Happiness is learning how to use the Hapi.js (Node.js) web framework to build reliable/scalable apps faster.
Hapi is the framework for rapidly building RESTful & Real-Time web applications and services with Node.js.
Whether you are building a very simple API for your website/mobile app or a large scale, cache heavy, secure e-commerce website, hapi has you covered. Hapi will help get your server developed quickly with its wide range of configurable options.
Most people/teams that have tried Hapi have embraced Hapi to build complete web applications. But if you are only building a REST API (e.g. for a mobile app) please read: https://github.com/dwyl/learn-api-design
Why Hapi instead of XYZ framework?Q: I already know how to build REST APIs in {framework-xyz} why learn a new framework?
A: If you are happy with your existing system & level of team productivity, stick with what you know. If not, learn [how to be] Hapi. (We have built Sites/APIs with both Express, Restify, Sails & Meteor and find Hapi has solved more "real world" problems and thus we end up writing less code. YMMV. See benefits below)
Q: Hapi looks like quite a steep learning curve, how long will it take me to learn?
A: You can get started immediately with the examples below, it will take approximately 60 mins to complete them all (after that add a couple of hours to read/learn further). The most important part is to try Hapi on a simple project to gain experience/confidence.
@ethanmick wrote a detailed post on why he prefers Hapi to Express: https://www.ethanmick.com/why-i-like-hapi-more-than-express/ --its worth a read. PDF
Beginner Friendly Examples/Apps to Learn From/WithWe have a few "beginner" example apps (with documentation & tests!) that will help you get started with something a bit more "real world":
For a list of examples see: https://github.com/dwyl?&query=example
Who (is using Hapi) ?The list of teams using Hapi.js to build their node.js apps grows every day! See: https://hapijs.com/community
While you should not make your decisions to use a given technology based on who else is using it, you should be aware that and if you need to answer the question: "Who is already using this in Production?" it's really useful to have a good list.
(Not essential before you start, however) You will benefit from having:
http module.First thing you should do to get familiar with Hapi is work through the makemehapi workshop.
(assumes some node.js prior knowledge but otherwise a gentle self-paced introduction)
Note: makemehapi currently uses Hapi v16. Some major changes were introduced to Hapi in v17. Differences between v16 and v17
Create a new folder on your local machine for your answers to makemehapi:
mkdir makemehapi && cd makemehapi
Install the workshop:
npm install -g makemehapi@latest
( if it fails to install see: https://stackoverflow.com/questions/16151018/npm-throws-error-without-sudo )
Once its installed, start the tutorial with the following command:
Try to complete all challenges.
If you get stuck, you can either google for the specific error you are seeing or if you are not "getting" it, you can always look at my answers in the /makemehapi directory of this repository or
the "official" solutions in the /makemehapi/exercises/{exercise-name}/solution directory
e.g: https://github.com/hapijs/makemehapi/tree/master/exercises/hello_hapi/solution
or if you still don't get it, ask us: https://github.com/dwyl/learn-hapi/issues
For the rest of the tutorial we will cover the various plugins and features we have used in the Hapi.js ecosystem that will help you getting up-and-running with Hapi!
Recap: Hello World in HapiOnce you have completed the makemehapi workshop, on your computer, create a new directory called "hapiapp". e.g:
mkdir hapiapp && cd hapiapp
Type out (or copy-paste) this code into a file called index.js
const Hapi = require('hapi');
const server = new Hapi.Server({port: 3000}); // tell hapi which TCP Port to "listen" on
server.route({
method: 'GET', // define the method this route will handle
path: '/{yourname*}', // this is how you capture route parameters in Hapi
handler: function(req, h) { // request handler method
return 'Hello ' + req.params.yourname + '!'; // reply with text.
}
});
async function startServer() {
await server.start() // start the Hapi server on your localhost
console.log('Now Visit: http://localhost:' + server.info.port + '/YOURNAME');
}
startServer();
module.exports = server;
Install Hapi:
npm init -y && npm install hapi --save
Run:
Visit: http://localhost:3000/YOURNAME (in your browser) you should see something like:
Validation is a fancy way of saying "checking" a value is the type / format and length you expect it to be.
e.g. imagine you ask people to input their phone number and some joker enters letters instead of numbers. The validation will display a message to the person informing the data is incorrect.
Joi is the validation library built by the same team as Hapi. Most people use Joi with Hapi, but given that it is a separate module, plenty of people use Joi independently; its well worth checking it out!
