Share more logic between iOS and Android

Now that you've implemented common logic using external dependencies, you can start adding more complex logic. Network requests and data serialization are the most popular use cases for sharing code using Kotlin Multiplatform. Learn how to implement these in your first application, so that after completing this onboarding journey you can use them in future projects.

The updated app will retrieve data over the internet from the SpaceX API and display the date of the last successful launch of a SpaceX rocket.

Add more dependencies

You'll need the following multiplatform libraries in your project:

kotlinx.coroutines, for using coroutines to write asynchronous code, which allows simultaneous operations.
kotlinx.serialization, for deserializing JSON responses into objects of entity classes used to process network operations.
Ktor, a framework as an HTTP client for retrieving data over the internet.

kotlinx.coroutines

To add kotlinx.coroutines to your project, specify a dependency in the common source set. To do so, add the following line to the build.gradle.kts file of the shared module:

kotlin {
   // ... 
   sourceSets {
       commonMain.dependencies {
           // ...
           implementation("org.jetbrains.kotlinx:kotlinx-coroutines-core:1.7.3")
       }
   }
}

The Multiplatform Gradle plugin automatically adds a dependency to the platform-specific (iOS and Android) parts of kotlinx.coroutines.

kotlinx.serialization

To use the kotlinx.serialization library, set up a corresponding Gradle plugin. To do that, add the following line to the existing plugins{} block at the very beginning of the build.gradle.kts file in the shared module:

plugins {
    // ...
    kotlin("plugin.serialization") version "1.9.23"
}

Ktor

You need to add the core dependency (ktor-client-core) to the common source set of the shared module. You also need to add supporting dependencies:

Add the ContentNegotiation functionality (ktor-client-content-negotiation), which allows serializing and deserializing the content in a specific format.
Add the ktor-serialization-kotlinx-json dependency to instruct Ktor to use the JSON format and kotlinx.serialization as a serialization library. Ktor will expect JSON data and deserialize it into a data class when receiving responses.
Provide the platform engines by adding dependencies on the corresponding artifacts in the platform source sets (ktor-client-android, ktor-client-darwin).

kotlin {
    // ...
    val ktorVersion = "2.3.7"

    sourceSets {
        commonMain.dependencies {
            // ...

            implementation("io.ktor:ktor-client-core:$ktorVersion")
            implementation("io.ktor:ktor-client-content-negotiation:$ktorVersion")
            implementation("io.ktor:ktor-serialization-kotlinx-json:$ktorVersion")
        }
        androidMain.dependencies {
            implementation("io.ktor:ktor-client-android:$ktorVersion")
        }
        iosMain.dependencies {
            implementation("io.ktor:ktor-client-darwin:$ktorVersion")
        }
    }
}

Synchronize the Gradle files by clicking Sync Now in the notification.

Create API requests

You'll need the SpaceX API to retrieve data, and you'll use a single method to get the list of all launches from the v4/launches endpoint.

Add a data model

In shared/src/commonMain/kotlin, create a new RocketLaunch.kt file in the project directory and add a data class which stores data from the SpaceX API:

import kotlinx.serialization.SerialName
import kotlinx.serialization.Serializable

@Serializable
data class RocketLaunch (
    @SerialName("flight_number")
    val flightNumber: Int,
    @SerialName("name")
    val missionName: String,
    @SerialName("date_utc")
    val launchDateUTC: String,
    @SerialName("success")
    val launchSuccess: Boolean?,
)

The RocketLaunch class is marked with the @Serializable annotation, so that the kotlinx.serialization plugin can automatically generate a default serializer for it.
The @SerialName annotation allows you to redefine field names, making it possible to declare properties in data classes with more readable names.

