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

37
tools/slice_audio.py
View File

@@ -1,30 +1,34 @@
import os,sys,numpy as np
import os
import sys
import numpy as np
import traceback
from scipy.io import wavfile
# parent_directory = os.path.dirname(os.path.abspath(__file__))
# sys.path.append(parent_directory)
from tools.my_utils import load_audio
from slicer2 import Slicer
def slice(inp,opt_root,threshold,min_length,min_interval,hop_size,max_sil_kept,_max,alpha,i_part,all_part):
os.makedirs(opt_root,exist_ok=True)
def slice(inp, opt_root, threshold, min_length, min_interval, hop_size, max_sil_kept, _max, alpha, i_part, all_part):
os.makedirs(opt_root, exist_ok=True)
if os.path.isfile(inp):
input=[inp]
input = [inp]
elif os.path.isdir(inp):
input=[os.path.join(inp, name) for name in sorted(list(os.listdir(inp)))]
input = [os.path.join(inp, name) for name in sorted(list(os.listdir(inp)))]
else:
return "输入路径存在但既不是文件也不是文件夹"
slicer = Slicer(
sr=32000, # 长音频采样率
threshold= int(threshold), # 音量小于这个值视作静音的备选切割点
min_length= int(min_length), # 每段最小多长，如果第一段太短一直和后面段连起来直到超过这个值
min_interval= int(min_interval), # 最短切割间隔
hop_size= int(hop_size), # 怎么算音量曲线，越小精度越大计算量越高（不是精度越大效果越好）
max_sil_kept= int(max_sil_kept), # 切完后静音最多留多长
threshold=int(threshold), # 音量小于这个值视作静音的备选切割点
min_length=int(min_length), # 每段最小多长，如果第一段太短一直和后面段连起来直到超过这个值
min_interval=int(min_interval), # 最短切割间隔
hop_size=int(hop_size), # 怎么算音量曲线，越小精度越大计算量越高（不是精度越大效果越好）
max_sil_kept=int(max_sil_kept), # 切完后静音最多留多长
)
_max=float(_max)
alpha=float(alpha)
for inp_path in input[int(i_part)::int(all_part)]:
_max = float(_max)
alpha = float(alpha)
for inp_path in input[int(i_part) :: int(all_part)]:
# print(inp_path)
try:
name = os.path.basename(inp_path)
@@ -32,7 +36,8 @@ def slice(inp,opt_root,threshold,min_length,min_interval,hop_size,max_sil_kept,_
# print(audio.shape)
for chunk, start, end in slicer.slice(audio): # start和end是帧数
tmp_max = np.abs(chunk).max()
if(tmp_max>1):chunk/=tmp_max
if tmp_max > 1:
chunk /= tmp_max
chunk = (chunk / tmp_max * (_max * alpha)) + (1 - alpha) * chunk
wavfile.write(
"%s/%s_%010d_%010d.wav" % (opt_root, name, start, end),
@@ -41,8 +46,8 @@ def slice(inp,opt_root,threshold,min_length,min_interval,hop_size,max_sil_kept,_
(chunk * 32767).astype(np.int16),
)
except:
print(inp_path,"->fail->",traceback.format_exc())
print(inp_path, "->fail->", traceback.format_exc())
return "执行完毕，请检查输出文件"
print(slice(*sys.argv[1:]))
print(slice(*sys.argv[1:]))