Allows:
* The HLS player to propose an "Audio only" resolution
* The live to output an "Audio only" resolution
* The live to ingest and output an "Audio only" stream
This feature is under a config for VOD videos and is enabled by default for lives
In the future we can imagine:
* To propose multiple audio streams for a specific video
* To ingest an audio only VOD and just output an audio only "video"
(the player would play the audio file and PeerTube would not
generate additional resolutions)
This commit introduce a new way to download videos:
* Add "/download/videos/generate/:videoId" endpoint where PeerTube can
mux an audio only and a video only file to a mp4 container
* The download client modal introduces a new default panel where the
user can choose resolutions it wants to download
* Comments and videos can be automatically tagged using core rules or
watched word lists
* These tags can be used to automatically filter videos and comments
* Introduce a new video comment policy where comments must be approved
first
* Comments may have to be approved if the user auto block them using
core rules or watched word lists
* Implement FEP-5624 to federate reply control policies
Breaking: YAML config `ip_view_expiration` is renamed `view_expiration`
Breaking: Views are taken into account after 10 seconds instead of 30
seconds (can be changed in YAML config)
Purpose of this commit is to get closer to other video platforms where
some platforms count views on play (mux, vimeo) or others use a very low
delay (instagram, tiktok)
We also want to improve the viewer identification, where we no longer
use the IP but the `sessionId` generated by the web browser. Multiple
viewers behind a NAT can now be able to be identified as independent
viewers (this method is also used by vimeo or mux)
* testing not removing old file and adding columb to db
* implement feature
* remove unnecessary config changes
* use only keptOriginalFileName, change keptOriginalFileName to keptOriginalFilename for consistency with with videoFile table, slight refactor with basename()
* save original video files to dedicated directory original-video-files
* begin implementing object storage (bucket) support
---------
Co-authored-by: chagai.friedlander <chagai.friedlander@fairkom.eu>
Co-authored-by: Ian <ian.kraft@hotmail.com>
Co-authored-by: Chocobozzz <me@florianbigard.com>
* Split methods in multiple classes
* Add JSONLD tags in embed too
* Index embeds but use a canonical URL tag (targeting the watch page)
* Remote objects don't include a canonical URL tag anymore. Instead we
forbid indexation
* Canonical URLs now use the official short URL (/w/, /w/p, /a, /c
etc.)