Develop Reference

Where to put reused values used for every emission in Observable/ Flowable?

I have some expensive operations that only need to be performed once (e.g. load/ download large files, load large ML models, or calculate optimized data structure based on some other data). I want to use this for every value the Observable/ Flowable generates:
The following code works, but it runs heavyProcessing() and heavyProcessing2() on the caller's thread. In my case, I can't choose what my callers thread (its the main thread because I am using WorkManager's RxWorker, which calls createWork from main). Therefore, start blocks the main thread. How do I get heavyProcessing to be performed in the background with RxJava and also available to the subsequent RxJava chain?
fun start(): Observable<Unit> {
val heavy = heavyProcessing() // the heavy value i want to use everywhere!
val anotherHeavyObject = heavyProcessing2()
val items = Observable.fromIterable(listOfHundredsOfItems)
.map { doSomeWork(it, heavy) }
.map { doSomeWork(it, anotherHeavyObject) }
}
My attempts has so far not worked:
Create a wrapper around the existing function: The issue with this code is the Observable returned by start() does not get observed, so the doSomeWork doesn't actually get done. I only know this because I put breakpoints in at doSomeWork, and it never gets called.
fun startInBackground(): Single<Unit> {
return Single.fromCallable {
start()
}
}
I've been trying to find ways of 'unnesting' the inner Observable (inside the Single), as that's probably the issue here. The inner Observable is not being observed.
This RxJava stuff is very unintuitive even after reading the guide

Yes, it was related to Deferred-dependent. The example in the docs state:
Sometimes, there is an implicit data dependency between the previous sequence and the new sequence that, for some reason, was not flowing through the "regular channels". One would be inclined to write such continuations as follows:
AtomicInteger count = new AtomicInteger();
Observable.range(1, 10)
.doOnNext(ignored -> count.incrementAndGet())
.ignoreElements()
.andThen(Single.defer(() -> Single.just(count.get())))
.subscribe(System.out::println);
Actually, all I needed the caller to do is:
Single.defer { start() }.map { doMoreWork() }
instead of
start().map { doMoreWork() }

Is it possible to implement an operator like delay but that also delays errors?

I'm trying for some time now to implement an extension function (just becuse it's easier to me) that is capable of delaying both normal item emissions and errors. The existing delay operators only delays normal item emissions, errors are delivered ASAP.
For context, I'm trying to immitate an Android LiveData's behavior (kinda). LiveDatas are a observable pattern implementation that is lifecycle aware. Their observers are only notified if they are in a state where they can process that emission. If they are not ready, the emission is cached in the livedata and delivered as soon as they become ready.
I created a BehaviourSubject that emits the state of my Activities and Fragments when it changes. With that I created a delay operator like this:
fun <T> Flowable<T>.delayUntilActive(): Flowable<T> = delay { lifecycleSubject.toFlowable(BackpressureStrategy.LATEST).filter { it.isActive } }
and then use it like this
myUseCase.getFlowable(Unit)
.map { it.map { it.toDisplayModel() } }
.delayUntilActive()
.subscribe({
view.displaySomethings(
}, { }).addTo(disposables)
So even if myUseCase emits when the view is not ready to display somethings, the emission won't reach onNext() until the view does become ready. The problem is that I also want the view to displayError() when onError is triggered, but that too is lifecycle sensitive. If the view isn't ready, the app will crash.
So I'm looking for a way to delay both emissions and errors (onComplete would be good too). Is this possible?
I tried some things with zip, onErrorReturn, delay inside delay, but nothing seemed right. I'd be equally unimpressed if this had a really easy solution I'm overlooking, or is impossible. Any ideas are welcome.
Bonus: any better way to do that for Single and Completable too? currently I'm just converting them to flowable.
Thanks in advance!

You can handle the error via onErrorResumeNext, then taking the same error and delaying it via delaySubscription until your desired signal to emit said error happens:
source
.onErrorResumeNext({ error ->
Observable.error(error)
.delaySubscription(lifecycleSubject.filter { it.Active } )
})

How to keep track of the number of emits in flowable?

Let's say I have a flowable, that some view is subscribed to and it's listening to the changes. I would like to add a custom method based on only the first emit of the flowable, but also keeping the other methods that listen to the changes. What is the best way to approach it?
The naive approach I have is to duplicate the flowable and convert it to Single or Completable to get the results, but it seems redundant.
Thank you.

Use .take(1). BTW also make sure that flowable is shared (otherwise some observers will miss events).

