
-        sentence = sentence.replace('ε', '艾普西龙')
-        sentence = sentence.replace('ζ', '捷塔')
-        sentence = sentence.replace('η', '依塔')
-        sentence = sentence.replace('θ', '西塔').replace('Θ', '西塔')
-        sentence = sentence.replace('ι', '艾欧塔')
-        sentence = sentence.replace('κ', '喀帕')
-        sentence = sentence.replace('λ', '拉姆达').replace('Λ', '拉姆达')
-        sentence = sentence.replace('μ', '缪')
-        sentence = sentence.replace('ν', '拗')
-        sentence = sentence.replace('ξ', '克西').replace('Ξ', '克西')
-        sentence = sentence.replace('ο', '欧米克伦')
-        sentence = sentence.replace('π', '派').replace('Π', '派')
-        sentence = sentence.replace('ρ', '肉')
-        sentence = sentence.replace('ς', '西格玛').replace('Σ', '西格玛').replace(
-            'σ', '西格玛')
-        sentence = sentence.replace('τ', '套')
-        sentence = sentence.replace('υ', '宇普西龙')
-        sentence = sentence.replace('φ', '服艾').replace('Φ', '服艾')
-        sentence = sentence.replace('χ', '器')
-        sentence = sentence.replace('ψ', '普赛').replace('Ψ', '普赛')
-        sentence = sentence.replace('ω', '欧米伽').replace('Ω', '欧米伽')
+        sentence = sentence.replace("①", "一")
+        sentence = sentence.replace("②", "二")
+        sentence = sentence.replace("③", "三")
+        sentence = sentence.replace("④", "四")
+        sentence = sentence.replace("⑤", "五")
+        sentence = sentence.replace("⑥", "六")
+        sentence = sentence.replace("⑦", "七")
+        sentence = sentence.replace("⑧", "八")
+        sentence = sentence.replace("⑨", "九")
+        sentence = sentence.replace("⑩", "十")
+        sentence = sentence.replace("α", "阿尔法")
+        sentence = sentence.replace("β", "贝塔")
+        sentence = sentence.replace("γ", "伽玛").replace("Γ", "伽玛")
+        sentence = sentence.replace("δ", "德尔塔").replace("Δ", "德尔塔")
+        sentence = sentence.replace("ε", "艾普西龙")
+        sentence = sentence.replace("ζ", "捷塔")
+        sentence = sentence.replace("η", "依塔")
+        sentence = sentence.replace("θ", "西塔").replace("Θ", "西塔")
+        sentence = sentence.replace("ι", "艾欧塔")
+        sentence = sentence.replace("κ", "喀帕")
+        sentence = sentence.replace("λ", "拉姆达").replace("Λ", "拉姆达")
+        sentence = sentence.replace("μ", "缪")
+        sentence = sentence.replace("ν", "拗")
+        sentence = sentence.replace("ξ", "克西").replace("Ξ", "克西")
+        sentence = sentence.replace("ο", "欧米克伦")
+        sentence = sentence.replace("π", "派").replace("Π", "派")
+        sentence = sentence.replace("ρ", "肉")
+        sentence = sentence.replace("ς", "西格玛").replace("Σ", "西格玛").replace("σ", "西格玛")
+        sentence = sentence.replace("τ", "套")
+        sentence = sentence.replace("υ", "宇普西龙")
+        sentence = sentence.replace("φ", "服艾").replace("Φ", "服艾")
+        sentence = sentence.replace("χ", "器")
+        sentence = sentence.replace("ψ", "普赛").replace("Ψ", "普赛")
+        sentence = sentence.replace("ω", "欧米伽").replace("Ω", "欧米伽")
         # 兜底数学运算，顺便兼容懒人用语
-        sentence = sentence.replace('+', '加')
-        sentence = sentence.replace('-', '减')
-        sentence = sentence.replace('×', '乘')
-        sentence = sentence.replace('÷', '除')
-        sentence = sentence.replace('=', '等')
+        sentence = sentence.replace("+", "加")
+        sentence = sentence.replace("-", "减")
+        sentence = sentence.replace("×", "乘")
+        sentence = sentence.replace("÷", "除")
+        sentence = sentence.replace("=", "等")
         # re filter special characters, have one more character "-" than line 68
-        sentence = re.sub(r'[-——《》【】<=>{}()（）#&@“”^_|\\]', '', sentence)
+        sentence = re.sub(r"[-——《》【】<=>{}()（）#&@“”^_|\\]", "", sentence)
         return sentence
 
     def normalize_sentence(self, sentence: str) -> str:
         # basic character conversions
         sentence = tranditional_to_simplified(sentence)
-        sentence = sentence.translate(F2H_ASCII_LETTERS).translate(
-            F2H_DIGITS).translate(F2H_SPACE)
+        sentence = sentence.translate(F2H_ASCII_LETTERS).translate(F2H_DIGITS).translate(F2H_SPACE)
 
         # number related NSW verbalization
         sentence = RE_DATE.sub(replace_date, sentence)
@@ -161,8 +159,7 @@ class TextNormalizer():
 
         sentence = RE_INTEGER.sub(replace_negative_num, sentence)
         sentence = RE_DECIMAL_NUM.sub(replace_number, sentence)
-        sentence = RE_POSITIVE_QUANTIFIERS.sub(replace_positive_quantifier,
-                                               sentence)
+        sentence = RE_POSITIVE_QUANTIFIERS.sub(replace_positive_quantifier, sentence)
         sentence = RE_DEFAULT_NUM.sub(replace_default_num, sentence)
         sentence = RE_NUMBER.sub(replace_number, sentence)
         sentence = self._post_replace(sentence)