- Rename variables and function to reflect model names. - Convert infrastructure data (JSON, M3U8) to model types. - Change algorithms to produce/consume `Source` model, in particular using generator functions to build a list of `Source`s rather than the opaque `rendition => variant => urls` mapping (this will make #7 very straight forward). - Download all master playlists after API call before selecting rendition/variants. Motivation for the last point: We use to offer rendition choosing right after the API call, before we download the appropriate master playlist to figure out the available variants. The problem with that is that ArteTV's codes for the renditions (given by the API) do not necessarily include complete languages information (if it is not French or German), for instance a original audio track in Portuguese would show as `VOEU-` (as in "EUropean"). The actual mention of the Portuguese would only show up in the master playlist. So, the new implementation actually downloads all master playlists straight after the API call. This is a bit wasteful, but I figured it was necessary to provide quality interaction with the user. Bonus? Now when we first prompt the user for rendition choice, we actually already know the available variants available, maybe we make use of that fact in the future... |
||
---|---|---|
samples | ||
src/delarte | ||
tests | ||
.gitignore | ||
LICENSE.md | ||
Makefile | ||
pyproject.toml | ||
README.md |
delarte
🎬 ArteTV downloader
💡 What is it ?
This is a toy/research project whose only goal is to familiarize with some of the technologies involved in multi-lingual video streaming. Using this program may violate usage policy of ArteTV website and we do not recommend using it for other purpose then studying the code.
ArteTV is a is a European public service channel dedicated to culture. Available programmes are usually available with multiple audio and subtitles languages.
🚀 Quick start
Install FFMPEG binaries and ensure it is in your PATH
$ ffmpeg -version
ffmpeg version N-109344-g1bebcd43e1-20221202 Copyright (c) 2000-2022 the FFmpeg developers
built with gcc 12.2.0 (crosstool-NG 1.25.0.90_cf9beb1)
Clone this repository
$ git clone https://git.afpy.org/fcode/delarte.git
$ cd delarte
Optionally create a virtual environnement
$ python3 -m venv .venv
$ source .venv/Scripts/activate
Install in edit mode
$ pip install -e .
Or install in edit mode with dev
dependencies if you intend to contribute.
$ pip install -e .[dev]
Now you can run the script
$ python3 -m delarte --help
or
$ delarte --help
ArteTV downloader.
usage: delarte [-h|--help] - print this message
or: delarte program_page_url - show available versions
or: delarte program_page_url version - show available resolutions
or: delarte program_page_url version resolution - download the given video
🔧 How it works
🏗️ The streaming infrastructure
Every video program have a program identifier visible in their web page URL:
https://www.arte.tv/es/videos/110139-000-A/fromental-halevy-la-tempesta/
https://www.arte.tv/fr/videos/100204-001-A/esprit-d-hiver-1-3/
https://www.arte.tv/en/videos/104001-000-A/clint-eastwood/
That program identifier enables us to query an API for the program's information.
The config API
For the last example the API call is as such:
https://api.arte.tv/api/player/v2/config/en/104001-000-A
The response is a JSON object:
{
"data": {
"id": "104001-000-A_en",
"type": "ConfigPlayer",
"attributes": {
"metadata": {
"providerId": "104001-000-A",
"language": "en",
"title": "Clint Eastwood",
"subtitle": "The Last Legend",
"description": "70 years of career in front of and behind the camera and still active at 90, Clint Eastwood is a Hollywood legend. A look back at his unique career through a portrait that explores the complexity of the Eastwood myth.",
"duration": { "seconds": 4652 },
...
},
"streams": [
{
"url": "https://.../104001-000-A_VOF-STE%5BANG%5D_XQ.m3u8",
"versions": [
{
"label": "English (Subtitles)",
"shortLabel": "OGsub-ANG",
"eStat": {
"ml5": "VOF-STE[ANG]"
}
}
],
...
},
{
"url": "https://.../104001-000-A_VOF-STF_XQ.m3u8",
"versions": [
{
"label": "French (Original)",
"shortLabel": "FR",
"eStat": {
"ml5": "VOF-STF"
}
}
],
...
},
{
"url": "https://.../104001-000-A_VOF-STMF_XQ.m3u8",
"versions": [
{
"label": "Original french version - closed captioning (FR)",
"shortLabel": "ccFR",
"eStat": {
"ml5": "VOF-STMF"
}
}
],
...
},
{
"url": "https://.../104001-000-A_VA-STA_XQ.m3u8",
"versions": [
{
"label": "German (Dubbed)",
"shortLabel": "DE",
"eStat": {
"ml5": "VA-STA"
}
}
],
...
},
{
"url": "https://.../104001-000-A_VA-STMA_XQ.m3u8",
"versions": [
{
"label": "German closed captioning ",
"shortLabel": "ccDE",
"eStat": {
"ml5": "VA-STMA"
}
}
],
...
}
],
...
}
}
}
Information about the program is detailed in data.attributes.metadata
and a list of available audio/subtitles combinations in data.attributes.streams
. In our code such a combination is referred to as a rendition (or version in the CLI).
Every such rendition has a reference to a master playlist file in .streams[i].url
and description of the audio/subtitle combination in .streams[i].versions[0]
.
