0.0.4 • Published 3 years ago

smart-assistant v0.0.4

Weekly downloads
-
License
ISC
Repository
github
Last release
3 years ago

Smart assistant

You can use this module for create assistants, chatbots, etc.

About

This is a library created to facilite the creation, training and usage of the NLP.js library.

Install

npm install --save smart-assistant

Or

yarn add --save smart-assistant

Train

You should train the assistant before use it.
we can train the assistant with two text files.

Create a new folder in your project.

cd /path/to/your/project/assistant
mkdir -p ./mind

Create a new text file in this folder and put the text in it.

touch ./mind/documents.txt

Create a new text file in this folder and put the text in it.

touch ./mind/answers.txt
echo "Hello world:introduction.programer" >> ./mind/documents.txt
echo "Hello:introduction" >> ./mind/documents.txt
echo "Hi:introduction" >> ./mind/documents.txt
echo "Hi there:introduction" >> ./mind/documents.txt
echo "introduction:HI" >> ./mind/documents.txt
echo "introduction:Hello" >> ./mind/documents.txt
echo "introduction:Hi there" >> ./mind/documents.txt
echo "introduction.programer:Hello programer" >> ./mind/documents.txt

Your folder struture should look like this:

.
├── mind
│   ├── documents.txt
│   └── answers.txt
└── index.js

Your documents text file should look like this:

Hello world:introduction.programer
Hello:introduction
Hi:introduction
Hi there:introduction

Your answers text file should look like this:

introduction:HI
introduction:Hello
introduction:Hi there
introduction.programer:Hello programer

Usage

const SmartAssistant = require('smart-assistant');

const assistant = new SmartAssistant(
  ['pt', 'en']  // List of supported languages
);

(async () => {

  await assistant.train(
    './model.nlp',// Path to the model, default: ./model.nlp
    './mind/documents.txt', // Path to the documents, default: ./mind/documents.txt
    './mind/answers.txt' // Path to the answers, default: ./mind/answers.txt
  );
  assistant.say('Trained');

  assistant.onConsole(); // Start the console
  // Or
  const answer = await assistant.answer('Hello world'); // Answer to a question
})();