Introduction
[!WARNING]
This demo includes examples for an unsupported version of Materialize (0.26.x).
In this tutorial, we are going to build a web application using AdonisJS and integrate it with Materialize to create a real-time dashboard based on streaming data using standard SQL.
Materialize lets you define views you want to maintain on your data, just as you would with any SQL table, and then get the results in real-time, unlike traditional databases which frequently act as if they've never been asked that question before.
Prerequisites
You need to have the following things installed before getting started:
What is Materialize
Materialize is a streaming database that takes data coming from sources like Kafka, PostgreSQL, S3 buckets, and more and allows you to effectively transform it in real-time using SQL.
Unlike a traditional database, Materialize is able to incrementally maintain views on top of streaming data, providing fresh and correct results as new data arrives. This means that, instead of recomputing the view from scratch every time it needs to be updated, it only does work proportional to the changes in its inputs, so it is fast and efficient.
In the context of web development, Materialize can be used as a backend to power real-time applications (as we'll see in this demo)!
Running a Materialize Demo
For the sake of this tutorial we are going to run the following Materialize Demo:
Materialize - Log Parsing Demo
The setup of the demo is the following:
We would not get into too much details here, but if you have not gone through this demo before, make sure to read it!
To run the demo, follow these steps:
First things first, before you could run the demo, you need to clone the repository:
- Clone the repository:
git clone https://github.com/bobbyiliev/mz-http-logs.git
- Once that is done, switch to the repository directory:
cd mz-http-logs
- Then start all services:
docker-compose up -d
With that, you would have your Materialize instance up and running. Next we will prepare our AdonisJS installation and use AdonisJS to create our Materialize sources and views!
What is AdonisJS
AdonisJS is a web framework for Node.js. It includes everything that you would need to create a fully functional web application or an API.
AdonisJS has been inspired by Laravel and it has its own ORM, Auth support, and a CLI tool called Ace which is very similar to Artisan.
At the end we would extend the Log parsing Materialize demo and have the following setup:
Install AdonisJS
Let's start by installing AdonisJS. To do so, you would need to run the following command:
npm init adonis-ts-app@latest hello-materialize
Once you run that, you will be asked to select a project structure. You will be able to choose between an API, Web App, and a minimal possible AdonisJS app:
CUSTOMIZE PROJECT
❯ Select the project structure … Press <ENTER> to select
api (Tailored for creating a REST API server)
❯ web (Traditional web application with server-rendered templates)
slim (A smallest possible AdonisJS application)
For this tutorial let's go with the web
app! Using your arrow keys select web
and hit enter.
After that you will be asked to choose a name for the project, I will leave this as hello-materialize
but feel free to choose a different name.
I will then press enter and say yes to the rest of the settings:
❯ Enter the project name · hello-materialize
❯ Setup eslint? (y/N) · y
❯ Configure webpack encore for compiling frontend assets? (y/N) › y
This will instantiate the project and might take up to a minute to complete:
Once ready, you can cd
into the new project directory:
cd hello-materialize
And then start the webserver:
node ace serve --watch
If you are coming from the Laravel world, this would be just like running php artisan serve
. The ace
CLI tool is just like artisan
and comes with a lot of the same functionalities.
To check all of the ace
commands, you can run: node ace
.
Installing Lucid
Lucid is the AdonisJS ORM. It is quite similar to Laravel Eloquent.
Lucid comes with an Active Record ORM, Query Builder, Migrations, Seeds, and Factories.
Let's go ahead and install it! To do so, just run the following command:
npm i @adonisjs/lucid
Once done, you would need to do a quick configuration.
Configuring Lucid
In order to configure Lucid, you need to run the following ace
command:
node ace configure @adonisjs/lucid
You will be asked to select the database driver that you want to use. As Materialize is wire-compatible with PostgreSQL, you can connect to it using any pg
driver; here, make sure to select PostgreSQL!
![AdonisJS lucid configuration](https://user-images.githubusercontent.com/21223421/142431728-ac88085b-34cb-4ebb-83c7-b0cae9fb455d.png)
Next, you will be asked to select where you want to display the configuration instructions. I chose In the terminal
, which prints out the necessary environment variables that you have to add to your .env
file.
