bacow/GPT-SoVITS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

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

Browse Source 
* ruff check --fix

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

* Change the link for G2PW Model

* update pytorch version and colab
This commit is contained in:
XXXXRT666
2025-04-07 09:42:47 +01:00
committed by GitHub
parent 9da7e17efe
commit 53cac93589
132 changed files with 8185 additions and 6648 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
GPT_SoVITS/text/g2pw/g2pw.py
View File

@@ -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()