benrucker / jermabot Goto Github PK

possibly with embed magic? maybe have a second page to list the sound files for play?

in order to improve cross-platform compatibility and ease of installation, the speak logic should be pulled into a robust voice interface. the interface will choose the best tts library to interface with on startup.

• Windows: sapi5
• linux: mycroft mimic if found
  • else, espeak 🤕

if none are found, it will throw a detailed exception at launch and/or at attempted invocation

and cave noises

no repro steps yet

turn on debug to find out why. repro by sitting in a channel with him for a while

the fix for this might be huge... we need to find a way to disable join sounds when a snap command is invoked, but reenable them once the snap stuff is over.

maybe we should make a new data structure to store specific info about a guild, such as if a snap is currently happening. this way, info can be tracked per guild and not affect other guilds.

maybe make a GuildInfo class that tracks state info using properties (is that the right term?) for a single guild and have a global list of GuildInfo objects.

not like another snap. instead this command should have the option to move everyone in a certain voice channel to another voice channel (eg $move #voicetarget #voicedest) OR move specific people to another channel (eg $move @el jefe$1234 @benjamin#5678 #voicedest). unsure if the syntax should force a notification for the people moved or if it should take in name fragments (ie $move jefe, ben #voicedest would function the same as the previous parenthetical)

Functionally identical to command play but constructs a looping source to keep the audio going. Might want to restrict to admins

Invoked when command volume is sent and the bot responds with the volume. Should update message to reflect change after up or down is hit.

This change also means that an argumentless call of volume should simply tell the user what the current volume is

Might want to rely on this issue being completed first: #6

Is your feature request related to a problem? Please describe.
i just DMd notsobot and half-expected for it to reply with a joke

Describe the solution you'd like
jerma should do this

Additional context
sigma

this way, instead of relying on the hacky way of just naming the file after someone, the join sounds will be separate from the rest of the sounds. Maybe the sounds should still be listed by $list and played with $play (with the username/nickname of the user). If this is combined with adding sounds through commands, the sounds can be named with user IDs instead of names, thus it can't be broken by a user changing their username and can be divided into guilds by folder.

allow start times through default youtube url handling. allow end times if reasonable to implement

Is your feature request related to a problem? Please describe.
jerma only fully works in whid. even my own testing server doesn't fully work with jermabot

Describe the solution you'd like
replace permissions logic that relies on specific whid roles with logic that checks a user for specific privileges (e.g. manage users)

Should then also make a way to set another guild as a 'parent' or something to gain their sounds and join sounds. That way, debugging will be easier in test servers.

bot leaves when a single person leaves

/play autofills possible sounds

if the leave sound fails to play, disconnect anyway

jerma was recently moved to a slower hdd; spin up time could be causing the delayed

solution: cache join sounds

possible implementation could get a list of members from connected guilds and preload any matching sound files.

a smarter idea might be to separate join sounds from regular ones, name them by ID, and add a user interface for them. otoh, this may cause confusion if join sounds are no longer playable by using someone's name. computational complexity will balloon if this is implemented poorly.

check for https://discordpy.readthedocs.io/en/latest/api.html#discord.VoiceState.afk

currently peeking at this for inspo.
expected functionality:

• $help
  • lists commands, command syntax, command descriptions
  • has a 'next page' button to view the sound list
  • has a 'previous page' button to go back to the help page
• $list
  • lists all available sounds
  • is a dm

• #56
  • steal the default discord.py HelpCommand logic for grabbing all the commands
  • rewrite the formatting and sending logic to use embeds and reactions
• $help <category>
  • splitting into separate issue #35
• add streaming notification thing
  • maybe incorporate into activity?
• add jerma joke command
  • download each joke and cut to length
  • OR get start time, end time, and link: store them locally and stream
• make jerma disconnect from voice when nobody is in the channel with him
  • this idea unashamedly stolen from hackerbot
• add commands to set your own join sound
  • add support for choosing an existing sound by name
  • add support for uploading a new sound file
  • make it so an admin has to approve of the change
• unify console output format
• add more colors to console output because colors are cool
• from Bertle: make dictionary of jerma word soundclips so he can actually speak in $speak

might be due to quitting the program ungracefully

before sweeping refactors have to happen, start off by bodging this into the system:
https://github.com/DXsmiley/dismock

after that is mostly working, the separation of logic and discord function calls should be separated. Thus, tests can be made that don't rely at all on the discord API

save the ears at night

when done uncomment out lines 64 & 65 in the jermahelp function and remove the other two.
also uncomment line 5 in help.py

Need to haves:

• Move to discord-supported interactions
  • see #59
• Open support server

Nice to haves:

• Pull functionality into idempotent functions
• Add tests 😓
• Migrate to a DB model
  • table for ginfo
  • table for guild id to sounds
  • table for guild id to movies
• Use a .env if it ain't broke
• Add CI/CD
• Make beta bot that's always running or easy to run
• improve TTS voice

it's your birthday today

cake and icecream

is on its way

current addsound usage involves sending a file and incorporating it into the list using its filename. proposed addition to the command allows the use of an argument to automatically rename the file after the user uploads it

in effect, it will be identical to uploading a sound, then immediately renaming it

Code layout is a mess. Refactor the commands and logic into cogs to separate them into more manageable bits.

Example list:

• Maintenance:
  • update, perish
• Scoreboard:
  • reset score, leaderboard (idk about +1)
• Sound player:
  • play (private), stop, volume
• TTS:
  • speak, birthday
• Fun
  • snap, fsmash
  • have this rely on the sounds cog maybe?
• Guild sounds:
  • play (public), addsound, rename, delete
• Control:
  • join, leave, help
  • Maybe join this with maint.