In my current Android project I have a dialog that retrieves a list of objects from a webservice and show these in a list. It has a problem though. The webservice (outside my control) is not the fastest, so the process takes a while and often the user changes the orientation of the device while this process is ongoing. For now the orientation change results in the original webservice call to be cancelled and a new one is created, but this is not how it should be done I know. I would like the dialog to able to continue the loading from the webservice when an orientation change happens, but I just can't get my head around how to do this. How is it possible to hook into an ongoing call and display this state in the view (the dialog)? Any recommendations are appreciated. I've played around with Android Architecture Components and kotlins sealed classes saving the viewstate in a livedata object that the view observes, but have not found a solution that I like.
I believe a lot of developers use RxJava for this kind of issue. Is this my only option?
Btw. at the moment I use MVP architecture in my project.
Edit
This is where I cancel the job - if the listener is null its cancelled (I'm using kotlin coroutines):
override fun getWorklist() {
job = onWorkerThread {
val result = repository.getResult().awaitResult()
onMainThread {
when (result) {
is Result.Ok -> listener?.onResult(result.getOrDefault(emptyList())) :? job.cancel()
// Any HTTP error
is Result.Error -> listener?.onHttpError(result.exception) :? job.cancel()
// Exception while request invocation
is Result.Exception -> listener?.onException(result.exception) :? job.cancel()
}
}
}
}
Edit 2
I've tried controlling the viewstate using this Android Arch ViewModel, but the MediatorLiveData object isn't triggered on changes of the viewstate object. And the view stays in Loading state:
class MainModel : ViewModel() {
private var job: Job? = null
val viewState = MutableLiveData<MainViewState>().apply { value = Loading("Warsaw") }
val dataGetter = MediatorLiveData<MainViewState>().apply {
addSource(viewState, {
Log.d("MainModel", "Viewstate is: " + viewState.value.toString() + ")...")
when(it) {
is Loading -> {
Log.d("MainModel", "Launching coroutine...")
job = launch(UI) {
Log.d("MainModel", "Loading...")
val items = Repository.getWorklist()
Log.d("MainModel", "Charts retrieved...")
this#MainModel.viewState.postValue(Success(it.items))
Log.d("MainModel", "Posted update to viewstate...")
}
}
}
})
}
override fun onCleared() {
Log.d("MainModel", "Clearing ViewModel")
job?.cancel()
}
}
interface ViewState
sealed class MainViewState : ViewState
class Loading() : MainViewState()
class Error(val error: String) : MainViewState()
class Success(val items: List<WorklistItem>) : MainViewState()
Related
This init block is in my ViewModel:
init {
viewModelScope.launch {
userRepository.login()
userRepository.user.collect {
_uiState.value = UiState.Success(it)
}
}
}
This is very similar to what's actually written on the app, but even this simple example doesn't work. After userRepository.login(), user which is a SharedFlow emits a new user state. This latest value DOES get collected within this collect function shown above, but when emitting a new uiState containing the result, the view does not get such update.
val uiState by viewModel.uiState.collectAsStateWithLifecycle()
Doing this for some reason, does not work. I suspect the issue is related to the lifecycle of the viewmodel, because when I treat the viewmodel as a singleton, this doesn't happen. It happens only when the viewmodel gets destroyed and then created a 2nd (or more) time(s).
What I'm trying to achieve is that the screen containing the view model is aware of the user state. Meaning that when I navigate to the screen, I want it to collect the latest user state, and then decide which content to show.
I also realize this is not the best pattern, most likely. I'm currently looking into a solution that holds the User as part of the app state and collecting per screen (given that it basically changes all or many screens and functionalities) so if you have any resources on an example on such implementation I'd be thankful. But I can't get my head around why this current implementation doesn't work so any light shed on the situation is much appreciated.
