henrymxu / evebot

E.V.E Bot is a Voice Enabled Discord Bot!

E.V.E Bot

E.V.E is a voice enabled all in one discord bot. Feel free to host your own instance!

If you are a developer and wish to contribute / customize E.V.E, check out the wiki!

Table of Contents

• Features
• Usage
• Installation
• Configurations
• Troubleshooting
• Support
• License

Features

• Play music!
• Start a custom radio!
• Clip or recite voice channel activity!
• Say anything you want with the bot!
• And much more!

Usage

1. Add the bot to your Discord Server
   • Once the bot has been started (see here), add it through the Discord Developer Portal
2. Use the ?help command to see the various command groups
   • Use the ?help command with a command name or group to see the command details (e.g ?help play)
3. Say one of the provided hotwords to start a voice command (Alexa, Hey Siri, and OK Google)
   • Wait for the chime to start saying your command
   • Don't worry about other people interrupting you

Installation

The installation process for E.V.E is quite long due to all the different services it requires.

These steps are if you wish to run the bot on your local machine (i.e. PC, laptop). E.V.E runs on macOS and Linux (Windows has not been tested).

Pre Installation

1. Retrieve all the required / desired tokens
2. Install the following external dependencies
   • ffmpeg
   • lame (pre-installed on macOS)
   • node / npm (node >=12.0.0)
3. Set up the environment variables and api tokens
   • Set up a json file with the key value pairs
   • An example can be found in example_keys.json
   • Environment variables can be used instead of a json file

Installation from NPM

1. Install the package (npm install -g evebot)
2. Start the bot using evebot <path to keys.json> (e.g evebot ~/Documents/keys.json)

Installation From Source

1. Clone / Download the git repository (git clone [email protected]:henrymxu/evebot.git)
2. Copy the json file you created above into the evebot directory
   • Rename this file to be keys.json
3. Inside a terminal, navigate to the newly cloned evebot directory
4. Install TypeScript (npm install -g typescript) and ts-node (npm install -g ts-node)
5. Install the dependencies using npm install
6. Start the bot using npm start

Local Installation with Docker

// TODO

Heroku Deployment

// TODO

Environment Variables

Name	Environment Variable	Description	Required
Bot Environment	bot_environment	Informs the bot to use either 'development' or 'production' settings. Defaults to development.	✗
Youtube Cookie	youtube_cookie	Required for the bot to bypass Youtube 429 Errors (Too many requests).	✗

General API Tokens

Name	Environment Variable	Description	Required
Discord Bot Token	discord_token	Required for the bot to login and communicate with discord.	✓
Youtube Search API Token	youtube_api_token	Required for the bot to search for song URLs (Hopefully this will be converted to a scraper).	✓
Genius API Token	genius_token	Required for the bot to find lyrics of songs.	✗
Spotify API Tokens	spotify_id	Required for the bot to create radio playlists, search albums and playlists.	✗
	spotify_secret		✗

Platform API Tokens

Details on how to authenticate and configure each provider can be found in each of their individual sections.

Summary

Name	Speech-to-Text	Text-to-Speech	Cloud Storage
IBM Watson	✓	✓	✗
Microsoft Azure	✓	✓	✗
Google Cloud	✓	✓	✗
Amazon Web Services	✗	✗	✓

IBM Watson

Environment Variable	Description
watson_token	API Key
watson_url	URL

Microsoft Azure

Environment Variable	Description
microsoft_token	KEY
microsoft_location	LOCATION (e.g eastus)

Google Cloud

Environment Variable	Description
google_keyFileCred	Contents of the Google Cloud JSON Credentials File

Amazon Web Services

Environment Variable	Description
AWS_REGION	AWS Region
AWS_ENDPOINT	Endpoint
AWS_ACCESS_KEY_ID	Access Key ID
AWS_SECRET_ACCESS_KEY	Secret Access Key

Configurations

Troubleshooting

• The bot doesn't seem to recognize your voice commands

  • Wait until you hear the chime from the bot after saying the hotword to begin your voice command
  • Make sure the voice command doesn't get cut off by the ending chime

• The bot seems to be unresponsive or not listening to any hotwords

  • Try using the rejoin command

• The bot is unable to find any songs to play

  • Try adding / updating the youtube_cookie to the bot

Support

• Feel free to create an issue on GitHub if there are any problems installing or using E.V.E!

License

This repository is licensed under Apache 2.0. You may host an instance of this bot for non-commercial and personal use free of charge.

evebot's Issues

Implement dynamic voice command length

Is your feature request related to a problem? Please describe.
Users are sometimes cut off when saying a voice command, this is due to the voice command prompt only allowing a fixed length.

Describe the solution you'd like
Allow dynamic voice command length (potentially up to a maximum to prevent extremely long speech to text streams).

Describe context of feature
Changes to the ConnectionHandler to implement this.

