Gamalon Botkit Middleware

This middleware plugin for Botkit allows developers to easily integrate Gamalon's classification service.

Quick Start

1. Run npm install.
2. Get your CLIENT_ID, CLIENT_SECRET and TREE_ID from Gamalon.
3. Create a .env file that should look as follows:

CLIENT_ID=<your Gamalon client id>

CLIENT_SECRET=<your Gamalon client secret>

TREE_ID=<your Gamalon tree id>

4. Run node .
5. Navigate to localhost:3000 to see the chatbot. It will spit out the classification results of any message you type.

API

skills/gamalon.js contains an example of how you might use the middleware.

• First you'll want to instantiate the middleware, passing in the client id, client secret, and tree id. This will tell the middleware which tree should be used for classification.

const gamalonMiddleware = require('../gamalonMiddleware')({
  clientId: process.env.CLIENT_ID,
  clientSecret: process.env.CLIENT_SECRET,
  treeId: process.env.TREE_ID,
});

• Once you've created the middleware, you can to use it. Gamalon's middleware can be used at any step of the incoming messages pipeline: ingest, normalize, categorize, receive. Below is an example of adding it to the receive step.

controller.middleware.receive.use(gamalonMiddleware);

• The middleware classifies the text from the message and adds the results to message.gamalon. It has the following fields:
  • error - A string that describes the error. If it is truthy, then an error has occurred and you should have your bot respond appropriately. Otherwise the classification results can be used.
  • intent - The name of the intent with the highest probability.
  • confidence - The confidence of the classification.
  • path - The path is the tree that leads to the intent.

controller.on('message_received', function(bot, message) {
  const { error, intent, confidence, path } = message.gamalon;

  //...
});

Tests

To run test use npm test. You must have your client id, client secret, and tree id in the .env file.

Using the middleware

• Copy the gamalonMiddleware directory into your project.
• Run npm install --save request