Configure the Materialize env variables
In order to let our AdonisJS application connect to Materialize, we need to change the PG_*
details in the .env
file.
With your favorite text editor, open the .env
file and update the PG_
environment variables to:
DB_CONNECTION=pg
PG_HOST=localhost
PG_PORT=6875
PG_USER=materialize
PG_PASSWORD=
PG_DB_NAME=materialize
This will allow AdonisJS to connect to Materialize just as it would when connecting to PostgreSQL.
One thing to keep in mind is that Materialize doesn’t yet support the full system catalog of PostgreSQL (we're working on it!), which means that ORMs like Lucid, Prisma, Sequelize, or TypeORM might fail during some attempts to interact with Materialize. As we work to broaden pg_catalog
coverage, the integration with these tools will gradually improve!
Creating a Controller
Let's create a controller where we will add the functionality that would allow us to connect to Materialize!
As the Materialize demo is simulating an application log with a lot of visitors, let's call our AdonisJS controller VisitorsController
:
node ace make:controller VisitorsController
This will create a controller file at:
app/Controllers/Http/VisitorsController.ts
Next, let's create the routes that we would need!
Creating the AdonisJS routes
Your routes file is stored at start/routes.ts
. In there we can specify our application URLs and map them to different controllers and methods!
We do not yet have the methods ready, but we know that we would need the following routes:
-
/source
: When visited, this route would create a Materialize source -
/view
: When visited, this route would create a materialized view -
/visitors
: This route would return an event stream with all of the latest changes to our materialized view -
/
: This will be the landing page where we will display the streaming data that we are getting from the/visitors
endpoint and Materialize
Open your routes file at start/routes.ts
and update it so that it has the following content:
import Route from '@ioc:Adonis/Core/Route'
Route.get('/', 'VisitorsController.index')
Route.get('/visitors', 'VisitorsController.visitors')
Route.get('/source', 'VisitorsController.source')
Route.get('/view', 'VisitorsController.view')
Next, let's add a method that would allow us to create a Materialize source as described in the Materialize Log Parsing Demo!
Creating a Materialize Source from logs
If you were accessing Materialize directly via a SQL client (like psql
), in order to access data from a continuously produced log file, you would execute the following statement:
CREATE SOURCE requests
FROM FILE '/log/requests' WITH (tail = true)
FORMAT REGEX '(?P<ip>\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}) - - \[(?P<ts>[^]]+)\] "(?P<path>(?:GET /search/\?kw=(?P<search_kw>[^ ]*) HTTP/\d\.\d)|(?:GET /detail/(?P<product_detail_id>[a-zA-Z0-9]+) HTTP/\d\.\d)|(?:[^"]+))" (?P<code>\d{3}) -';
Let's see how we could do that via AdonisJS!
First, open the app/Controllers/Http/VisitorsController.ts
file with your favorite text editor.
The file would have the following content initially:
// import { HttpContextContract } from '@ioc:Adonis/Core/HttpContext'
export default class VisitorsController {}
There are a few things that we would want to do:
- Import Lucid:
import Database from '@ioc:Adonis/Lucid/Database'
- Then inside the VisitorsController class, let's create a method called
source
and
// import { HttpContextContract } from '@ioc:Adonis/Core/HttpContext'
import Database from '@ioc:Adonis/Lucid/Database'
export default class VisitorsController {
public async source({request, response}) {
//Using Ludic to connect to Materialize, we are executing a CREATE SOURCE statement
const res = await Database.rawQuery(
`CREATE SOURCE requests
FROM FILE '/log/requests' WITH (tail = true)
FORMAT REGEX '(\\?P<ip>\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}) - - \[(\\?P<ts>[^]]+)\] "(\\?P<path>(\\?:GET /search/\\\?kw=(\\?P<search_kw>[^ ]*) HTTP/\d\.\d)|(\\?:GET /detail/(\\?P<product_detail_id>[a-zA-Z0-9]+) HTTP/\d\.\d)|(\\?:[^"]+))" (\\?P<code>\d{3}) -';`
);
return res;
}
}
Now, if you were to visit the /source
URL via your browser (http://127.0.0.1:3333/source
) it would create your Materialize source:
Creating a Materialize View
You may be familiar with materialized views from the world of traditional databases like PostgreSQL, which are essentially cached queries. The unique feature here is the materialized view we are about to create is automatically kept up-to-date.