An example: Type out (or copy-paste) this code into a file called hellovalidate.js
// Start this app from your command line with: node hellovalidate.js
// then visit: http://localhost:3000/YOURNAME
const Hapi = require('hapi'),
Joi = require('joi');
const server = new Hapi.Server({ port: 3000 });
server.route({
method: 'GET',
path: '/{yourname*}',
config: { // validate will ensure YOURNAME is valid before replying to your request
validate: {
params: {
yourname: Joi.string().min(2).max(40).alphanum().required()
}
},
handler: function (req, h) {
return 'Hello '+ req.params.yourname + '!';
}
}
});
async function startServer() {
await server.start(); // start the Hapi server on your localhost
console.log('Now Visit: http://localhost:' + server.info.port + '/YOURNAME');
}
startServer();
Now try entering an invalid name: http://localhost:3000/T
You should see a Validation Error:
This might not look like a "Friendly" Error message. But as we will see later, it provides all the information we need in our Client/App and we can display a more user-friendly error to people.
Joi has many more useful validation methods. We will use a few of them later on when we build our example app.
If you're new to Test Driven Development (TDD) read our Beginners' TTD Tutorial: https://github.com/dwyl/learn-tdd (first)
and then come back to this tutorial!
If you've done functional or unit testing in previous programming projects you will be at home with Lab.
Lab borrows heavily from Mocha, so if you followed our learn-mocha tutorial this should all be familiar.
(Using the code we wrote above in the Validation with Joi section with a minor addition) An example of testing with Lab:
const Lab = require("lab"); // load Lab module
const lab = exports.lab = Lab.script(); //export test script
const Code = require("code"); //assertion library
const server = require("../examples/hellovalidate.js");
lab.experiment("Basic HTTP Tests", function() {
// tests
lab.test("GET /{yourname*} (endpoint test)", async function() {
const options = {
method: "GET",
url: "/Timmy"
};
// server.inject lets you simulate an http request
const response = await server.inject(options);
Code.expect(response.statusCode).to.equal(200); // Expect http response status code to be 200 ("Ok")
Code.expect(response.result).to.have.length(12); // Expect result to be "Hello Timmy!" (12 chars long)
await server.stop();
});
});
First we create a test suite for our test Lab.experiment (the first argument is the name of the test suite "Basic HTTP Tests")
Next is a basic test that calls the only route we have /{yourname} in this case GET /Timmy.
We expect to receive a 200 http status code and the response body to be the text "Hello Timmy!".
"scripts": {
"test": "lab -c"
}
The result should look something like this:
Note how the test script has a -c (coverage) flag above this give us the code coverage.
We have 100% code coverage so we can move on to our next test/feature!
Note on Testing: Tape is Simpler than Lab+CodeHow do you think we would write a test for an error? (hint: have a look inside ./test/test.js and see the second test :)
While Lab is really Good and is the "official" testing framework used by Hapi, we prefer
the simplicity of tape; we find our tests are simpler to write/read/understand. #YMMV Also we prefer to use a separate & specialised module for tracking test coverage: istanbul which we find does a better job at tracking coverage...
The preceding Lab test can be re-written (simplified) in Tape as:
const test = require('tape');
const server = require("../index.js"); // our index.js from above
test("Basic HTTP Tests - GET /{yourname*}", async function(t) { // t
const options = {
method: "GET",
url: "/Timmy"
};
// server.inject lets you similate an http request
const response = await server.inject(options);
t.equal(response.statusCode, 200); // Expect http response status code to be 200 ("Ok")
t.equal(response.result.length, 12); // Expect result to be "Hello Timmy!" (12 chars long)
await server.stop();
t.end(); // t.end() is required to end the test in tape
});
These tests are functionally equivalent in that they test exactly the same outcome. Decide for yourself which one you prefer for readability and maintainability in your projects.
For our Tape Tutorial see: https://github.com/dwyl/learn-tape
Making sure your code is working as you expect it to (over time).
Integrating Hapi with Travis CIIf you are new to Travis-CI or need a refresher see: https://github.com/dwyl/learn-travis
We have Travis-CI enabled for all our hapi.js based projects:
So if you need an example to follow, check out our repos!
And as always, if you have any questions, ask!
Boom makes custom errors easier in Hapi. Imagine you have a page or item of content (photo, message, etc.) that you want to protect from public view (only show to someone who is logged in).
First install boom:
npm install boom --save
Next write another test in ./test/test.js (If you aren't used to "Test First" - trust the process...)
lab.experiment("Authentication Required to View Photo", function() {
// tests
lab.test("Deny view of photo if unauthenticated /photo/{id*} ", async function() {
const options = {
method: "GET",
url: "/photo/8795"
};
// server.inject lets you simulate an http request
const response = await server.inject(options);
Code.expect(response.statusCode).to.equal(401); // Expect http response status code to be 200 ("Ok")
Code.expect(response.result.message).to.equal("Please log-in to see that"); // (Don't hard-code error messages)
});
});
When you run npm test you will see a fail:
Next we want to make this test pass and we'll use Boom to get our custom error message.
