#!/usr/bin/env python3

import numpy as np

from scipy.io import wavfile

import os, sys

sys.path.insert(0, '/Users/kaede/tts/GPT-SoVITS')

from tools.slicer2 import Slicer

from tools.my_utils import load_audio

def slice_folder(input_dir, output_dir, sr=32000):

os.makedirs(output_dir, exist_ok=True)

slicer = Slicer(sr=sr, threshold=-40, min_length=3000, min_interval=200, hop_size=20, max_sil_kept=2000)

files = sorted([f for f in os.listdir(input_dir) if f.endswith(('.mp3', '.wav', '.m4b', '.flac'))])

total = 0

for fname in files:

fpath = os.path.join(input_dir, fname)

name = os.path.splitext(fname)[0]

try:

audio = load_audio(fpath, sr)

for chunk, start, end in slicer.slice(audio):

tmp_max = np.abs(chunk).max()

if tmp_max > 1:

chunk /= tmp_max

out_name = f"{name}_{start:010d}_{end:010d}.wav"

wavfile.write(os.path.join(output_dir, out_name), sr, (chunk * 32767).astype(np.int16))

total += 1

print(f" {fname}: OK ({total} total)")

except Exception as e:

print(f" {fname}: FAIL - {e}")

return total

if __name__ == "__main__":

datasets = [

("/Users/kaede/tts/training_data/YouTube-Taiwan", "/Users/kaede/tts/training_data/sliced/YouTube-Taiwan"),

("/Users/kaede/tts/training_data/YouTube-HeroStory", "/Users/kaede/tts/training_data/sliced/YouTube-HeroStory"),

("/Users/kaede/tts/training_data/YouTube-JinYong", "/Users/kaede/tts/training_data/sliced/YouTube-JinYong"),

]

for inp, out in datasets:

if os.path.exists(inp):

n = len([f for f in os.listdir(inp) if f.endswith(('.mp3', '.wav', '.m4b', '.flac'))])

if n > 0:

print(f"Slicing {inp} ({n} files)...")

slice_folder(inp, out)

print(f" -> {out}")

else:

print(f"SKIP {inp}: no audio files")

else:

print(f"SKIP {inp}: not found")