Future Audio Work

[Drapersniper]

This issue is to gather features to be suggested for or worked on for Audio this supersedes and replaces #2494 due to the changes added by #2904

Open Issues:

• [ ] Lavalink status (#1811)
• [ ] Document Lavalink setup with multiple instances (#2843) (This could also send a message to the bot owner)

TODO:

• [ ] Queue: move songs up and down the list by number of songs (#621)
• [ ] Break Search command into multiple subcommands
• [ ] Wastebin lists of links to tracks
• [ ] Remove multiple songs at once
• [ ] Make Queue aware when enqueuing playlists and/or albums
• [ ] Add more dispatches
• [ ] Add more public APIs to allow for more Audio add-ons

Requests:

• [ ] Make playlist remember where it was stopped? (#1544)
• [ ] Google Play Music Support (https://github.com/MarshallOfSound/Google-Play-Music-Desktop-Player-UNOFFICIAL-/blob/master/docs/PlaybackAPI_WebSocket.md)
• [ ] SFX support
• [ ] [p]prev Can go back multiple songs > This will be easier if #366 is addressed
• [ ] Play multiple songs at once [p]play 1|2|3 for example
• [ ] Local: M3U/M3U8/PLS/XSPF/WPL/ASX/iTunes Playlists support (#620) (make custom parser or contribute to Lavaplayer/Lavalink)
• [ ] Parse tracks out of a M3U and PLS url
• [ ] Repeat the whole queue
• [ ] [p]play and [p]playlist append commands to bulk enque multiple songs

  Maybe:

• [ ] A thumbsup/thumbsdown on a np message to make a song play more frequently on a playlist in tandem with the playlist checking on #1544, + skip stats (playlist playing stat saving/management)
• [ ] Lyrics support
• [ ] Expand Spotify playlist support to use SoundCloud API
• [ ] Jukebox price to be charged per song instead of per command?
• [ ] Pillow Audio 😝
• [ ] Add RPC functionality; information on doing this at #3741
• [ ] add [p]makeplaylist that enqueue n number of recently played tracks in the server
• [ ] add [p]playstric <author> || <track name> which would allow using Global Cache and JDA cache directly.
• [ ] Add special casing for Live stream playlists

Next Todo

• [ ] Autojoin/bind (a command to bind the bot to a text channel and a voice channel for the duration of the audio session), also setting a lock on the audio set notify channel via this method, pinned queue
• [ ] Configurable limit on max concurrent audio streams (#272)
• [ ] More Spotify functionality using their API
  • Search for Artists, Playlists and Albums
  • Use Song Metadata for info [ ] Spotify Search command
• [ ] Absolute Vote toggle (Voting system that DJ's can't bypass)
• [ ] Auto Join a music channel

In Edge

• [x] [p]audioset role To accept a list of roles
• [x] Set maximum allowed volume
• [x] Effects
• [x] Channel specific volume settings
• [x] Configurable maximum queue size - Places where the Queue is Controllate already been added, now need to add a command to set it. (Note to self, Hard coded > Bot Owner > Server Mods)
• [x] Add a [p]mixplaylist <url> command based on the newly discovered behaviour of youtube mix playlists - https://github.com/Cog-Creators/Red-DiscordBot/issues/2940#issuecomment-583879466

Done

• [x] Persist queues on reboot/reload Done in #3242 but want to make it avaliable as part of #3459

What about sfx?

• sfx can be found (unapproved) at: https://github.com/baiumbg/baiumbg-Cogs . Use at your own risk.

Want to suggest something?

• Just leave a comment below!

[aikaterna]

Please keep arguments off this topic. Thanks.

[Ryonez]

Given the size of queues and potential risk from it I will not remove the cooldown here unless there's a lot of push back. If there is then I'm removing the 3 scopes and making playlist only creatable by bot owners

Could you please go over the potential risks? We could also have it so the bot owner gets to bypass the cooldown with a prompt maybe.