We are using .streams[i].versions[0].eStat.ml5
as our rendition key:
VOF-STE[ANG]
English (Subtitles)VOF-STF
French (Original)VOF-STMF
Original french version - closed captioning (FR)VA-STA
German (Dubbed)VA-STMA
German closed captioning- ...
The master playlist
As defined in HTTP Live Streaming, for example:
#EXTM3U
...
#EXT-X-STREAM-INF:BANDWIDTH=2335200,AVERAGE-BANDWIDTH=1123304,VIDEO-RANGE=SDR,CODECS="avc1.4d401e,mp4a.40.2",RESOLUTION=768x432,FRAME-RATE=25.000,AUDIO="program_audio_0",SUBTITLES="subs"
medias/104001-000-A_v432.m3u8
#EXT-X-STREAM-INF:BANDWIDTH=4534432,AVERAGE-BANDWIDTH=2124680,VIDEO-RANGE=SDR,CODECS="avc1.4d0028,mp4a.40.2",RESOLUTION=1920x1080,FRAME-RATE=25.000,AUDIO="program_audio_0",SUBTITLES="subs"
medias/104001-000-A_v1080.m3u8
#EXT-X-STREAM-INF:BANDWIDTH=4153392,AVERAGE-BANDWIDTH=1917840,VIDEO-RANGE=SDR,CODECS="avc1.4d401f,mp4a.40.2",RESOLUTION=1280x720,FRAME-RATE=25.000,AUDIO="program_audio_0",SUBTITLES="subs"
medias/104001-000-A_v720.m3u8
#EXT-X-STREAM-INF:BANDWIDTH=1445432,AVERAGE-BANDWIDTH=726160,VIDEO-RANGE=SDR,CODECS="avc1.4d401e,mp4a.40.2",RESOLUTION=640x360,FRAME-RATE=25.000,AUDIO="program_audio_0",SUBTITLES="subs"
medias/104001-000-A_v360.m3u8
#EXT-X-STREAM-INF:BANDWIDTH=815120,AVERAGE-BANDWIDTH=429104,VIDEO-RANGE=SDR,CODECS="avc1.42e00d,mp4a.40.2",RESOLUTION=384x216,FRAME-RATE=25.000,AUDIO="program_audio_0",SUBTITLES="subs"
medias/104001-000-A_v216.m3u8
...
#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="program_audio_0",LANGUAGE="fr",NAME="VOF",AUTOSELECT=YES,DEFAULT=YES,URI="medias/104001-000-A_aud_VOF.m3u8"
#EXT-X-MEDIA:TYPE=SUBTITLES,GROUP-ID="subs",NAME="English",DEFAULT=YES,AUTOSELECT=YES,FORCED=NO,LANGUAGE="en",URI="medias/104001-000-A_st_VO-ANG.m3u8"
...
This file show the a list of video variants URIs (one per video resolution). Each of them has
- exactly one video media playlist reference
- exactly one audio media playlist reference
- at most one subtitles media playlist reference
The video and audio media playlist
As defined in HTTP Live Streaming, for example:
#EXTM3U
#EXT-X-TARGETDURATION:6
#EXT-X-VERSION:7
#EXT-X-MEDIA-SEQUENCE:1
#EXT-X-INDEPENDENT-SEGMENTS
#EXT-X-PLAYLIST-TYPE:VOD
#EXT-X-MAP:URI="104001-000-A_v1080.mp4",BYTERANGE="28792@0"
#EXTINF:6.000,
#EXT-X-BYTERANGE:1734621@28792
104001-000-A_v1080.mp4
#EXTINF:6.000,
#EXT-X-BYTERANGE:1575303@1763413
104001-000-A_v1080.mp4
#EXTINF:6.000,
#EXT-X-BYTERANGE:1603739@3338716
104001-000-A_v1080.mp4
#EXTINF:6.000,
#EXT-X-BYTERANGE:1333835@4942455
104001-000-A_v1080.mp4
...
This file shows the list of segments the server expect to serve.
The subtitles media playlist
As defined in HTTP Live Streaming, for example:
#EXTM3U
#EXT-X-VERSION:7
#EXT-X-TARGETDURATION:4650
#EXT-X-MEDIA-SEQUENCE:1
#EXT-X-PLAYLIST-TYPE:VOD
#EXTINF:4650,
104001-000-A_st_VO-ANG.vtt
#EXT-X-ENDLIST
This file shows the file containing the subtitles data.
⚙️The process
- Get the config API object for the program identifier.
- Select a rendition.
- Get the master playlist.
- Select a variant.
- Download audio, video and subtitles media content.
- convert
VTT
subtitles toSRT
- convert
- Figure out the output filename from metadata.
- Feed the all the media to
ffmpeg
for muxing
📽️ FFMPEG
The multiplexing (muxing) the video file is handled by ffmpeg. The script expects ffmpeg to be installed in the environnement and will call it as a subprocess.
Why not use FFMPEG direcly with the HLS master playlist URL ?
So we can be more granular about renditions and variants that we want.
Why not use VTT
subtitles direcly ?
Because it fails 😒.
Why not use FFMPEG direcly with the media playalist URLs and let it do the download ?
Because some programs would randomly fail 😒. Probably due to invalid segmentation on the server.
📌 Dependences
🤝 Help
For sure ! The more the merrier.