Cloud Firestore + Authentication FTW

The last post on this site was about how Firebase’s Cloud Firestore was a great option for saving data, especially in new or prototype applications.

In a previous post I also showed how you can use Firebase’s handy authentication mechanism to easily register and authenticate users to your site.

This post will attempt to show how the Firestore and authentication mechanisms can be combined to store user preference information.

Authentication Recap

In the auth post we saw how easy it was to setup and that, when a user authenticates, they have a unique id when you can then use.

firebase.auth().onAuthStateChanged(function(user) {
if (user) {
// User is signed in. uid is now available to use
var uid = user.uid;
}
});

The code above shows how we can watch for when a user logs in. After they have logged in we can then use the uid as a unique identifier for each logged in user. The next step is to start creating some data in our Firestore that we can link to this authenticated user each time they log in.

Save User Related Data

Saving user related data e.g. profile information, with Cloud Firestore is easy. Once you have an authenticated user’s unique identifier you can start to save documents in your Firestore with the UID as an identifier. Each time the user logs in you just look up the document in your Firestore to retrieve the related information about this user/login.

firebase.auth().onAuthStateChanged(function(user) {

var dbUser = db.collection('users’)

.doc(user.uid).set(

{

email: user.email,

someotherproperty: ‘some user preference’

});

});

In the code above we wait for the user to log in, retrieve their UID and create a document in the Firestore using the uid as the index. From this point we can add as many properties to this document that we need, and retrieve the information anytime the user logs in.

Simple stuff! Give Firebase a go, you won’t regret it.

The Problem

Recently I was tasked with providing fast full-text searching, across multiple tables, for a reasonably big database (~2m records).

Immediately I reached for some sort of Solr or Elastic solution. Being the lazy developer that I am, I chose the path of least resistance (most samples online on how to get started), and chose Elastic [edit: would seriously consider Angolia for my next project, if I had less than 10k records to index].

Getting Elastic set up was relatively simple. Having access to a few AWS credits, I used the AWS ElasticSearch service because I didn’t want the responsibility for maintaining the server.

Now to populate with some data.

Taking data from a database and putting it in an Elastic index must be a common scenario with lots of examples and help, right?

Well, not quite as common as I thought, and not as easy as I was hoping.

I was “hoping” for a nice simple user interface; point at my database, select what data I want, import into Elastic, done. I thought, with the relative maturity of Elastic, that it would be baked into the product, or someone would have created an add-on already. At the least, I thought, there would be a plethora of examples to show how to get data from a database to Elastic.

What I found was lacking, and its the reason why I’m making this post to help others looking to do something similar.

Logstash

The currently recommended solution for indexing database data is Logstash (up until 2015 the recommended way was through Rivers).

Logstash. Looks great from the diagram

With Logstash you can connect to your data, filter it, transform it, and then add it to Elastic. Great!, just what I was looking.

Logstash takes quite a bit of work to set up and configure though. The setup is fairly straightforward; just download and install on Windows or Linux. After installation, there is no UI with Logstash, just access to a command line command, to which you supply configurations.

Unfortunately Logstash does not have a way to connect with databases out of the box. To connect logstash to your database you will first need to jump through a couple of hoops:

1. You will need to install the JDBC plugin for Logstash, the instructions on how to do this can be found here.
2. The JDBC plugin does not include any database drivers so you will need to find the driver you need and then download. I am mostly connecting to Postgres so I need to download the Postgres JDBC driver. Make a note of where you download this driver to, you will need the path to the driver to your configuration later.

Providing you have no issues with your Java installation, you should now be ready to go.

Logstash Configurations

Most of your time will be spent fine tuning configurations. Simple configurations are straight forward i.e. connect to a single table with basic field types and import into Elastic. Anything outside these parameters can be challenging.

A basic configuration to connect to a Postgres database and import into a new Elastic index could look something like this:

# file: simple-out.conf
input {
    jdbc {
        # Postgres jdbc connection string to our database, mydb
        jdbc_connection_string = "jdbc:postgresql://localhost:5432/mydb"
        # The user we wish to execute our statement as
        jdbc_user = "postgres"
        # password
        jdbc_password = "mypassword"
        # The path to our downloaded jdbc driver
        jdbc_driver_library = "/path/to/postgresql-9.4-1201.jdbc41.jar"
        # The name of the driver class for Postgresql
        jdbc_driver_class = "org.postgresql.Driver"
        # our query
        statement = "SELECT * from contacts"
    }
}
output {
    elasticsearch {
        protocol = http
        index = "contacts"
        document_type = "contact"
        document_id = "%{uid}"
        host = "https://urltomyelasticinstance"
    }
}

The configuration file above can be run with the following Logstash command:

"\pathtologstashbindirectory\logstash.bat" -f simple-out.conf

1. Connect to the postgres database.
2. Get all the data from the ‘contacts’ table.
3. Connect to your elastic instance.
4. Create a new index called ‘contacts’
5. Create a new document type called ‘contact’
6. Automatically create mappings for the fields in the contacts table.
7. Populate the index with the data.

