I'm trying to handle JWT-authentication in gRPC on my backend. I can extract the JWT in an interceptor but how do I access it in my service? I think it should be done with a CoroutineContextServerInterceptor but this doesn't work:
val jwtKey: Context.Key<String> = Context.key("jwtKey")
fun main() {
ServerBuilder.forPort(8980).intercept(UserInjector).addService(MyService).build().start().awaitTermination()
}
object UserInjector : CoroutineContextServerInterceptor() {
override fun coroutineContext(call: ServerCall<*, *>, headers: Metadata): CoroutineContext {
val jwtString = headers.get(Metadata.Key.of("jwt", Metadata.ASCII_STRING_MARSHALLER))
println("coroutineContext: $jwtString")
return GrpcContextElement(Context.current().withValue(jwtKey, jwtString))
}
}
object MyService : MyServiceGrpcKt.MyServiceCoroutineImplBase() {
override suspend fun testingJWT(request: Test.MyRequest): Test.MyResponse {
println("testingJWT: ${jwtKey.get()}")
return Test.MyResponse.getDefaultInstance()
}
}
Output:
coroutineContext: something
testingJWT: null
I think you'll need to propagate that in its own coroutine context element.
class JwtElement(val jwtString: String) : CoroutineContext.Element {
companion object Key : CoroutineContext.Key<JwtElement>
override val key: CoroutineContext.Key<JwtElement>
get() = Key
}
object UserInjector : CoroutineContextServerInterceptor() {
override fun coroutineContext(call: ServerCall<*, *>, headers: Metadata): CoroutineContext {
val jwtString = headers.get(Metadata.Key.of("jwt", Metadata.ASCII_STRING_MARSHALLER))
println("coroutineContext: $jwtString")
return JwtElement(jwtString)
}
}
object MyService : MyServiceGrpcKt.MyServiceCoroutineImplBase() {
override suspend fun testingJWT(request: Test.MyRequest): Test.MyResponse {
println("testingJWT: ${coroutineContext[JwtElement]}")
return Test.MyResponse.getDefaultInstance()
}
}
Related
Does anyone know how to program the override function convertForReceive of a custom Multipart.FormData converter?
I want to convert the multipart request to my class with the converter but I don't know how it works.
I have:
Application.kt
install(ContentNegotiation) {
json()
register(ContentType.MultiPart.FormData, CustomMultipartConverter)
}
CustomMultipartConverter
object CustomMultipartConverter: ContentConverter {
override suspend fun convertForReceive(context: PipelineContext<ApplicationReceiveRequest, ApplicationCall>): Any? {
TODO("Not yet implemented")
}
override suspend fun convertForSend(
context: PipelineContext<Any, ApplicationCall>,
contentType: ContentType,
value: Any
): Any? {
TODO("Not yet implemented")
}
}
REQUEST CLASS
class CreatePostRequest(
val text: String,
val image: File? = null
)
ROUTE
route("v1/posts") {
authenticate {
route("create") {
val authJWT = call.authentication.principal as JWTAtuh
val request = call.receive<CreatePostRequest>()
//myCode
call.respond(HttpStatusCode.OK)
}
}
}
You can take SerializationConverter as a reference:
override suspend fun convertForReceive(context: PipelineContext<ApplicationReceiveRequest, ApplicationCall>): Any? {
val request = context.subject
val channel = request.value as? ByteReadChannel ?: return null
val charset = context.call.request.contentCharset() ?: defaultCharset
val serializer = format.serializersModule.serializer(request.typeInfo)
val contentPacket = channel.readRemaining()
return when (format) {
is StringFormat -> format.decodeFromString(serializer, contentPacket.readText(charset))
is BinaryFormat -> format.decodeFromByteArray(serializer, contentPacket.readBytes())
else -> {
contentPacket.discard()
error("Unsupported format $format")
}
}
}
My app using room as a database and retrofit as a network calling api.
i am observing database only as a single source of truth. every thing is working fine. But i am not finding solution of one scenario.