EDIT
This is what I have in mind for the repository
private val _user = MutableSharedFlow<User>()
override val user: Flow<User> = _user
override suspend fun login() {
delay(2000)
_user.emit(LoggedUser.aLoggedUser())
}
override suspend fun logout() {
delay(2000)
_user.emit(GuestUser)
}
For your case better to use this pattern:
ViewModel class:
sealed interface UserUiState {
object NotLoggedIn : UserUiState
object Error : UserUiState
data class LoggedIn(val user: User) : UserUiState
}
class MyViewModel #Inject constructor(
userRepository: UserRepository
) : ViewModel() {
val userUiState = userRepository.login()
.map { user ->
if (user != null)
UserUiState.LoggedIn(user)
else
UserUiState.Error
}
.stateIn(
scope = viewModelScope,
started = SharingStarted.WhileSubscribed(5_000),
initialValue = UserUiState.NotLoggedIn
)
}
Repository class:
class UserRepository {
fun login(): Flow<User?> = flow {
val user = TODO("Your code to get user")
if (isSuccess) {
emit(user)
} else {
emit(null)
}
}
}
Your screen Composable:
#Composable
fun Screen() {
val userUiState by viewModel.userUiState.collectAsStateWithLifecycle()
when (userUiState) {
is UserUiState.LoggedIn -> { TODO("Success code") }
UserUiState.NotLoggedIn -> { TODO("Waiting for login code") }
UserUiState.Error -> { TODO("Error display code") }
}
}
How it works: login() in repository returns autorized user flow which can be used in ViewModel. I use UserUiState sealed class to handle possible user states. And then I convert User value in map {} to UserUiState to display it in the UI Layer. Then Flow of UserUiState needs to be converted to StateFlow to obtain it from the Composable function, so I made stateIn.
And of course, this will solve your problem
Tell me in the comments if I got something wrong or if the code does not meet your expectations
Note: SharedFlow and StateFlow are not used in the Data Layer like you do.
EDIT:
You can emiting flow like this if you are working with network:
val user = flow of {
while (true) {
// network call to get user
delay(2000)
}
}
If you use Room you can do this in your dao.
#Query(TODO("get actual user query"))
fun getUser(): Flow<User>
It is a better way and it recommended by android developers YouTube channel
My repo has the following function:
override fun getTopRatedMoviesStream(): Flow<List<Movie>>
I have the following Result wrapper:
sealed interface Result<out T> {
data class Success<T>(val data: T) : Result<T>
data class Error(val exception: Throwable? = null) : Result<Nothing>
object Loading : Result<Nothing>
}
fun <T> Flow<T>.asResult(): Flow<Result<T>> {
return this
.map<T, Result<T>> {
Result.Success(it)
}
.onStart { emit(Result.Loading) }
.catch { emit(Result.Error(it)) }
}
And finally, my ViewModel has the following UiState logic:
data class HomeUiState(
val topRatedMovies: TopRatedMoviesUiState,
val isRefreshing: Boolean
)
#Immutable
sealed interface TopRatedMoviesUiState {
data class Success(val movies: List<Movie>) : TopRatedMoviesUiState
object Error : TopRatedMoviesUiState
object Loading : TopRatedMoviesUiState
}
class HomeViewModel #Inject constructor(
private val movieRepository: MovieRepository
) : ViewModel() {
private val topRatedMovies: Flow<Result<List<Movie>>> =
movieRepository.getTopRatedMoviesStream().asResult()
private val isRefreshing = MutableStateFlow(false)
val uiState: StateFlow<HomeUiState> = combine(
topRatedMovies,
isRefreshing
) { topRatedResult, refreshing ->
val topRated: TopRatedMoviesUiState = when (topRatedResult) {
is Result.Success -> TopRatedMoviesUiState.Success(topRatedResult.data)
is Result.Loading -> TopRatedMoviesUiState.Loading
is Result.Error -> TopRatedMoviesUiState.Error
}
HomeUiState(
topRated,
refreshing
)
}
.stateIn(
scope = viewModelScope,
started = WhileUiSubscribed,
initialValue = HomeUiState(
TopRatedMoviesUiState.Loading,
isRefreshing = false
)
)
fun onRefresh() {
viewModelScope.launch(exceptionHandler) {
movieRepository.refreshTopRated()
isRefreshing.emit(true)
isRefreshing.emit(false)
}
}
The issue is the TopRatedMoviesUiState.Loading state is only emitted once on initial load but not when user pulls to refresh and new data is emitted in movieRepository.getTopRatedMoviesStream(). I understand that it is because .onStart only emits first time the Flow is subscribed to.
Do I somehow resubscribe to Flow when refresh is performed? Refresh does not always return new data from repo so how in this case, how do I avoid duplicate emission?
You emit TopRatedMoviesUiState.Loading in onStart. So what you describe is totally expected. Loading is emitted when the stream starts. I.e. when you start collecting. In your case by stateIn.
Looking at it from another perspective, how should your result wrapper know that the repository is currently loading?
And that's also the answer. Only two places in your code know that you are reloading.
Either subscribe to a fresh flow whenever you call refreshTopRated and complete the flow after emitting the final result.
Or emit a loading state right from the repository before you start loading.
I'd prefer the later one.
Neither solution will save you from emitting the same result again and again. For that, you'd need a new state like 'Unchanged' that your repository emits when it finds no new data. But please evaluate if this optimization is required. What is the cost of an extra emission?
