What we need for this tutorial
- NPM and NodeJS installed
- node-nlp module
Setting up our project
First create a folder in which you want to create your project, I created folder named chatbot_nodejs.
Initialized repo in that folder with command npm init and install modules with npm i node-nlp.
Create a file named index.js in that folder.
Now our project is setup.
Creating Intents
Create a folder named intents inside our project folder. We will create our intents in JSON, So I created two json file named greetings.bye.json and greetings.hello.json. Inside it, I structured our questions and answers like this.
{
"questions": [
"goodbye",
"bye take care",
"see you later",
"bye for now",
"i must go"
],
"answers": [
"see you soon!",
"Till next time",
"bye bye",
"have a great day"
]
}
{
"questions": [
"hello",
"hi",
"howdy",
"Greetings"
],
"answers": [
"Hey there!",
"Hello",
"Hi",
"Greetings"
]
}
You can add more as you like by creating new json files and structuring your intents like this.
Training and Saving chatbot
Create a file named train.js in our project folder. Let’s write some code to train the model.
// Let's start with importing `NlpManager` from `node-nlp`. This will be responsible for training, saving, loading and processing.
const { NlpManager } = require("node-nlp");
// Creating new Instance of NlpManager class.
const manager = new NlpManager({ languages: ["en"] });
// Let's import fs module to read our json files.
const fs = require("fs");
// Let's read all our intents files in the folder intents
const files = fs.readdirSync("./intents");
// Looping through the files and Parsing the string to object and passing it to manager instance to train and process it.
for (const file of files) {
let data = fs.readFileSync(`./intents/${file}`);
data = JSON.parse(data);
const intent = file.replace(".json", "");
for (const question of data.questions) {
manager.addDocument("en", question, intent);
}
for (const answer of data.answers) {
manager.addAnswer("en", intent, answer);
}
}
// let's create a function that will be responsible for Training and saving the manager instance.
async function train_save(){
await manager.train();
manager.save();
}
// Calling the above function
train_save();
Our train.js file is now complete, Now let’s create our index.js files.
// Let's start with importing `NlpManager` from `node-nlp`. This will be responsible for training, saving, loading and processing.
const { NlpManager } = require("node-nlp");
console.log("Starting Chatbot ...");
// Creating new Instance of NlpManager class.
const manager = new NlpManager({ languages: ["en"] });
// Loading our saved model
manager.load();
// Loading a module readline, this will be able to take input from the terminal.
var readline = require("readline");
var rl = readline.createInterface(process.stdin, process.stdout);
console.log("Chatbot started!");
rl.setPrompt("> ");
rl.prompt();
rl.on("line", async function (line) {
// Here Passing our input text to the manager to get response and display response answer.
const response = await manager.process("en", line);
console.log(response.answer);
rl.prompt();
}).on("close", function () {
process.exit(0);
});
Now our files are complete. All we need to do now is create two scripts train and start. Open package.json and Add two scripts.
"scripts": {
"train": "node train.js",
"start": "node index.js"
},
trainscript will run thetrain.jsfile.startscript will run theindex.jsfile.
Now our project is complete.
You can download this source code from my Github.