Like for the first time when user open app it do following operations
fetch data from db
fetch data from server
because currently database is empty so it sends empty result to observer which hide progress bar . i want to discard that event and send result to observer when server dump data to database. even server result is empty. so progress bar should always hide once their is confirmation no data exists.
in other words application should always rely on database but if it empty then it should wait until server response and then notify observer.
this is my code
observer
viewModel.characters.observe(viewLifecycleOwner, Observer {
Log.e("status is ", "${it.message} at ${System.currentTimeMillis()}")
when (it.status) {
Resource.Status.SUCCESS -> {
binding.progressBar.visibility = View.GONE
if (!it.data.isNullOrEmpty()) adapter.setItems(ArrayList(it.data))
}
Resource.Status.ERROR -> {
Toast.makeText(requireContext(), it.message, Toast.LENGTH_SHORT).show()
binding.progressBar.visibility = View.GONE
}
Resource.Status.LOADING ->
binding.progressBar.visibility = View.VISIBLE
}
})
ViewModel
#HiltViewModel
class CharactersViewModel #Inject constructor(
private val repository: CharacterRepository
) : ViewModel() {
val characters = repository.getCharacters()
}
Repository
class CharacterRepository #Inject constructor(
private val remoteDataSource: CharacterRemoteDataSource,
private val localDataSource: CharacterDao
) {
fun getCharacters() : LiveData<Resource<List<Character>>> {
return performGetOperation(
databaseQuery = { localDataSource.getAllCharacters() },
networkCall = { remoteDataSource.getCharacters() },
saveCallResult = { localDataSource.insertAll(it.results) }
)
}
}
Utility function for all api and database handling
fun <T, A> performGetOperation(databaseQuery: () -> LiveData<T>,
countQuery: () -> Int,
networkCall: suspend () -> Resource<A>,
saveCallResult: suspend (A) -> Unit): LiveData<Resource<T>> =
liveData(Dispatchers.IO) {
emit(Resource.loading())
val source = databaseQuery().map { Resource.success(it,"database") }.distinctUntilChanged()
emitSource(source)
val responseStatus = networkCall()
if (responseStatus.status == SUCCESS) {
saveCallResult(responseStatus.data!!)
} else if (responseStatus.status == ERROR) {
emit(Resource.error(responseStatus.message!!))
}
}
LocalDataSource
#Dao
interface CharacterDao {
#Query("SELECT * FROM characters")
fun getAllCharacters() : LiveData<List<Character>>
#Insert(onConflict = OnConflictStrategy.REPLACE)
suspend fun insertAll(characters: List<Character>)
#Insert(onConflict = OnConflictStrategy.REPLACE)
suspend fun insert(character: Character)
}
DataSource
class CharacterRemoteDataSource #Inject constructor(
private val characterService: CharacterService
): BaseDataSource() {
suspend fun getCharacters() = getResult { characterService.getAllCharacters() }}
}
Base Data Source
abstract class BaseDataSource {
protected suspend fun <T> getResult(call: suspend () -> Response<T>): Resource<T> {
try {
Log.e("status is", "started")
val response = call()
if (response.isSuccessful) {
val body = response.body()
if (body != null) return Resource.success(body,"server")
}
return error(" ${response.code()} ${response.message()}")
} catch (e: Exception) {
return error(e.message ?: e.toString())
}
}
private fun <T> error(message: String): Resource<T> {
Timber.d(message)
return Resource.error("Network call has failed for a following reason: $message")
}
}
Character Service
interface CharacterService {
#GET("character")
suspend fun getAllCharacters() : Response<CharacterList>
}
Resource
data class Resource<out T>(val status: Status, val data: T?, val message: String?) {
enum class Status {
SUCCESS,
ERROR,
LOADING
}
companion object {
fun <T> success(data: T,message : String): Resource<T> {
return Resource(Status.SUCCESS, data, message)
}
fun <T> error(message: String, data: T? = null): Resource<T> {