That being said, here are some more questions, challenging your code. Hopefully guiding you to a solid implementation:
Why do you need the isRefreshing StateFlow? Is there a difference in the UI wether you initially load or wether you refresh?
Why do you need topRated to be a StateFlow? Wouldn't a regular Flow do?
Emitting two items to a StateFlow in succession (i.e. isRefreshing.emit(true); isRefreshing.emit(false) ) might loose the first emission [1]. As for state, only the most recent value is relevant. States are not events!
[1]: StateFlow has a replay buffer of 1 and a buffer overflow policy of DROP_OLDEST. See State flow is a shared flow in kotlinx-coroutines-core/kotlinx.coroutines.flow/StateFlow
Okay so I've been using StateFlow with Room database for a while. Now I have one common case. At the start of my app I have a logic that if ROOM database is empty I should show an EmptyContent(), otherwise I will show the ListContent() from ROOM database.
Now every time I launch the app, I'm always getting that EmptyContent() shown for a HALF a second maybe, and then the ListContent() is displayed. After that when I'm using the app everything works normal. But at that app launch time, while ROOM database is working I guess, that EmptyContent() is shown for just a small amount of period (Because my StateFlow default value is an empty list), and after that the actual LIST from Database is displayed.
Now I have one solution for that, to just use delay() function inside a Coroutine, to wait for example 200MS and then trigger the function for reading the DATABASE, because those 200MS are enough for ROOM database to actually get the value and update my STATE FLOW variable with the actual data instead of using that StateFlow default value for a half second at the beginning.
Is that a good solution, I must ask? Because I'm using coroutine, the thread is not blocked, and I'm just waiting until ROOM database updates my STATE FLOW variable the second time.
#Composable
fun displayContent(
tasks: List<ToDoTask>,
ListContent: #Composable () -> Unit
) {
val scope = rememberCoroutineScope()
var counter by remember { mutableStateOf(0)}
LaunchedEffect(Unit){
scope.launch {
delay(200)
counter = 1
}
}
if(counter == 1){
if (tasks.isNotEmpty()) {
ListContent()
} else {
EmptyContent()
}
}
}
My suggestion would be map your expected states.
For instance:
sealed class RequestState<out T> {
object Idle : RequestState<Nothing>()
object Loading : RequestState<Nothing>()
data class Success<T>(val data: T) : RequestState<T>()
data class Error(
val t: Throwable,
var consumed: Boolean = false
) : RequestState<Nothing>()
}
And your function would be something like:
#Composable
fun YourScreen() {
val requestState = viewModel.screenData.collectAsState()
when (requestState) {
is Idle ->
// This is the default state, do nothing
is Loading ->
// Display some progress indicator
is Success ->
YourListScreen(requestState.data) // Show the list
is Error ->
// Display an error.
}
LaunchedEffect(Unit) {
viewModel.loadData()
}
}
Of course, in your view model you must emit these values properly...
class YourView: ViewModel() {
private val _screenData =
MutableStateFlow<RequestState<List<ToDoTask>>>(RequestState.Idle)
val screenDatae: StateFlow<RequestState<List<ToDoTask>>> = _screenData
fun loadData() {
_screenData.value = Loading
try {
// load the data from database
_screenData.value = Success(yourLoadedData)
} catch (e: Exception) {
_screenData.value = Error(e)
}
}
}
I have a usecase:
Open app + disable network -> display error
Exit app, then enable network, then open app again
Expected: app load data
Actual: app display error that meaning state error cached, liveData is not emit
Repository class
class CategoryRepository(
private val api: ApiService,
private val dao: CategoryDao
) {
val categories: LiveData<Resource<List<Category>>> = liveData {
emit(Resource.loading(null))
try {
val data = api.getCategories().result
dao.insert(data)
emit(Resource.success(data))
} catch (e: Exception) {
val data = dao.getCategories().value
if (!data.isNullOrEmpty()) {
emit(Resource.success(data))
} else {
val ex = handleException(e)
emit(Resource.error(ex, null))
}
}
}
}
ViewModel class
class CategoryListViewModel(
private val repository: CategoryRepository
): ViewModel() {
val categories = repository.categories
}
Fragment class where LiveDate obsever
viewModel.apply {
categories.observe(viewLifecycleOwner, Observer {
// live data only trigger first time, when exit app then open again, live data not trigger
})
}
can you help me explain why live data not trigger in this usecase and how to fix? Thankyou so much
Update
I have resolved the above problem by replace val categories by func categories() at repository class. However, I don't understand and can't explain why it works properly with func but not val.