The wrong way of doing this is to explicitly hard-code the response for this route. The right way is to create a generic route which responds to any request for a photo with any id. And since we don't currently have any authentication set up, we mock (fake) it. (Don't worry we will get to the authentication in the next step...)
const Boom = require('boom');
server.route({
method: 'GET',
path: '/photo/{id*}',
config: { // validate will ensure `id` is valid before replying to your request
validate: { params: { id: Joi.string().max(40).min(2).alphanum() } },
handler: function (req, h) {
// until we implement authentication we are simply returning a 401:
return Boom.unauthorized('Please log-in to see that');
// the key here is our use of the Boom.unauthorised method
}
}
});
Our test passes but the point was to show that returning errors with specific messages is easy with Boom.
Have a look at https://github.com/hapijs/boom for more error response options. We will be using these later as we build our app. Let's move on to authentication.
For a more user-friendly approach to error-handling see: https://github.com/dwyl/hapi-error
Application logging can often be an afterthought developers only implement after they have a production bug which is crashing their API/App and they are scrambling to try and "debug" it.
Thankfully, it Hapi has first-class support for logging with the good module.
We have written a little example you can use to get started: examples/hellogood.js
Run it locally with node examples/hellogood.js then visit http://localhost:3000/hello/yourname in your browser.
Note: Good is not yet compatible with Hapi 17, so this code will only run if you are using v16. See here for more details
You should expect to see something like this:
There are good examples including logging use http (e.g. to a 3rd party logging tool) in the Good repo: https://github.com/hapijs/good/tree/master/examples
Again, if you have any questions, ask
Authentication is the process of determining whether someone or something is, in fact, who or what it is declared to be.
Authentication (or "Auth") is something many novice (naive?) developers attempt to write themselves. (I was once that kid... trust me, we have bigger fish to fry, use a well-written/tested library!)
We have 4 options:
The go-to solution for 3rd party authentication in hapi is bell: https://github.com/hapijs/bell. There are a few good examples in the repo: https://github.com/hapijs/bell/tree/master/examples.
Most apps don't need caching from "Day 1" (because you don't know upfront where your app's bottlenecks are going to be...).
But, once again, the team that brought you Hapi.js have solved the problem of caching, see: https://github.com/hapijs/catbox/ and https://hapijs.com/tutorials/caching
Using Socket.io with Hapi for Real-Time AppsWe use Redis for blazing fast application and data caching. Hapi.js Catbox makes this very easy!
Using Socket.io with Hapi.js could not be easier! To help you get started we've built a fully-functional chat application with tests (now featured on the hapijs.com Resources page), which demonstrates the power of Real-Time data-synching in your apps.
There were some major changes introduced to Hapi when version 17 was released. For a full list, see the version 17 release notes, but here are the major differences relevant to this guide:
async functions. This means that instead of passing a callback to the function, and having that called when the function is finished, hapi functions return a promise that can either be resolved, or called synchronously with await. For more on async functions, see MDN and this guideserver.connection() replaced with options passed directly into the server when it's created. A small change, but important to update. Before, we created our server with:var server = new Hapi.server();
and passed our options to:
server.connection({port: 8000});
Now, we just pass our options straight away, and no longer need to call the connection method:
const server = new Hapi.server({port: 8000});
reply() interface replaced with a new lifecycle methods interface. You no longer have to call reply when sending a response from a handler. You can now just:And the reply parameter to your handler has been replaced with a response toolkit (h) containing helpers from hapi core and your plugins.
Not all of the hapi plugins have been updated to work with v17 yet (For example Bell, and Good), so be careful if you decide to upgrade an existing project.
The previous version of this tutorial and code examples for Hapi 16 can be found here: https://github.com/dwyl/learn-hapi/tree/b58495ea002a9f3f8af8d183f6004d2b483f4591
Please Suggest Improvements!If you want to extend this tutorial or simply request additional sections, open an issue on GitHub: https://github.com/dwyl/learn-hapi/issues
Background Reading / WatchingGitHub Repo: https://github.com/hapijs/hapi (has documentation)
An ecosystem of tools and best practices for the working hapijs developer (up-to-date with last version of hapi): https://hapipal.com/ https://hapipal.com/getting-started
Restify vs Express performance: https://stackoverflow.com/questions/17589178/why-should-i-use-restify
REST API in Express: https://pixelhandler.com/posts/develop-a-restful-api-using-nodejs-with-express-and-mongoose
Hapi API Reference: https://github.com/hapijs/hapi/blob/master/API.md
Selected StackOverflow Questions & Answers:
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4