DevTut

# RxSwift

# Disposing

After the subscription was created, it is important to manage its correct deallocation.

The docs told us that

If a sequence terminates in finite time, not calling dispose or not using addDisposableTo(disposeBag) won't cause any permanent resource leaks. However, those resources will be used until the sequence completes, either by finishing production of elements or returning an error.

There are two ways of deallocate resources.

  1. Using disposeBags and addDisposableTo operator.
  2. Using takeUntil operator.

In the first case you manually pass the subscription to the DisposeBag object, which correctly clears all taken memory.

let bag = DisposeBag()
Observable.just(1).subscribeNext { 
    print($0)
}.addDisposableTo(bag)

You don't actually need to create DisposeBags in every class that you create, just take a look at RxSwift Community's project named NSObject+Rx (opens new window). Using the framework the code above can be rewritten as follows:

Observable.just(1).subscribeNext { 
    print($0)
}.addDisposableTo(rx_disposeBag)

In the second case, if the subscription time coincides with the self object lifetime, it is possible to implement disposing using takeUntil(rx_deallocated):

let _ = sequence
    .takeUntil(rx_deallocated)
    .subscribe {
        print($0)
    }

# RxSwift basics

FRP, or Functional Reactive Programming, has some basic terms which you need to know.

Every piece of data can be represented as Observable, which is an asynchronous data stream. The power of FRP is in representation synchronous and asynchronous events as streams, Observables, and providing the same interface to work with it.

Usually Observable holds several (or none) events that holds the date - .Next events, and then it can be terminated successfully (.Success) or with an error (.Error).

Let's take a look at following marble diagram:

--(1)--(2)--(3)|-->

In this example there is a stream of Int values. As time moves forward, three .Next events occurred, and then the stream terminated successfully.

--X->

The diagram above shows a case where no data was emitted and .Error event terminates the Observable.

Before we move on, there are some useful resources:

  1. RxSwift (opens new window). Look at examples, read docs and getting started.
  2. RxSwift Slack room (opens new window) has a few channels for education problem solving.
  3. Play around with RxMarbles (opens new window) to know what operator does, and which is the most useful in your case.
  4. Take a look on this example (opens new window), explore the code by yourself.

# Creating observables

RxSwift offers many ways to create an Observable, let's take a look:

import RxSwift

let intObservale = Observable.just(123) // Observable<Int>
let stringObservale = Observable.just("RxSwift") // Observable<String>
let doubleObservale = Observable.just(3.14) // Observable<Double>

So, the observables are created. They holds just one value and then terminates with success. Nevertheless, nothing happening after it was created. Why?

There are two steps in working with Observables: you observe something to create a stream and then you subscribe to the stream or bind it to something to interact with it.

Observable.just(12).subscribe {
    print($0)
}

The console will print:

.Next(12)
.Completed()

And if I interested only in working with data, which take place in .Next events, I'd use subscribeNext operator:

Observable.just(12).subscribeNext {
    print($0) // prints "12" now
}

If I want create an observable of many values, I use different operators:

Observable.of(1,2,3,4,5).subscribeNext {
    print($0)
}
// 1
// 2
// 3
// 4
// 5

// I can represent existing data types as Observables also:
[1,2,3,4,5].asObservable().subscribeNext { 
    print($0) 
}
// result is the same as before.

And finally, maybe I want an Observable that does some work. For example, it is convenient to wrap a network operation into Observable<SomeResultType>. Let's take a look of do one can achieve this:

Observable.create { observer in    // create an Observable ...
    MyNetworkService.doSomeWorkWithCompletion { (result, error) in
        if let e = error {
            observer.onError(e)    // ..that either holds an error
        } else {
            observer.onNext(result) // ..or emits the data
            observer.onCompleted()  // ..and terminates successfully.
        }
    }
    return NopDisposable.instance // here you can manually free any resources
                                //in case if this observable is being disposed.
}

# Bindings

Observable.combineLatest(firstName.rx_text, lastName.rx_text) { $0 + " " + $1 }
.map { "Greetings, \($0)" }
.bindTo(greetingLabel.rx_text)

Using the combineLatest operator every time an item is emitted by either of two Observables, combine the latest item emitted by each Observable. So in this way we combine the result of the two UITextField's creating a new message with the text "Greetings, \($0)" using string interpolation to later bind to the text of a UILabel.

We can bind data to any UITableView and UICollectionView in an very easy way:

viewModel
.rows
.bindTo(resultsTableView.rx_itemsWithCellIdentifier("WikipediaSearchCell", cellType: WikipediaSearchCell.self)) { (_, viewModel, cell) in
    cell.title = viewModel.title
    cell.url = viewModel.url
}
.addDisposableTo(disposeBag)

That’s an Rx wrapper around the cellForRowAtIndexPath data source method. And also Rx takes care of the implementation of the numberOfRowsAtIndexPath, which is a required method in a traditional sense, but you don’t have to implement it here, it’s taken care of.

# RxCocoa and ControlEvents

RxSwift provides not only the ways to control your data, but to represent user actions in a reactive way also.

RxCocoa contains everything you need. It wraps most of the UI components' properties into Observables, but not really. There are some upgraded Observables called ControlEvents (which represent events) and ControlProperties (which represent properties, surprise!). These things holds Observable streams under the hood, but also have some nuances:

  • It never fails, so no errors.
  • It will Complete sequence on control being deallocated.
  • It delivers events on the main thread (MainScheduler.instance).

Basically, you can work with them as usual:

button.rx_tap.subscribeNext { _ in   // control event
    print("User tapped the button!")
}.addDisposableTo(bag)

textField.rx_text.subscribeNext { text in // control property
    print("The textfield contains: \(text)")
}.addDisposableTo(bag)
// notice that ControlProperty generates .Next event on subscription
// In this case, the log will display 
// "The textfield contains: "
// at the very start of the app.

This is very important to use: as long as you use Rx, forget about the @IBAction stuff, everything you need you can bind and configure at once. For example, viewDidLoad method of your view controller is a good candidate to describe how the UI-components work.

Ok, another example: suppose we have a textfield, a button, and a label. We want to validate text in the textfield when we tap the button, and display the results in the label. Yep, seems like an another validate-email task, huh?

First of all, we grab the button.rx_tap ControlEvent:

----()-----------------------()----->

Here empty parenthesis show user taps. Next, we take what's written in the textField with withLatestFrom operator (take a look at it here (opens new window), imagine that upper stream represents user taps, bottom one represents text in the text field).

button.rx_tap.withLatestFrom(textField.rx_text)

----("")--------------------("123")--->
//  ^ tap   ^ i wrote 123    ^ tap

Nice, we have a stream of strings to validate, emitted only when we need to validate.

Any Observable has such familiar operators as map or filter, we'll take map to validate the text. Create validateEmail function yourself, use any regex you want.

button.rx_tap                                // ControlEvent<Void>
        .withLatestFrom(textField.rx_text)   // Observable<String>
        .map(validateEmail)                  // Observable<Bool>
        .map { (isCorrect) in
            return isCorrect ? "Email is correct" : "Input the correct one, please"
        }                                    // Observable<String>
        .bindTo(label.rx_text)              
        .addDisposableTo(bag)

Done! If you need more custom logic (like showing error views in case of error, making a transition to another screen on success...), just subscribe to the final Bool stream and write it there.

← Swift Package Manager →