Refactor: moved text/ to style_bert_vits2/text_processing/(language)/

app.py

@@ -27,8 +27,8 @@ from style_bert_vits2.constants import (
 )
 from style_bert_vits2.logging import logger
 from style_bert_vits2.models.infer import InvalidToneError
+from style_bert_vits2.text_processing.japanese import normalize_text
 from style_bert_vits2.text_processing.japanese.g2p_utils import g2kata_tone, kata_tone2phone_tone
-from style_bert_vits2.text_processing.japanese.normalizer import normalize_text
 
 
 # Get path settings

server_editor.py

@@ -42,8 +42,8 @@ from style_bert_vits2.constants import (
 )
 from style_bert_vits2.logging import logger
 from style_bert_vits2.text_processing import bert_models
+from style_bert_vits2.text_processing.japanese import normalize_text
 from style_bert_vits2.text_processing.japanese.g2p_utils import g2kata_tone, kata_tone2phone_tone
-from style_bert_vits2.text_processing.japanese.normalizer import normalize_text
 from style_bert_vits2.text_processing.japanese.user_dict import (
     apply_word,
     update_dict,

style_bert_vits2/text_processing/__init__.py

@@ -67,18 +67,17 @@ def clean_text(
 
     # Changed to import inside if condition to avoid unnecessary import
     if language == Languages.JP:
-        from style_bert_vits2.text_processing.japanese.g2p import g2p
-        from style_bert_vits2.text_processing.japanese.normalizer import normalize_text
+        from style_bert_vits2.text_processing.japanese import g2p, normalize_text
         norm_text = normalize_text(text)
         phones, tones, word2ph = g2p(norm_text, use_jp_extra, raise_yomi_error)
     elif language == Languages.EN:
-        from ...text import english as language_module
-        norm_text = language_module.normalize_text(text)
-        phones, tones, word2ph = language_module.g2p(norm_text)
+        from style_bert_vits2.text_processing.english import g2p, normalize_text
+        norm_text = normalize_text(text)
+        phones, tones, word2ph = g2p(norm_text)
     elif language == Languages.ZH:
-        from ...text import chinese as language_module
-        norm_text = language_module.normalize_text(text)
-        phones, tones, word2ph = language_module.g2p(norm_text)
+        from style_bert_vits2.text_processing.chinese import g2p, normalize_text
+        norm_text = normalize_text(text)
+        phones, tones, word2ph = g2p(norm_text)
     else:
         raise ValueError(f"Language {language} not supported")
 

style_bert_vits2/text_processing/chinese/__init__.py

@@ -66,7 +66,7 @@ def replace_punctuation(text):
     return replaced_text
 
 
-def g2p(text):
+def g2p(text: str) -> tuple[list[str], list[int], list[int]]:
     pattern = r"(?<=[{0}])\s*".format("".join(PUNCTUATIONS))
     sentences = [i for i in re.split(pattern, text) if i.strip() != ""]
     phones, tones, word2ph = _g2p(sentences)
@@ -168,7 +168,7 @@ def _g2p(segments):
     return phones_list, tones_list, word2ph
 
 
-def normalize_text(text):
+def normalize_text(text: str) -> str:
     numbers = re.findall(r"\d+(?:\.?\d+)?", text)
     for number in numbers:
         text = text.replace(number, cn2an.an2cn(number), 1)

style_bert_vits2/text_processing/english/__init__.py

@@ -369,7 +369,7 @@ def normalize_numbers(text):
     return text
 
 
-def normalize_text(text):
+def normalize_text(text: str) -> str:
     text = normalize_numbers(text)
     text = replace_punctuation(text)
     text = re.sub(r"([,;.\?\!])([\w])", r"\1 \2", text)
@@ -419,7 +419,7 @@ def text_to_words(text):
     return words
 
 
-def g2p(text):
+def g2p(text: str) -> tuple[list[str], list[int], list[int]]:
     phones = []
     tones = []
     phone_len = []

style_bert_vits2/text_processing/japanese/__init__.py

@@ -0,0 +1,2 @@
+from style_bert_vits2.text_processing.japanese.g2p import g2p  # type: ignore
+from style_bert_vits2.text_processing.japanese.normalizer import normalize_text  # type: ignore