Edit: Response in Discord:

in short, your bot can get ddosed and/or you could possibly corrupt your playlist files

[aikaterna]

I have already commented on this in support in the main server.

[Mathizsias]

Songs when played with !local start are not pre-shuffled and always appear to start with the same song from the folder. Even though the bot is configured to !shuffle. The same !shuffle works fine for !play.

[fixator10]

There is some probability that lavalink WS shutdown happens indefinitely, unless SIGKILL happens.

Approximate reproduce steps:

1. Run Red with audio with some usage for some time (probably let it play on some servers)
2. Run `await bot.shutdown(restart=True)`
3. [V3/edge]: Red will stop on `[INFO] red.Audio.WS.LLServer: Shutdown Lavalink WS.`
4. ^ that will continue indefinitely until SIGKILL is given:
   [2021-04-04 00:39:53] [INFO] red.Audio.WS.LLServer: Shutdown Lavalink WS.
   [2021-04-04 00:40:44] [INFO] red.main: Signals.SIGTERM received. Quitting...
   [2021-04-04 00:40:44] [INFO] red.main: Shutting down with exit code: ExitCodes.SHUTDOWN
   [2021-04-04 00:40:44] [INFO] red.main: Please wait, cleaning up a bit more
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: State 'stop-sigterm' timed out. Killing.
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Killing process 7935 (python3.8) with signal SIGKILL.
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Main process exited, code=killed, status=9/KILL
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Failed with result 'timeout'.

[Drapersniper]

There is some probability that lavalink WS shutdown happens indefinitely, unless SIGKILL happens.

Approximate reproduce steps:

1. Run Red with audio with some usage for some time (probably let it play on some servers)
2. Run `await bot.shutdown(restart=True)`
3. [V3/edge]: Red will stop on `[INFO] red.Audio.WS.LLServer: Shutdown Lavalink WS.`
4. ^ that will continue indefinitely until SIGKILL is given:
   [2021-04-04 00:39:53] [INFO] red.Audio.WS.LLServer: Shutdown Lavalink WS.
   [2021-04-04 00:40:44] [INFO] red.main: Signals.SIGTERM received. Quitting...
   [2021-04-04 00:40:44] [INFO] red.main: Shutting down with exit code: ExitCodes.SHUTDOWN
   [2021-04-04 00:40:44] [INFO] red.main: Please wait, cleaning up a bit more
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: State 'stop-sigterm' timed out. Killing.
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Killing process 7935 (python3.8) with signal SIGKILL.
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Main process exited, code=killed, status=9/KILL
   Apr 04 04:40:54 systemd[1]: kurisuv3.service: Failed with result 'timeout'.

I meant here https://github.com/Cog-Creators/Red-DiscordBot/pull/4565 Fixator, but its fine. i just tested edge and no longer reproducing if the issue persist after the latest update give me a shout on the PR.

[aikaterna]

Just adding a link back to https://github.com/Cog-Creators/Red-DiscordBot/issues/3492 here as I saw it wasn't linked. I ran into this myself today where I missed the loading message where audio was loading in Spotify tracks.

• play a spotify playlist
• bot is still enqueing
• run [p]stop
• bot stops playback of current song, keeps feeding in converted songs as it keeps going

[Technetium1]

It would be nice to be able to delete large amounts of tracks from the playlist. For example 40-50,52 or 40-50|52. Might be better to use the bar to keep format with the current request for playing multiple tracks, but it feels most natural to use a comma for both.

[Drapersniper]

Note to self: You fucking moron yes its jacks fault but you should make better usage on destination in the arg parse logic ... fix yo shit.

[ltzmax]

Add a possible way to toggle notify to a channel, example [p]audioset notify #testing.

[aikaterna]

Add a possible way to toggle notify to a channel, example [p]audioset notify #testing.

This is already listed in the main description of this topic as "setting a lock on the audio set notify channel via this method".

[Drapersniper]

Closing, this one is becoming ungodly; if i missed anything from #5075 feel free to bring it over there.