# Using a REST API
REST APIs are extremely popular on the web and allow you to freely grab a site's data if it has an available API over an HTTP connection.
If you've ever seen a music bot that accepts a YouTube query instead of just a video's URL, then you've seen a REST API in action. discord.js uses the Discord API, so you've probably used an API yourself.
# Making HTTP requests with Node
In these examples, we will be using node-fetch (opens new window), an excellent library for making HTTP requests.
To install node-fetch, run the following command:
npm install node-fetch
# Skeleton code
To start off, you're just going to be using this skeleton code:
const { Client, MessageEmbed } = require('discord.js');
const client = new Client();
const prefix = '!';
client.once('ready', () => {
console.log('Ready!');
});
client.on('message', async message => {
if (!message.content.startsWith(prefix) || message.author.bot) return;
const args = message.content.slice(prefix.length).trim().split(/ +/);
const command = args.shift().toLowerCase();
// ...
});
client.login('your-token-goes-here');
TIP
We're going to take advantage of destructuring in this tutorial to maintain readability.
# Using node-fetch
node-fetch is a lightweight, Promise-based module that brings the Fetch API (opens new window), which is available in browsers, to node. If you aren't already familiar with Promises, you should read up on them here.
In this tutorial, we'll be making a bot with two API-based commands using the random.cat (opens new window) and Urban Dictionary (opens new window) APIs.
To require node-fetch, you'd do:
const fetch = require('node-fetch');
# Random Cat
Random cat's API is available at https://aws.random.cat/meow and returns a JSON (opens new window) response. To actually fetch data from the API, you're going to do the following:
fetch('https://aws.random.cat/meow').then(response => response.json());
It may seem like this does nothing, but what it's doing is launching a request to the random.cat server. The server is returning some JSON that contains a file property, which is a string containing a link to a random cat. node-fetch returns a response object, which we can change into JSON with response.json(). Next, let's implement this into a command. The code should look similar to this:
if (command === 'cat') {
const { file } = await fetch('https://aws.random.cat/meow').then(response => response.json());
message.channel.send(file);
}
So, here's what's happening in this code:
- You're sending a
GETrequest to random.cat. - random.cat sees your request and gets a random file from their database.
- random.cat then sends that file's URL as a JSON object that contains a link to the image.
- node-fetch receives the response and deserializes it with
response.json(). - You then send the object's
fileproperty in Discord.
WARNING
The response will only be parsed if the server's Content-Type header includes application/json. In some cases you may have to apply the .text() method instead of .json() and JSON.parse() it yourself.
# Urban Dictionary
Urban Dictionary's API is available at https://api.urbandictionary.com/v0/define, accepts a term parameter, and returns a JSON response.
First, you're going to need to fetch data from the API. To do this, you'd do:
const querystring = require('querystring');
if (command === 'urban') {
if (!args.length) {
return message.channel.send('You need to supply a search term!');
}
const query = querystring.stringify({ term: args.join(' ') });
const { list } = await fetch(`https://api.urbandictionary.com/v0/define?${query}`).then(response => response.json());
}
Here, we use Node's native querystring module (opens new window) to create a query string (opens new window) for the URL so that the Urban Dictionary server can parse it and know what to search.
If you were to do !urban hello world, then the URL would become https://api.urbandictionary.com/v0/define?term=hello%20world since the string gets encoded.
You can get the respective properties from the returned JSON. If you were to view it in your browser, it usually looks like a bunch of mumbo jumbo. If it doesn't, great! If it does, then you should get a JSON formatter/viewer. If you're using Chrome, JSON Formatter (opens new window) is one of the more popular extensions. If you're not using Chrome, search for "JSON formatter/viewer <your browser>" and get one.
Now, if you look at the JSON, you can see that it's a list property, which is an array of objects containing various definitions for the term (maximum 10). Something you always want to do when making API-based commands is to handle no results. So, let's throw a random term in there (e.g. njaksdcas) and then look at the response. The list array should then be empty. Now you are ready to start writing!
As explained above we want to check if the API returned any answers for our query like so:
if (!list.length) {
return message.channel.send(`No results found for **${args.join(' ')}**.`);
}
After making sure that there are results, you will use those results. For now, let's send back the definition and nothing more. It's as simple as:
message.channel.send(list[0].definition);
Here, you are only getting the first object from the array of objects called list and grabbing its definition property.
If you've followed the tutorial, you should have something like this:
Now, let's just make this an embed.
We are also going to be defining a utility function at the top of our file so that our embed doesn't error when the field value is over 1024 characters. Here is a bit of code to do that:
const trim = (str, max) => ((str.length > max) ? `${str.slice(0, max - 3)}...` : str);
The following snippet is how we'll be structuring the embed:
const [answer] = list;
const embed = new MessageEmbed()
.setColor('#EFFF00')
.setTitle(answer.word)
.setURL(answer.permalink)
.addFields(
{ name: 'Definition', value: trim(answer.definition, 1024) },
{ name: 'Example', value: trim(answer.example, 1024) },
{ name: 'Rating', value: `${answer.thumbs_up} thumbs up. ${answer.thumbs_down} thumbs down.` }
);
message.channel.send(embed);
Now, if you do that same command again, you should get this:
# Resulting code
If you want to compare your code to the code we've constructed so far, you can review it over on the GitHub repository here (opens new window).