Connect HTTP client

In shared/src/commonMain/kotlin, create a new RocketComponent class in the project directory.
Add the httpClient property to retrieve rocket launch information through an HTTP GET request:
import io.ktor.client.* import io.ktor.client.plugins.contentnegotiation.* import io.ktor.serialization.kotlinx.json.* import kotlinx.serialization.json.Json class RocketComponent { private val httpClient = HttpClient { install(ContentNegotiation) { json(Json { prettyPrint = true isLenient = true ignoreUnknownKeys = true }) } } }
- The ContentNegotiation Ktor plugin and the JSON serializer deserialize the result of the GET request.
- The JSON serializer here is configured in a way that it prints JSON in a more readable manner with the prettyPrint property. It is more flexible when reading malformed JSON with isLenient, and it ignores keys that haven't been declared in the rocket launch model with ignoreUnknownKeys.
Add the getDateOfLastSuccessfulLaunch() suspending function to RocketComponent:
private suspend fun getDateOfLastSuccessfulLaunch(): String { // ... }
Call the httpClient.get() function to retrieve information about rocket launches:
import io.ktor.client.request.* import io.ktor.client.call.* private suspend fun getDateOfLastSuccessfulLaunch(): String { val rockets: List<RocketLaunch> = httpClient.get("https://api.spacexdata.com/v4/launches").body() }
- httpClient.get() is also a suspending function because it needs to retrieve data over the network asynchronously without blocking threads.
- Suspending functions can only be called from coroutines or other suspending functions. This is why getDateOfLastSuccessfulLaunch() was marked with the suspend keyword. The network request is executed in the HTTP client's thread pool.
Update the function again to find the last successful launch in the list:
private suspend fun getDateOfLastSuccessfulLaunch(): String { val rockets: List<RocketLaunch> = httpClient.get("https://api.spacexdata.com/v4/launches").body() val lastSuccessLaunch = rockets.last { it.launchSuccess == true } }
The list of rocket launches is sorted by date from oldest to newest.
Convert the launch date from UTC to your local date and format the output:
import kotlinx.datetime.Instant import kotlinx.datetime.TimeZone import kotlinx.datetime.toLocalDateTime private suspend fun getDateOfLastSuccessfulLaunch(): String { val rockets: List<RocketLaunch> = httpClient.get("https://api.spacexdata.com/v4/launches").body() val lastSuccessLaunch = rockets.last { it.launchSuccess == true } val date = Instant.parse(lastSuccessLaunch.launchDateUTC) .toLocalDateTime(TimeZone.currentSystemDefault()) return "${date.month} ${date.dayOfMonth}, ${date.year}" }
The date will be in the "MMMM DD, YYYY" format, for example, OCTOBER 5, 2022.
Add another suspending function, launchPhrase(), which will create a message using the getDateOfLastSuccessfulLaunch() function:
suspend fun launchPhrase(): String = try { "The last successful launch was on ${getDateOfLastSuccessfulLaunch()} 🚀" } catch (e: Exception) { println("Exception during getting the date of the last successful launch $e") "Error occurred" }

Create the flow

You can use flows instead of suspending functions. They emit a sequence of values instead of a single value that suspending functions return.

Open the Greeting.kt file in the shared/src/commonMain/kotlin directory.
Add a rocketComponent property to the Greeting class. The property will store the message with the last successful launch date:
private val rocketComponent = RocketComponent()
Change the greet() function to return a Flow:
import kotlinx.coroutines.delay import kotlinx.coroutines.flow.Flow import kotlinx.coroutines.flow.flow import kotlin.time.Duration.Companion.seconds fun greet(): Flow<String> = flow { emit(if (Random.nextBoolean()) "Hi!" else "Hello!") delay(1.seconds) emit("Guess what this is! > ${platform.name.reversed()}") delay(1.seconds) emit(daysPhrase()) emit(rocketComponent.launchPhrase()) }
- The Flow is created here with the flow() builder function, which wraps all the statements.
- The Flow emits strings with a delay of one second between each emission. The last element is only emitted after the network response returns, so the exact delay depends on your network.

Add internet access permission

To access the internet, the Android application needs the appropriate permission. Since all network requests are made from the shared module, it makes sense to add the internet access permission to its manifest.

Update your composeApp/src/androidMain/AndroidManifest.xml file with the access permission:

<?xml version="1.0" encoding="utf-8"?>
<manifest xmlns:android="http://schemas.android.com/apk/res/android">
    <uses-permission android:name="android.permission.INTERNET"/>
    ...
</manifest>

Update Android and iOS apps

You've already updated the API of the shared module by changing the return type of the greet() function to Flow. Now you need to update native (iOS, Android) parts of the project so that they can properly handle the result of calling the greet() function.

Android app

As both the shared module and the Android application are written in Kotlin, using shared code from Android is straightforward.

Introduce a view model

Now that the application is becoming more complex, it's time to introduce a view model to the Android activity called MainActivity. It invokes the App() function that implements the UI. The view model will manage the data from the activity and won't disappear when the activity undergoes a lifecycle change.

