Skip to main content

Plugin Hubtype Babel

For more information, refer to GitHub.

What Does This Plugin Do?​

This plugin allows you to integrate Hubtype Babel in your Botonic project. It works like any other AI/NLU services plugin, like Dialogflow, Watson, etc.

Once installed within your project, this plugin sends all text inputs that users are sending to your bot and queries Hubtype Babel API. The plugin retrieves the call results to enhance the input object. You can then use this data in your routes and actions.

Here is an example of an input object received from a user:

{
    "type": "text",
    "data": "Where is my order?"
}

Example of the same input object after being processed by this plugin:

{
    "type": "text",
    "data": "I want to talk with an agent"
    "intent": "help"
    "confidence": 0.9987
    "intents": [
      {
        "intent": "help",
        "confidence": 0.9987
      },
      {
        "intent": "greeting",
        "confidence": 0.0010
      },
      {
        "intent": "insult",
        "confidence": 0.0003
      }
    ],
    "hasSense": true,
    "entities": []  // Currently not supported
}

Setup​

  1. Install the plugin from npm (or yarn):
npm i --save @botonic/plugin-hubtype-babel
  1. Add it to the src/plugins.js file defining the projectId of the Project you want to use:
export const plugins = [
  {
    id: 'hubtype-babel',
    resolve: require('@botonic/plugin-hubtype-babel'),
    options: {
      projectId: '6800762a-03b5-419e-be97-67feac7bc5b9',
    },
  },
]

Use​

You can use it in your routes like any other NLP plugins:

export const routes = [{ intent: 'order-location', action: OrderLocation }]

Or you can use it in your actions:

import React from 'react'
import { Text } from '@botonic/react'

export default class OrderLocation extends React.Component {
  static async botonicInit({ input }) {
    if (
      input.intents &&
      input.intents.length > 0 &&
      input.intents[0].score > 0.6
    )
      return { response: input.intents[0].attributeArrays.ANSWER_TEXT }
  }
  render() {
    return this.props.response ? (
      <Text>{this.props.response}</Text>
    ) : (
      <Text>I don't know what you're talking about</Text>
    )
  }
}

Plugin Options​

  • projectId: Id of the project used to predict the intent.
  • [includeHasSense]: Whether to include if the input text has sense or not.
    • Default: False.
  • [automaticBotMessagePrefix]: Prefix of an automatic bot message added to the input text. If a message has this prefix, inference will not be executed.
    • Default: '[Automatic Bot Message]'.
  • [host]: Host uri where to request inference.
  • [authToken]: Authorization Token to being able to run inference. Only needed when using the plugin locally.