droff v0.43.6

Simple Discord client powered by RxJS and Axios

• API documentation: https://tim-smart.github.io/droff/droff/

Note: Looking for a Discord library that is more functional in nature?
Check out: https://github.com/tim-smart/dfx

Goals

• Lightweight - Simple Axios wrapper for the REST API with a lean Gateway API wrapper powered by RxJS.
• Functional - Favour functional composition over inheritence.
• Scalable - Every component of the library can be scaled seperately, allowing for easy horizontal distribution of your bot.

Packages

Install

yarn add droff

Usage

Basic ping example. Look at droff-interactions and droff-commands for examples that work with slash commands etc.

Please note that you have to subscribe to client.effects$ for the client to function. This essentially starts the client.

import { createClient, Intents } from "droff";
import * as Rx from "rxjs";
import * as RxO from "rxjs/operators";

const client = createClient({
  token: process.env.DISCORD_BOT_TOKEN!,
  gateway: {
    intents: Intents.GUILD_MESSAGES,
  },
});

const pings$ = client.fromDispatch("MESSAGE_CREATE").pipe(
  RxO.filter((msg) => msg.content === "!ping"),
  RxO.flatMap((msg) =>
    client.createMessage(msg.channel_id, {
      message_reference: { message_id: msg.id },
      content: "Pong!",
    }),
  ),
);

// Subscribe to our side effects
Rx.merge(client.effects$, pings$).subscribe();

Gateway proxy

Larger bots may want to seperate the websocket handling from the bot logic, for zero downtime deployments.

To do this you would pipe the gateway dispatch events into a event streaming tool, like Apache Kafka or Rabbitmq, then subscribe to the events in your bot logic.

See example/gateway-proxy.ts for an example.

Also see the Redis example.

REST proxy

Larger bots may want to funnel all Discord HTTP requests through a single proxy server, to simplify rate limiting.

See example/proxy.ts for an example.

Caching

Droff will only activate the caches that you use. So by default nothing is cached.

To use a cache, you call one of the cache factory methods, optionally passing in a store implementation.

import * as Rx from "rxjs";
import { createClient, Intents } from "../src/mod";

const client = createClient({
  token: process.env.DISCORD_BOT_TOKEN!,
  gateway: {
    // You will need to enable some intents for the corresponding cache.
    intents: Intents.GUILD_EMOJIS | Intents.GUILD_MEMBERS,
  },
});

const [roleCache, roleCacheEffects$] = client.rolesCache();

// Subscribe to the cache effects if you want to populate the cache from the
// gateway events
Rx.merge(client.effects$, roleCacheEffects$).subscribe();

// You can then use the cache:
roleCache.getForParent("guild id xxx").then((map) => map.get("role id xxx"));

What's missing

Just a heads up that is a relatively new client library. You should note that:

• There isn't much documentation
• There isn't anything implemented for Voice channels
• Some caches will be missing
• No benchmarking / optimization has been done at this point

Pull requests are more than welcome :)