database_conver_script.py

import glob
import os
from pydub import AudioSegment

database_path = "./Taiwan-Tongues-ASR-CE-dataset-zhtw/train"
database_list = []

for i in glob.glob(os.path.join(database_path, "mp3", "*.mp3")):
    filename = os.path.basename(i).split(".")[0]
    filetext = os.path.join(database_path, "txt", filename + ".txt")
    conver_filename = os.path.join(database_path, "wav", filename + ".wav")

    sound = AudioSegment.from_mp3(i)
    sound = sound.set_frame_rate(32000)
    sound = sound.set_channels(1)
    sound = sound.set_sample_width(2)
    sound.export(conver_filename, format="wav")

    filecontent = ""
    with open(filetext, "r") as f:
        for line in f.readlines():
            filecontent += line

    database_list.append(f"{conver_filename}|tw_slicer_opt|ZH|{filecontent}")

with open(os.path.join(database_path, "text.list"), "w", encoding="utf-8") as f:
    f.write("\n".join(database_list))