Let's do the same thing as before, but to create a materialized view based on our file source! To do that, let's create a method called view
with the following content:
Add this right after the end of the
source
method
public async view({request, response}) {
//Using Ludic to connect to Materialize, we are executing a CREATE VIEW statement
const res = await Database.rawQuery(
`CREATE OR REPLACE MATERIALIZED VIEW unique_visitors AS
SELECT count(DISTINCT ip) FROM requests;`
);
return res;
}
Our materialized view would show the count of the unique visitors flowing through our demo application.
To create the view, visit the /view
URL via your browser (eg. http://127.0.0.1:3333/view
).
With that, our view will be created and we can move on to the next step!
Creating an event stream
You can query the new materialized view, that we've just created, as usual with a standard SELECT
statement:
SELECT * FROM unique_visitors;
However, in order to take full advantage of the incrementally updated materialized view right from our AdonisJS app, rather than querying Materialize with a standard SELECT
to get the state of the view at a point in time, we will use a TAIL
statement to request a stream of updates as the view changes.
public async visitors({request, response}) {
// First we set a header to identify that this would be an event stream
response.response.setHeader('Content-Type', 'text/event-stream');
// Then we declare a TAIL cursor
await Database.rawQuery('BEGIN');
await Database.rawQuery('DECLARE visitors_c CURSOR FOR TAIL unique_visitors');
// Finally we use FETCH in a loop to retrieve each batch of results as soon as it is ready
while (true) {
const res = await Database.rawQuery('FETCH ALL visitors_c');
response.response.write(`data: ${JSON.stringify(res.rows)}\n\n`)
}
}
For more information about TAIL
, make sure to check out the official documentation here:
If you were to now visit the /visitors
URL via your browser, you would see the following output:
Next, let's create a view where we would use the /visitors
endpoint as an event source and continuously update our web page.
Displaying the number of unique visitors on the frontend
First, before we get started, make sure that you've executed the following command to configure Encore which is used to compile and serve the frontend assets for your AdonisJS app:
node ace configure encore
Then create a new file at:
resources/views/visitors.edge
And add the following content:
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Materialize and AdonisJS</title>
@entryPointStyles('app')
@entryPointScripts('app')
</head>
<body>
<main>
<div>
<h1 class="title"> Hi there! </h1>
<p class="subtitle">
The number of unique visitors is: <strong><span id="count"></span></strong>
</p>
</div>
</main>
<script>
var eventSource = new EventSource("http://127.0.0.1:3333/visitors");
const count = 0;
eventSource.onmessage = function(e) {
const data = JSON.parse(e.data)
//const count = omit(data, 'mz_timestamp', 'mz_diff', 'mz_progressed')
const { mz_diff, mz_progressed } = data;
data.forEach(entry => {
if(entry.mz_diff == -1){
console.log('Old count: ' + entry.count)
} else {
console.log('New count: ' + entry.count)
let countDiv = document.getElementById("count");
countDiv.innerHTML = entry.count;
}
})
};
</script>
</body>
</html>
A quick rundown of the main things to keep in mind:
-
new EventSource
: First we define a newEventSource
and specify our/visitors
endpoint. -
eventSource.onmessage
: Then we listen for new messages to show in the EventStream. -
JSON.parse(e.data)
: After that we parse our data -
data.forEach
: Finally we run a loop and update the total unique visitors counter on the page.
Now if you were to visit your AdonisJS application you would see the following output:
As you can see, rather than making a huge amount of AJAX requests, we just tap into the stream and keep our web page up to date with the latest changes from Materialize!
Conclusion
This is pretty much it! You've now built a web application using AdonisJS that connects to Materialize and pulls the number of unique visitors from your application as new data is logged.
As a next step, make sure to head over to the Materialize Docs and try out some of the available demos:
To learn more about AdonisJS, you can also find the documentation here:
You can find a link to the AdonisJS source code from this demo here:
Streaming Data wiht Materialize and AdonisJS Demo files
To learn more about Streaming SQL, make sure to check out this post here:
Streaming SQL: What is it, why is it useful?
Hope that this was helpful!
Top comments (0)