return Resource(Status.ERROR, data, message)
}
fun <T> loading(data: T? = null): Resource<T> {
return Resource(Status.LOADING, data, "loading")
}
}
}
CharacterList
data class CharacterList(
val info: Info,
val results: List<Character>
)
What is the best way by that i ignore database if it is empty and wait for server response and then notify observer
I am trying to listen to my ViewModels MutableStateFlow from my FlutterSceneView. But I get the following error when trying to set the listener from the views init:
Suspend function 'listenToBackgroundColor' should be called only from a coroutine or another suspend function
class FlutterSceneView(context: Context, private val viewModel: FlutterSceneViewModelType): PlatformView {
private val context = context
private val sceneView = SceneView(context)
init {
listenToBackgroundColor() // Error here
}
private suspend fun listenToBackgroundColor() {
viewModel.colorFlow.collect {
val newColor = Color.parseColor(it)
sceneView.setBackgroundColor(newColor)
}
}
}
My ViewModel:
interface FlutterSceneViewModelType {
var colorFlow: MutableStateFlow<String>
}
class FlutterSceneViewModel(private val database: Database): FlutterSceneViewModelType, ViewModel() {
override var colorFlow = MutableStateFlow<String>("#FFFFFF")
init {
listenToBackgroundColorFlow()
}
private fun listenToBackgroundColorFlow() {
database.backgroundColorFlow.watch {
colorFlow.value = it.hex
}
}
}
the .watch call is a helper I have added so that this can be exposed to iOS using Kotlin multi-platform, it looks as follows but I can use collect instead if necessary:
fun <T> Flow<T>.asCommonFlow(): CommonFlow<T> = CommonFlow(this)
class CommonFlow<T>(private val origin: Flow<T>) : Flow<T> by origin {
fun watch(block: (T) -> Unit): Closeable {
val job = Job()
onEach {
block(it)
}.launchIn(CoroutineScope(Dispatchers.Main + job))
return object : Closeable {
override fun close() {
job.cancel()
}
}
}
}
I resolved this by using viewModel context:
private fun listenToBackgroundColor() {
viewModel.colorFlow.onEach {
val newColor = Color.parseColor(it)
sceneView.setBackgroundColor(newColor)
}.launchIn(viewModel.viewModelScope)
}
I had to import the following into my ViewModel:
import androidx.lifecycle.ViewModel
import androidx.lifecycle.viewModelScope
from:
implementation("androidx.lifecycle:lifecycle-viewmodel-ktx:2.2.0")
I want to test a method of my ViewModel that collects a Flow. Inside the collector a LiveData object is mutated, which I want to check in the end. This is roughly how the setup looks:
//Outside viewmodel
val f = flow { emit("Test") }.flowOn(Dispatchers.IO)
//Inside viewmodel
val liveData = MutableLiveData<String>()
fun action() {
viewModelScope.launch { privateAction() }
}
suspend fun privateAction() {
f.collect {
liveData.value = it
}
}
When I now call the action() method in my unit test, the test finishes before the flow is collected. This is how the test might look:
#Test
fun example() = runBlockingTest {
viewModel.action()
assertEquals(viewModel.liveData.value, "Test")
}
I am using the TestCoroutineDispatcher via this Junit5 extension and also the instant executor extension for LiveData:
class TestCoroutineDispatcherExtension : BeforeEachCallback, AfterEachCallback, ParameterResolver {
#SuppressLint("NewApi") // Only used in unit tests
override fun supportsParameter(parameterContext: ParameterContext?, extensionContext: ExtensionContext?): Boolean {
return parameterContext?.parameter?.type === testDispatcher.javaClass
}
override fun resolveParameter(parameterContext: ParameterContext?, extensionContext: ExtensionContext?): Any {
return testDispatcher
}
private val testDispatcher = TestCoroutineDispatcher()
override fun beforeEach(context: ExtensionContext?) {
Dispatchers.setMain(testDispatcher)
}
override fun afterEach(context: ExtensionContext?) {
Dispatchers.resetMain()
