examples/jik_trim/step_1_download_audio.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
from urllib.parse import urlparse

from pathlib import Path

import requests
import pandas as pd

from project_settings import project_path


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--audio_file",
        default="audio.xlsx",
        type=str,
    )
    parser.add_argument(
        "--output_dir",
        default=(project_path / "temp/audio_trim/origin").as_posix(),
        type=str,
    )
    args = parser.parse_args()
    return args


def main():
    args = get_args()

    output_dir = Path(args.output_dir)
    output_dir.mkdir(parents=True, exist_ok=True)

    df = pd.read_excel(args.audio_file)

    for i, row in df.iterrows():
        name = row["name"]
        scene_id = row["scene_id"]
        audio_id = row["audio_id"]
        audio_url = row["audio_url"]

        schema = urlparse(audio_url)
        path = schema.path

        filename = output_dir / path[1:]
        filename.parent.mkdir(parents=True, exist_ok=True)

        resp = requests.get(audio_url)
        with open(filename.as_posix(), "wb") as f:
            f.write(resp.content)
    return


if __name__ == "__main__":
    main()