README

Ebony Framework

An easy to use, module-based, multi-channel chatbot framework.

Features

Separation in independent, portable, easy to customize modules
Separate adapters for interfacing with communication channels (Facebook Messenger, Slack, Viber, SMS etc.)
Dialogue context management system
Route-based rules, with middlewares
Uses MongoDB for storing user information

Getting Started

For now, there is no detailed documentation for Ebony Framework. You can easily create an ebony-based Facebook Messenger Bot with the Messenger Bot Template. If you have any questions and/or suggestions feel free to open an issue, or create a pull request.

Prerequisites

You'll need to have NodeJS/NPM on your system, as well as Typescript installed either in your node_modules folder or globally in your system.

Install Ebony Framework

$ npm i ebony-framework

Install an adapter

(For now only a Messenger Platform adapter is available)

$ npm i ebony-messenger-adapter

Initialize the bot

import { Bot } from 'ebony-framework';

// Import the adapter (for example the Messenger Adapter) as well as the user model
import { MessengerAdapter, MessengerUser } from 'ebony-messenger-adapter';

// ... Here import the modules you'll use...
import botModule from './modules/botModule';

const adapters = [
    // Here you initialize the adapters one by one.
    // For example the Messenger Adapter:
    new MessengerAdapter({
        pageId: /* Your FB Page ID */,
        pageToken: /* Your Page Token */ ,
        appSecret: /* Your App Secret */,
        webhookKey: /* Your Webhook Key */
    });
];

export const bot = new Bot<MessengerUser>(adapters, {
    mongodbUri: /* Your MongoDB Connection URI */
});

bot.addModule(botModule);

bot.start({
    port: 3000 // You can substitute "3000" with the PORT of your choice.
});

Channel Adapters

For now, only a Messenger Platform adapter is publicly available. If you want to use the Ebony Framework for another channel, open an issue to discuss the development of a new adapter.

Ebony Messenger Adapter

Modules

A module is a portable collection of actions, rules and middlewares that is imported into the chatbot on runtime.

See Modules.md for more information.

Action

An action is a function that when triggered, sends a message to the user.

Rules

There are multiple types of rules inside a module. A rule could be context based (ContextRouter), postback based (PostbackRouter) or even a regex (TextMatcher). These rules, work similarly to routes in a web framework. They point to an action that is triggered when the rule is satisfied.

Middlewares

Middlewares run before (preMiddlewares) and after (postMiddlewares) an action is triggered.

Next Steps

Use redis for user context storing
Decouple database logic from framework (add support for multiple databases)
Create Slack Adapter
Add support for multiple adapters per bot (Allow the same bot to answer through many channels)
Refine the User model, support true extendable user models
Add tests

Authors

Christos Panagiotakopoulos - Initial Work

License

This project is licensed under the MIT License - see the LICENSE file for details.

Usage no npm install needed!