testDispatcher.cleanupTestCoroutines()
}
}
class InstantExecutorExtension : BeforeEachCallback, AfterEachCallback {
override fun beforeEach(context: ExtensionContext?) {
ArchTaskExecutor.getInstance()
.setDelegate(object : TaskExecutor() {
override fun executeOnDiskIO(runnable: Runnable) = runnable.run()
override fun postToMainThread(runnable: Runnable) = runnable.run()
override fun isMainThread(): Boolean = true
})
}
override fun afterEach(context: ExtensionContext?) {
ArchTaskExecutor.getInstance().setDelegate(null)
}
}
You can try either,
fun action() = viewModelScope.launch { privateAction() }
suspend fun privateAction() {
f.collect {
liveData.value = it
}
}
#Test
fun example() = runBlockingTest {
viewModel.action().join()
assertEquals(viewModel.liveData.value, "Test")
}
or
fun action() {
viewModelScope.launch { privateAction()
}
suspend fun privateAction() {
f.collect {
liveData.value = it
}
}
#Test
fun example() = runBlockingTest {
viewModel.action()
viewModel.viewModelScope.coroutineContext[Job]!!.join()
assertEquals(viewModel.liveData.value, "Test")
}
You could also try this,
suspend fun <T> LiveData<T>.awaitValue(): T? {
return suspendCoroutine { cont ->
val observer = object : Observer<T> {
override fun onChanged(t: T?) {
removeObserver(this)
cont.resume(t)
}
}
observeForever(observer)
}
}
#Test
fun example() = runBlockingTest {
viewModel.action()
assertEquals(viewModel.liveData.awaitValue(), "Test")
}
So what I ended up doing is just passing the Dispatcher to the viewmodel constructor:
class MyViewModel(..., private val dispatcher = Dispatchers.Main)
and then using it like this:
viewModelScope.launch(dispatcher) {}
So now I can override this when I instantiate the ViewModel in my test with a TestCoroutineDispatcher and then advance the time, use testCoroutineDispatcher.runBlockingTest {}, etc.
Having code to get from two different endpoints, but the response json have same scheme and the response handler code are same.
internal interface DataApi {
#GET("/api/api_1")
fun getData_1(#QueryMap params: Map<String, String>?): Call<DataMpdel>
#GET("/api/api_2")
fun getData_2(#QueryMap params: Map<String, String>?): Call<DataMpdel>
}
private lateinit var dataApi: DataApi
init {
dataApi = createDataApi()
}
private fun createDataApi() : DataApi {
val restAdapter = Retrofit.Builder()
.baseUrl(baseUrl)
.client(okHttpClient)
.addConverterFactory(getGsonConverterFactory())
.build()
return restAdapter.create(DataApi::class.java)
}
It could use if/else to call the different endpoint, but that will duplicate the handler code.
Update: create the handler function and in both onResposne() just call that function.
Any suggestion to be better with shared handler code?
private fun fetchFirst(){
if (!UseDataApi_1) {
//
dataApi. getData_1(getQueryParams()).enqueue(object : Callback<DataModel> {
override fun onResponse(call: Call<DataModel>?, response: Response<DataModel>?) {
// same handler code
}
override fun onFailure(call: Call<DataModel>?, throwable: Throwable?) {
// same error handler code
}
})
} else {
//
dataApi. getData_1(getQueryParams()).enqueue(object : Callback<DataModel> {
override fun onResponse(call: Call<DataModel>?, response: Response<DataModel>?) {
// same handler code
}
override fun onFailure(call: Call<DataModel>?, throwable: Throwable?) {
// same error handler code
}
})
}
}
If you want to minimize the duplicate code, create the callback as a class object and use it for both calls (something like this):
private val callback = object : Callback<DataModel> {
override fun onResponse(call: Call<DataModel>?, response: Response<DataModel>?) {
// handler code
}
override fun onFailure(call: Call<DataModel>?, throwable: Throwable?) {
// error handler code
}
}
private fun fetchFirst() {
if (UseDataApi_1) {
dataApi.getData_1(getQueryParams()).enqueue(callback)
} else {
dataApi.getData_2(getQueryParams()).enqueue(callback)
}
}