Even More Productivity?

After experiencing the immense productivity gains from using Ionic for mobile development, surely using Ionic 2 would be even more productive, right?!!?.

Sadly. No. Not yet, at least.

I had hoped to be writing a post telling everyone how awesome Ionic 2 is.

I had hoped to be falling over myself to tell you about how quickly it was to get started, the incredible workflow, the intuitiveness of the framework, the terseness of the code.

Instead, what follows is a moan colored by disappointment and frustration from using Ionic 2, Typescript & Angular 2.

Ionic, Angular 2 And Typescript

Ionic is a cross platform mobile app framework. It uses Html & JavaScript (Angular) to create mobile apps that can be run on Android and/or IOs.

Ionic 2 is more of the same, but with a few notable differences. You will, inevitably, be using Angular 2 and Typescript to create your apps.

Angular is a JavaScript framework which makes creating single page web apps easier than just using plain JavaScript. Angular 2 is google’s second attempt at the framework, and a significant departure from the first.

Typescript is a superscript of JavaScript. It allows you to write tidier, more elegant, maintainable, JavaScript.

Now, both Angular 2 and Typescript are great in my book, but they are ‘new’. This leads me to my first gripe about the Ionic 2 stack, the learning curve:

Gripe 1: The Learning Curve

There are several steep learning curves involved with the Ionic 2 stack.

Typescript is a big departure from vanilla JavaScript. The syntax should not be alien to anyone who currently uses ES6/ES2015 but it will still take some getting used to.

The other steep learning curve is with Angular 2. Although the syntax in Angular 2 is more elegant and easier to learn than Angular 1, they both have steep learning curves.

Also, there are also changes with Ionic 2 which are a departure from Ionic 1 which you must learn. However these changes are not as severe as the changes in the Angular 2 and using Typescript.

Gripe 2: Beta

Ionic 2 is a beta.

Angular 2 is just out of beta.

Typescript is relatively new.

If you regularly program via stack overflow ‘copy and paste’ you will be frustrated by the lack of help and the ‘out-of-date’ help that you find.

Gripe 3: The Showstopper: Transpiling & Bundling

Ok, here’s the thing that’s eating me most: transpiling & bundling.

This is not a gripe with Ionic 2 per se; it is a gripe about modern web development.

Before, with Ionic 1, using the web tools to debug your app, feedback was instant. You made a change to your code and the changes were reflected immediately on screen.

Now, with Ionic 2, when you save, all your code needs compiled from typescript to something the browser can understand. On top of that, Ionic 2 now bundles your application for use in the browser every time a change, no matter how small, is made.

The upshot, is that when you save even a small project, the wait before you get feedback is greatly increased. I have spent my time waiting 30 seconds upwards waiting for a simple change in my JavaScript to take effect on screen!.

(Maybe I am using it wrong? Please, please, correct me if I am.)

This is a BIG problem for me. The thing I liked most about Ionic 1 (and web development in general) was how immediate the feedback loop was.

Things were so much more productive in the web world compared to using compiled languages. Working with a scripting language had its issues, but it felt so productive not having to wait on the compiler to finish.

Transpilers & bundlers have added a compile step to web development which affects productivity.

Largely due to my opinion on the frustrating feedback loop with Ionic 2, I will not be using Ionic 2 for new projects yet.

Sure, there are a lot of benefits from using Typescript to create a maintainable code base and productivity gains from using using Angular 2, but these are currently outweighed by the negatives.

Ionic 1 is still more productive for me and I will be continuing to use it in the near future.

As always, this is just my opinion. If I am missing something obvious, or have got this horribly wrong, please correct me in the comments below….

Code Run Through

This post follows on from posts about hybrid mobile app development using the new WordPress REST API and Ionic: Part 1, Part 2

What follows is a run through of the import parts of the code, for the application shown in Part 2.