I think you can use share operator for that. Share operator makes a Connectable Observable. And then Connectable Observable publishes items each subscribes.
val o = Flowable.fromArray(1, 2, 3, 4, 5)
.map {
println("heavy operation")
it + it
}
.share() // publish the changes
.subscribeOn(Schedulers.computation()) // for testing. change what you want
o.take(1).subscribe { println("Special work: $it") } // take one
o.subscribe { println("Normal work: $it") }
Result
heavy operation
Special work: 2
Normal work: 2
heavy operation
Normal work: 4
heavy operation
Normal work: 6
heavy operation
Normal work: 8
heavy operation
Normal work: 10

Observable startWith isn't emitted in doOnNext

I have a number of Observables that are used for network requests in my app. Since so much is the same, I apply an Observable transformation to them:
/**
* Creates a transformer that applies the schedulers and error handling for all of the observables in this ViewModel.
*/
private fun applyTransformations(): Observable.Transformer<NetworkState, NetworkState> {
return Observable.Transformer { observable ->
observable
.subscribeOn(Schedulers.io())
.observeOn(AndroidSchedulers.mainThread())
.onErrorReturn { NetworkState.Error(it) }
.doOnNext { publishState(it) }
.startWith(NetworkState.Loading())
}
}
The goals I am trying to accomplish with the above:
Apply consistent schedulers
Handle any error by returning an instance of my sealed class.
Handle any onNext by publishing the state returned by the observable.
Start off by sending a Loading state.
This works mostly fine, but what I've noticed is that while I call startWith and a loading state, it is never actually handled by doOnNext(). In other words, publishState() is never called for my loading state.
Where I set up the observables, I don't bother to add a subscriber, because the doOnNext() above is all that I'll need:
val subscription = repository.getInstagramPhotos(count)
.map { mapIGPhotoResponse(it) }
.compose(applyTransformations())
.subscribe()
If I were to supply a subscriber above, though, it would handle the loading state. It would also handle two onNext() calls - one for the subscriber supplied, and one for the doOnNext in the transform.
Is there a way to modify this startWith call to emit to whatever I've specified in doOnNext? I'm using RxJava 1.
Edit: Just to clarify some more, if I track what's emitted I expect to see two things. Loading -> Success. What I actually see is just Success. If I supply a subscriber to the observable I see Loading -> Success -> Success.

startWith should be before doOnNext.
Rxjava methods, though they look like they use the builder pattern, actually don't. They return a new observable each time an operator is applied. In your case, your doOnNext observable completes before your start with observable, so it's consumer isn't called with what you supply in startWith.
Ideally, you should go with:
observable
.startWith(NetworkState.Loading())
.doOnNext { publishState(it) }
.onErrorReturn { NetworkState.Error(it) }
.subscribeOn(Schedulers.io())
.observeOn(AndroidSchedulers.mainThread())
Also, be careful with subscribing with no Consumer for onEror should it happen. Since you have nothing to consume the onError, RxJava will crash your app since it has nothing to notify for the error. Consider replacing the doOnNext with a Success Consumer in subscribe, and an empty Consumer for the error if you want to ignore it.
Also doOnNext is typically used for side effects, such as logging and the sort, they're more of a convenience than true functional operators.

How to achieve buffering logic with publish sequence in RxJava2?

I want to achieve the following with RxJava and as I may not have enough knowledge in this area would like to have some help :)
I need to create a PublishSubject which would emit events with the following sequence:
Emit 1, 2, 3
Buffer 4 in subscribe's completion if a certain condition is not satisfied (may be a network connection for example or some other condition).
For 5, 6 ... buffer after 4 if the condition is not satisfied yet.
Repeat to emit 4 after some time when the condition is satisfied.
If trying to emit 5,6 and the condition is satisfied, then instead of buffering 5, 6 ... after 4, just emit 4 and then 5, 6, 7 , 8 ...
The last 2 points are necessary because the sequence of emitting is really important, which makes difficulties for me to achieve to this.
I hope I could describe what I want to achieve :)
Findings: After asking this question I've done some findings and achieved the following:
private Observable observable = publishSubject
.observeOn(Schedulers.io())
.map(Manager::callNew)
.doOnError(throwable -> Logger.e(throwable, "Error occurred"))
.retryWhen(throwableObservable -> throwableObservable
.zipWith(Observable.range(1, 10), (n, i) -> i)
.flatMap(retryCount -> {
long retrySeconds = (long) Math.pow(2, retryCount);
Logger.d("The call has been failed retrying in %s seconds. Retry count %s", retrySeconds, retryCount);
return Observable.timer(retrySeconds, TimeUnit.SECONDS)
.doOnNext(aLong -> {
C24Logger.d("Timer was completed. %s", aLong);
})
.doOnComplete(() -> Logger.d("Timer was completed."));
}));
The problem is here with PublishSubject. Because it already has emitted all the items, it emits only new ones for retryWhen. If I use ReplaySubject them it emits also the old completed items too for the new retryWhen re-subscribe, which I do not need anymore.
Is there a way to use the ReplaySubject to remove the completed items from the buffer?

You want to be able to turn buffering on and off, depending upon an external condition. Perhaps the simplest way to do it is use the buffer() operator to continually buffer items based on the condition.
(I have removed stuff from the observer chain)
private Observable observable = publishSubject
.publish( obs -> obs.buffer( obs.filter( v -> externalCondition( v ) ) ) )
.flatMapIterable( bufferedList -> bufferedList )
.subscribe( ... );
The publish() operator allows multiple observer chains to subscribe to the incoming observer chain. The buffer() operator monitors the observable that emits a value only when the external condition is true.
When the external condition is true, buffer() will emit a series of lists with only a single element. When the condition goes false, buffer() starts buffering up the results, and when the condition goes true again, all the buffered items are emitted as a list. The flatMapIterable() step will take each item out of the buffer and emit it separately.