TPP.Core

This is the core executable project that ties all components together.

You can set yourself up and run it to connect to twitch chat and execute a bunch of basic commands, like !reddit or !stopnew. More elaborate features are still under active development.

run the project

First, ensure the project builds and runs on your system by executing it once. For example, to see all command line options, do

dotnet run -- --help

Then, try to test your yet non-existing base config file using testconfig. It will tell you what file you need and how to create it.

dotnet run -- testconfig

For local testing, you need to remove all but the console entry from the Chat.Connections config. You will be able to use the console as a simulated chat.

If you want to connect to an actual Twitch chat you need to keep the twitch entry of the Chat.Connections list and change a few configurations:

1. In the twitch entry, Username and Password contain the credentials of some Twitch account that will be the chat bot. You can obtain an oauth token from here. These configs are still required because not all chat functionality is migrated to the Twitch API or EventSub yet. In the future, you will only need to do 2.) and 3.).

2. In the twitch entry, AppClientId and AppClientSecret contain the credentials of some

   Twitch Developer Application. You can create your own application under any account you like. If you are TPP staff, this typically is the official "Twitch Plays Pokemon" application, which is owned by FelkCraft and has the AppClientId 1iinedru2u5as4ss7pzn237olcrgk2o.

3. In the twitch entry, RefreshToken contains the refresh token of some user that authenticated against the app configured in 2.), and can be obtained from e.g. twitchtokengenerator.com. You should use the same account as you did in 1.).

   If you are TPP staff, you can also use twitchplayspokemon.tv/custom_scopes, which uses the official "Twitch Plays Pokemon" application for authenticating. For a list of scopes you need to authenticate, see the section "Twitch Authorization Scopes".

4. No chat messages are actually being sent by default. To change this, in the twitch entry set Channel to some unpopulated twitch channel, preferably the bot's own channel, and add the channel name to SuppressionOverrides. You may also add your own name to SuppressionOverrides to be able to receive whispers.

5. Add your name to Chat.DefaultOperatorNames to be able to do stuff requiring elevated privileges, for example issuing the !stopnew command.

All unchanged entries can be deleted. Missing configurations revert to their default value. A minimal configuration may look like this:

{
  "$schema": "./config.schema.json",
  "Chat": {
    "Connections": [{"Type": "console"}]
  }
}

Ensure that you have a properly configured MongoDB server running. See the TPP.Persistence.MongoDB project for instructions.

Finally, you run the project, e.g. in dualcore mode:

dotnet run -- start -m dualcore

Twitch Authorization Scopes

The following scopes are currently in use:

scope	used for
chat:read	Read messages from chat (via IRC/TMI).
chat:edit	Send messages to chat (via IRC/TMI).
user:bot	Appear in chat as bot.
user:read:chat	Read messages from chat. (via EventSub)
user:write:chat	Send messages to chat. (via Twitch API)
user:manage:whispers	Sending and receiving whispers.
moderator:read:chatters	Read the chatters list in the channel (e.g. for badge drops).
moderator:read:followers	Read the followers list (currently old core).
moderator:manage:banned_users	Timeout, ban and unban users (tpp automod, mod commands).
moderator:manage:chat_messages	Delete chat messages (tpp automod, purge invalid bets).
moderator:manage:chat_settings	Change chat settings, e.g. emote-only mode (mod commands).
channel:read:subscriptions	Reacting to incoming subscriptions

modes

You can run different modes by specifying its mode name in the dotnet run -- start -m <mode> command. Some modes require an additional mode-specific config file, but just trying to run them will tell you how to create it. Currently the following modes are supported:

mode	description
dualcore	This is is a simple mode meant for replacing some functionality that has been ported and subsequently removed from the old python core.
match	This mode runs basic match cycles. You need the old overlay to see anything. See the below section on old overlay compatibility.
run	This mode processed inputs from chat and offers them for consumption through a http endpoint. It also sends them to the old overlay.
dummy	This mode purposely does nothing for testing purposes.

The modes match and run require a mode-specific configuration file, which you can test and generate similar to the base config by passing an additional --mode or -m option. See the --help output for more details on that.

old overlay compatibility mode

The new core (this repository) does not contain any visual components. But it does speak the same websocket event dialect the current "old" overlay speaks, so to see anything at http://localhost:5000/overlay the overlay must be started from the old core with python -m tpp overlay. It will then be able to connect to the new core's websocket to receive overlay events.

proper publishing

dotnet run is a development command that always implicitly restores dependencies, builds the project, and then executes it. Running it this way causes a slow startup and prevents the application from handling SIGTERM events on linux.

If you want faster startup times or graceful SIGTERM handling, publish the project first. For example, making a release build and running the resulting executable may look like this:

dotnet publish -c Release
../artifacts/publish/TPP.Core/release/TPP.Core.exe --help