By now you should have a [very basic] Ionic app running in your browser. The app will allow you to:

• Log into your remote, WordPress REST API enabled, website.
• Make a post from the mobile app to your WordPress site.

Pretty simple stuff.

The code that performs the magic is pretty simple too.

*I have created both an Ionic 1 and Ionic 3 repo for the App code. However, below I describe the structure for the Ionic 1 repo only.

Ionic Project Structure (Ionic 1)

Ionic Folder Stucture

As you can see from the folder structure below there are quite a few folders in our Ionic App.

However, the important files and folders are as follows:

• www/js: This folder contains all the Javascript code for the app.
• www/js/app.js: The main entry point for the app. Things like routes are configured here.
• www/js/controllers.js: The logic behind what happens on the screens is placed inside this file e.g. what should happen when the login button is pressed.
• www/js/services.js: Reusable services to perform business logic e.g. authentication and connection to WordPress
• www/templates: This folder contains the html for screen in the app.
• ionic.config.json: This file contains configuration options for the app. In Part 2, we changed a setting in this file to point to our WordPress site.

The folder structure is much like any other Angular application, so we will head straight to the code to see what the key lines are:

App.js

In app.js we configure the Angular app before it runs.

In this file we state:

• What our routes are i.e. what are the urls for our app, what html should be loaded and what Javascript files should control the html.

// Ionic uses AngularUI Router which uses the concept of states
// Learn more here: https://github.com/angular-ui/ui-router
// Set up the various states which the app can be in.
// Each state's controller can be found in controllers.js
 $stateProvider
 .state('login', {
 url: '/login',
 templateUrl: 'templates/login.html',
 controller: 'LoginCtrl'
 })
 .state('report', {
 url: '/report',
 templateUrl: 'templates/report.html',
 controller: 'ReportCtrl'
 })
 // if none of the above states are matched, use this as the fallback
 $urlRouterProvider.otherwise('/login');

• Any injectors/interceptors we want to use. We will see later that, when we get an authentication token, we want this to be automatically passed to every subsequent HTTP call we make. An interceptor is a way of achieving this.

Controllers.js

The controllers.js file contains the logic for the Login and Report screens.

The LoginController has one function: try to login when the user clicks the Login button

$scope.login = function () {
 // contact our login service with the data from the username and password fields
 LoginService.loginUser($scope.data.username, $scope.data.password).then(function (data) {
 // if it is a success, go to the Report screen
 $state.go('report');
 }, function (data) {
 // if there is an error pop it up onscreen
 var alertPopup = $ionicPopup.alert({
 title: 'Login failed!',
 template: 'Please check your credentials!'
 });
 });
 }

The ReportController has a single function too: try to post the score and report data to WordPress (though our WordPress service).

$scope.createReport = function () {
 // show a saving... message while we contact the service
 $ionicLoading.show({
 template: 'Saving...'
 });
 // pass through the values from the score and report fields to the service
 WordPressService.createReport($scope.data.score, $scope.data.report).then(success, failure);
 }

Both controllers are very simple, it is in the Services.js where the real work is performed.

Services.js

The services.js is the work horse of the app, it contains 3x services:

• LoginService. This service is used to contact a WordPress REST API end point and request a JWT authentication token.

// the important bit, contact the end point and ask for a token
 $http.post('/server/wp-json/jwt-auth/v1/token', data).error(function (error) {
 failure(error);
 }).success(function (data) {
 // you are now logged in, save to session storage, the auth interceptor will pick up
 // and add to each request
 $window.sessionStorage.token = data.token;
 success(data);
 });

• WordPressService. The WordPress service contacts our WordPress REST Api and tries to create a post.

var data = {
 title: score,
 excerpt: report,
 content: report,
 status: 'publish'
 };
 // the important bit, make a request to the server to create a new post
 // The Authentication header will be added to the request automatically by our Interceptor service
 $http.post('/server/wp-json/wp/v2/posts', data).error(function (error) {
 deferred.reject(error);
 }).success(function (data) {
 deferred.resolve(data);
 });

• AuthInterceptor. The interceptor service checks localstorage to see if we were given a token, if we have, then it adds it to every Http request.

request: function (config) {
 config.headers = config.headers || {};
 //if there is a token, add it to the request
 if ($window.sessionStorage.token) {
 config.headers.Authorization = 'Bearer ' + $window.sessionStorage.token;
 }
 return config;
 }

Any Questions?

That’s basically all there is to it.

If you have any questions, or any amendments that I can make to the Github repo, then please comment below….