Why does this happen? This happens because your ViewModel has not been killed yet. The ViewModel on cleared() is called when the Fragment is destroyed. In your case your app is not killed and LiveData would just emit the latest event already set. I don't think this is a case to use liveData builder. Just execute the method in the ViewModel when your Fragment gets in onResume():
override fun onResume(){
viewModel.checkData()
super.onResume()
}
// in the viewmodel
fun checkData(){
_yourMutableLiveData.value = Resource.loading(null)
try {
val data = repository.getCategories()
repository.insert(data)
_yourMutableLiveData.value = Resource.success(data)
} catch (e: Exception) {
val data = repository.getCategories()
if (!data.isNullOrEmpty()) {
_yourMutableLiveData.value = Resource.success(data)
} else {
val ex = handleException(e)
_yourMutableLiveData.value = Resource.error(ex,null)
}
}
}
Not sure if that would work, but you can try to add the listener directly in onResume() but careful with the instantiation of the ViewModel.
Small advice, if you don't need a value like in Resource.loading(null) just use a sealed class with object
UPDATE
Regarding your question that you ask why it works with a function and not with a variable, if you call that method in onResume it will get executed again. That's the difference. Check the Fragment or Activity lifecycle before jumping to the ViewModel stuff.
I am using MVVM in my app. When you enter a query and click search button, the chain is as follows: Fragment -> ViewModel -> Repository -> API -> Client. The client is where HTTP requests are made. But there is one thing here, the client needs to make a call and get a key from the server at initialization. Therefore, to prevent any call before it this first call completes, I need to be able to observe it from Fragment so that I can disable search button. Since each component in the chain can communicate with adjacent components, all components should have a state.
I am thinking to implement a StatefulComponent class and make all components to extend it:
open class StatefulComponent protected constructor() {
enum class State {
CREATED, LOADING, LOADED, FAILED
}
private val currentState = MutableLiveData(State.CREATED)
fun setState(newState: State) {
currentState.value = newState
}
val state: LiveData<State> = currentState
val isLoaded: Boolean = currentState.value == State.LOADED
val isFailed: Boolean = currentState.value == State.FAILED
val isCompleted: Boolean = isLoaded || isFailed
}
The idea is that each component observers the next one and updates itself accordingly. However, this is not possible for ViewModel since it is already extending ViewModel super class.
How can I implement a solution for this problem?
The most common approach is to use sealed class as your state, so you have any paramaters as you want on each state case.
sealed class MyState {
object Loading : MyState()
data class Loaded(data: Data) : MyState()
data class Failed(message: String) : MyState()
}
On your viewmodel you will have only 1 livedata
class MyViewModel : ViewModel() {
private val _state = MutableLiveData<MyState>()
val state: LiveData<MyState> = _state
fun load() {
_state.postCall(Loading)
repo.loadSomeData(onData = { data ->
_state.postCall(Loaded(data))
}, onError = { error -> _state.postCall(Failed(error.message)) })
}
// coroutines approach
suspend fun loadSuspend() {
_state.postCall(Loading)
try {
_state.postCall(Loaded(repo.loadSomeDataSupend()))
} catch(e: Exception) {
_state.postCall(Failed(e.message))
}
}
}
And on the fragment, just observe the state
class MyFragment : Fragment() {
...
onViewCreated() {
viewModel.state.observer(Observer {
when (state) {
// auto casts to each state
Loading -> { button.isEnabled = false }
is Loaded -> { ... }
is Failed -> { ... }
}
}
)
}
}
As João Gouveia mentioned, we can make stateful components quite easily using kotlin's sealed classes.
But to make it further more useful, we can introduce Generics! So, our state class becomes StatefulData<T> which you can use pretty much anywhere (LiveData, Flows, or even in Callbacks).
sealed class StatefulData<out T : Any> {
data class Success<T : Any>(val result : T) : StatefulData<T>()
data class Error(val msg : String) : StatefulData<Nothing>()
object Loading : StatefulData<Nothing>()
}
I've wrote an article fully explaining this particular implementation here
https://naingaungluu.medium.com/stateful-data-on-android-with-sealed-classes-and-kotlin-flow-33e2537ccf55
If you are using the composable ... You can use produce state
#Composable
fun PokemonDetailScreen(
viewModel: PokemonDetailVm = hiltViewModel()
) {
/**
* This takes a initial state and with that we get a coroutine scope where we can call a API and assign the data into the value
*/
val pokemonInfo = produceState<Resource<Pokemon>>(initialValue = Resource.Loading()) {
value = viewModel.getPokemonInfo(pokemonName)
}.value
}