Finish README

• Missing detailed installation guide
• Troubleshooting
• Improve keys setup

Requesting help on an aliased command doesn't return the correct keyword

Describe the bug
If a command has an alias, and the alias is used in the help command, the help command returns an incorrect description

To Reproduce
e.g play is aliased with test, help test is used, the following message will use test as the keyword instead of play

Expected behaviour
help message uses the correct keyword (maybe it should show aliases in the message as well)

Standardize Acknowledgments

• Errors on command execution should be thrown with an acknowledgement type or string
• This error should be handled by the dispatcher
• Execution function should return a Promise, that contains an appropriate acknowledgement type or string

Skip multiple songs at once

Is your feature request related to a problem? Please describe.
Want to skip multiple songs at once

Describe the solution you'd like
Allow skip command to skip multiple songs at once

Describe context of feature
Command improvement

Lint and Prettify Project

• Use eslint and prettier to format the project

Add option to start radio after playing a song

Is your feature request related to a problem? Please describe.
After playing a song with the play command with nothing in queue, the discord channel the bot was playing is filled with awkward silence

Describe the solution you'd like
I would like to have the option to start a spotify radio after my requested song is played to fill in this void.

Describe context of feature
Add an option to start a spotify radio in the play command
ex. -play blackpink lovesick girls -radio

Additional context
u mam

Snowboy Deprecation

With Snowboy deprecation coming soon, and Porcupine's introduction of Alexa / OK Google / Hey Siri, Snowboy likely does not to be included / used in the library.

Pros:

• Reduce node dependencies
• Reduce linux dependencies

Cons:

• One less reference for hotword detection

Improve Table messages

• Large table messages are split across multiple messages, this is unappealing and might be considered spammy in some servers
• Once #28 has been closed, add an option to scroll through the table by reacting to the message (similar to Groovy)

Implement Unit Tests

• Add unit tests (using jest?)
• Mocking discordjs might be hard
• Some helpful articles / documentation for implementing tests

https://medium.com/@RupaniChirag/writing-unit-tests-in-typescript-d4719b8a0a40
https://jestjs.io/docs/en/getting-started
https://medium.com/codeclan/mocking-es-and-commonjs-modules-with-jest-mock-37bbb552da43
https://flaviocopes.com/jest/#mocking
https://dev.to/jackcaldwell/mocking-es6-class-methods-with-jest-bd7

Improve Bot Acknowledgement

Is your feature request related to a problem? Please describe.
More standardized bot acknowledgement (i.e reacting with emojis)

Describe the solution you'd like
Define different acknowledgement types, allow users to customize the emojis that represent each acknowledgement type

Describe context of feature
Improvement to existing feature

Additional context
Users should be able to customize the emojis through configs

Shuffle Queue Feature

• Add option to shuffle queue
• Add this option to queueing playlists / albums as well

Check if config contains all properties of default config on load

• Check if the guild config contains all required properties by comparing against default config when loaded.
• This allows E.V.E bot to have updates to the config and ensure that previously generated configs will be updated properly

Resolve Playlist / Album songs later

Issue

Currently when a playlist / album is queued, every song is passed through YTDL-core to determine whether or not a valid song can be found. This can cause issues when large playlists / albums are queued as it takes a while to resolve all these songs.

Fix

Possibly only resolve the first 5? or so songs in the playlist / album, then as songs are removed from the queue, resolve another one in the background

Voice Receiving causing crash when user on browser joins

Describe the bug
DiscordJS attempts to decode corrupted packet (from user joining from browser). This error is not handled to the bot crashes

To Reproduce
Have user join on browser and say something, bot will then crash with error:

/evebot/node_modules/prism-media/src/opus/Opus.js:64
    return this.encoder.decode(buffer, Opus.name === 'opusscript' ? null : this._options.frameSize);
                        ^
TypeError: The compressed data passed is corrupted
    at Decoder._decode (evebot/node_modules/prism-media/src/opus/Opus.js:64:25)
    at Decoder._transform (evebot/node_modules/prism-media/src/opus/Opus.js:189:20)
    at Decoder.Transform._read (_stream_transform.js:191:10)
    at Decoder.Transform._write (_stream_transform.js:179:12)
    at doWrite (_stream_writable.js:403:12)
    at writeOrBuffer (_stream_writable.js:387:5)
    at Decoder.Writable.write (_stream_writable.js:318:11)
    at Readable.ondata (_stream_readable.js:695:22)
    at Readable.emit (events.js:310:20)
    at Readable.EventEmitter.emit (domain.js:482:12)

Expected behaviour
Should not crash

Bot Configuration:

• OS: [e.g. macOS, Ubuntu]
• SpeechEngines (if applicable)
• HotwordEngines (if applicable)
• Database (if applicable)

Additional context

Created Issue on DiscordJS Repo here

Implement option to opt-out of voice commands

Is your feature request related to a problem? Please describe.
Some people may not wish to have the bot always be listening to them

Describe the solution you'd like
Provide a way to opt out of the bot listening to them

Describe context of feature
Configuration command (i.e add user ids to array)

Additional context
Users would not be able to access voice commands

Improve Bot Responses / Implement max length of various text inputs

• Some bot responses are extremely long, they should be truncated in order to improve the overall visual quality
• Some user inputs are not restricted in text length (e.g macro command name, alias, etc.), these should also be limited

Implement feature to allow user to acknowledge bot messages

Is your feature request related to a problem? Please describe.
No way for bot to ask user to do something (i.e if radio is done, allow user to react to a message and bot will retrieve some more songs)

Describe the solution you'd like
Allow commands to request users to acknowledge a response in order to complete a commnad

Describe context of feature
Bot responds with a message, user reacts to it, bot then performs an action based on reaction

Additional context

Determine command execution time

• Curious to see how long command execution takes (from message to response)
• Should be straight forward to implement using dispatcher and command execution
• Would be nice to store these execution times in order to find improvements