more code refactor

GPT_SoVITS/text/cleaner.py

@@ -1,29 +1,27 @@
 from text import chinese, japanese, cleaned_text_to_sequence, symbols, english
 
-language_module_map = {
-    'zh': chinese,
-    "ja": japanese,
-    'en': english
-}
+language_module_map = {"zh": chinese, "ja": japanese, "en": english}
 special = [
-    ('%', 'zh', "SP"),
-    ('￥', 'zh', "SP2"),
-    ('^', 'zh', "SP3"),
+    ("%", "zh", "SP"),
+    ("￥", "zh", "SP2"),
+    ("^", "zh", "SP3"),
     # ('@', 'zh', "SP4")#不搞鬼畜了，和第二版保持一致吧
 ]
+
+
 def clean_text(text, language):
     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)
     language_module = language_module_map[language]
     norm_text = language_module.text_normalize(text)
-    if(language=="zh"):
+    if language == "zh":
         phones, word2ph = language_module.g2p(norm_text)
         assert len(phones) == sum(word2ph)
         assert len(norm_text) == len(word2ph)
     else:
         phones = language_module.g2p(norm_text)
-        word2ph=None
+        word2ph = None
 
     for ph in phones:
         assert ph in symbols
@@ -41,17 +39,17 @@ def clean_special(text, language, special_s, target_symbol):
     new_ph = []
     for ph in phones:
         assert ph in symbols
-        if ph == ',':
+        if ph == ",":
             new_ph.append(target_symbol)
         else:
             new_ph.append(ph)
     return new_ph
 
+
 def text_to_sequence(text, language):
     phones = clean_text(text)
     return cleaned_text_to_sequence(phones)
 
-if __name__ == '__main__':
-    print(clean_text("你好%啊啊啊额、还是到付红四方。", 'zh'))
-
 
+if __name__ == "__main__":
+    print(clean_text("你好%啊啊啊额、还是到付红四方。", "zh"))