VideoDownload

VideoDownload tool for Youtube/BiliBili

Basic Usage
- With source code
- With Docker
Additional Configuration
Celery Mode
Contributing
Thanks

Basic Usage

With source code

Requirements:

Aria2 (for Youtube)
FFmpeg (for Youtube、BiliBili)
Poetry (as package management)
Python 3.8.3+

Setup:

git clone https://github.com/dxsooo/VideoDownload.git
cd VideoDownload

# unix like systems
poetry export -f requirements.txt --without-hashes | pip install -r /dev/stdin

# for windows powershell or cmd
poetry export -f requirements.txt --output requirements.txt --without-hashes
pip install -r requirements.txt

Run:

python download.py -u <VideoURL>

VideoURL is the video playing url and the url pattern should follow:

source	url pattern
Youtube	https://www.youtube.com/watch?v=xxxx
BiliBili	https://www.bilibili.com/video/BVxxxx or http://www.bilibili.com/video/avxxxx

The video is saved in videos/ of the current path and named with video id.

There are more options:

% python download.py --help
usage: download.py [-h] -u URL [-d DIR]

options:
  -h, --help         show this help message and exit
  -u URL, --url URL  video playing url
  -d DIR, --dir DIR  video save directory

With Docker

You can easily download video by docker:

docker run -t -v/path/to/save:/app/videos dxsooo/video-download:0.4.11 download.py -u <VideoURL>

Additional Configuration

As YT-DLP is used in youtube download, a method to config its options is provided. You can add an ini file in configs/youtube.ini with content like:

[ydl_opts]
max_filesize=314572800 ; limit file size, 300M
subtitleslang=en.* ; set subtitles langs to en, default is zh

You can also mount the ini file to docker container for configuration.

Celery Mode

VideoDownload can also work as a Celery worker that receive download tasks, making it possible and convenient to deploy in distribute systems and integrate with cloud-native services.

It is recommended to use by docker:

docker run -d --name video-downloader-1 \
    -e BROKER=${YOUR_CELERY_BROKER} \
    -e BACKEND=${YOUR_CELERY_BACKEND} \
    -v /path/to/save:/app/videos \
    --entrypoint=celery \
    dxsooo/video-download:0.4.11 -A celery_worker worker -c 4 -E

For BiliBili, as some deps could not run with multi process, concurrency(-c) should be 1. But it is ok to run multi docker containers to walk around.

And then you can send task by one of the following methods:

1.with Celery, Python only

# pip install celery
from celery import Celery
celery = Celery(broker="<YOUR_CELERY_BROKER>")
celery.send_task('celery_worker.download',("<VideoURL>",))

2.use REST API from celery flower, language-independent

Start flower by:

docker run -d --name video-downloader-flower \
    -e BROKER=${YOUR_CELERY_BROKER} \
    -e BACKEND=${YOUR_CELERY_BACKEND} \
    --entrypoint=celery \
    -p 5555:5555 \
    dxsooo/video-download:0.4.11 -A celery_worker flower

Example request:

POST /api/task/async-apply/celery_worker.download HTTP/1.1
Accept: application/json
Accept-Encoding: gzip, deflate, compress
Content-Type: application/json; charset=utf-8
Host: localhost:5555

{
    "args": ["<VideoURL>"]
}

if you want to use custom queue instead of the default celery, start worker with -Q <queue>, and send request refer mher/flower#456

Read the source code for more parameters to control task.

Thanks

Bilix
YT-DLP