Add the following dependencies to your composeApp/build.gradle.kts file:
androidMain.dependencies { // ... implementation("androidx.lifecycle:lifecycle-viewmodel-ktx:2.6.2") implementation("androidx.lifecycle:lifecycle-runtime-compose:2.6.2") implementation("androidx.lifecycle:lifecycle-viewmodel-compose:2.6.2") }
In composeApp/src/androidMain/kotlin, create a new MainViewModel Kotlin class:
import androidx.lifecycle.ViewModel class MainViewModel : ViewModel() { // ... }
This class extends Android's ViewModel class, which ensures the correct behavior regarding lifecycle and configuration changes.
Create a greetingList value of the StateFlow type and its backing property:
import kotlinx.coroutines.flow.MutableStateFlow import kotlinx.coroutines.flow.StateFlow class MainViewModel : ViewModel() { private val _greetingList = MutableStateFlow<List<String>>(listOf()) val greetingList: StateFlow<List<String>> get() = _greetingList }
- StateFlow here extends the Flow interface but has a single value or state.
- The private backing property _greetingList ensures that only clients of this class can access the read-only greetingList property.
In the init function of the View Model, collect all the strings from the Greeting().greet() flow:
import androidx.lifecycle.viewModelScope import kotlinx.coroutines.launch class MainViewModel : ViewModel() { private val _greetingList = MutableStateFlow<List<String>>(listOf()) val greetingList: StateFlow<List<String>> get() = _greetingList init { viewModelScope.launch { Greeting().greet().collect { phrase -> //... } } } }
Since the collect() function is suspended, the launch coroutine is used within the view model's scope. This means that the launch coroutine will run only during the correct phases of the view model's lifecycle.
Inside the collect trailing lambda, update the value of _greetingList to append the collected phrase to the list of phrases in list:
import kotlinx.coroutines.flow.update class MainViewModel : ViewModel() { //... init { viewModelScope.launch { Greeting().greet().collect { phrase -> _greetingList.update { list -> list + phrase } } } } }
The update() function will update the value automatically.

Use the view model's flow

In composeApp/src/androidMain/kotlin, locate the App.kt file and update it, replacing the previous implementation:
import androidx.lifecycle.compose.collectAsStateWithLifecycle import androidx.compose.runtime.getValue import androidx.lifecycle.viewmodel.compose.viewModel @Composable fun App(mainViewModel: MainViewModel = viewModel()) { MaterialTheme { val greetings by mainViewModel.greetingList.collectAsStateWithLifecycle() Column( modifier = Modifier.padding(all = 20.dp), verticalArrangement = Arrangement.spacedBy(8.dp), ) { greetings.forEach { greeting -> Text(greeting) Divider() } } } }
- The collectAsStateWithLifecycle() function calls on greetingList to collect the value from the view model's flow and represent it as a composable state in a lifecycle-aware manner.
- When a new flow is created, the compose state will change and display a scrollable Column with greeting phrases arranged vertically and separated by dividers.
To see the results, re-run your composeApp configuration in Android Studio:

iOS app

For the iOS part of the project, you'll make use of the Model–view–viewmodel pattern again to connect the UI to the shared module, which contains all the business logic.

The module is already imported in the ContentView.swift file with the import Shared declaration.

Introducing a view model

Go back to your iOS app in Xcode.
In iosApp/iOSApp.swift, update the entry point for your app:
@main struct iOSApp: App { var body: some Scene { WindowGroup { ContentView(viewModel: ContentView.ViewModel()) } } }
In iosApp/ContentView.swift, create a ViewModel class for ContentView, which will prepare and manage data for it:
import SwiftUI import Shared struct ContentView: View { @ObservedObject private(set) var viewModel: ViewModel var body: some View { ListView(phrases: viewModel.greetings) .onAppear { self.viewModel.startObserving() } } } extension ContentView { @MainActor class ViewModel: ObservableObject { @Published var greetings: Array<String> = [] func startObserving() { // ... } } } struct ListView: View { let phrases: Array<String> var body: some View { List(phrases, id: \.self) { Text($0) } } }
- ViewModel is declared as an extension to ContentView, as they are closely connected.
- ViewModel has a greetings property that is an array of String phrases. SwiftUI connects the view model (ContentView.ViewModel) with the view (ContentView).
- ContentView.ViewModel is declared as an ObservableObject.
- The @Published wrapper is used for the greetings property.
- The @ObservedObject property wrapper is used to subscribe to the view model.

Now the view model will emit signals whenever this property changes.

Choose a library to consume flows from iOS

In this tutorial, you can choose between the SKIE and KMP-NativeCoroutines libraries to help you work with flows in iOS. Both are open-source solutions that support cancellation and generics with flows, which the Kotlin/Native compiler doesn't yet provide by default.

The SKIE library augments the Objective-C API produced by the Kotlin compiler: SKIE transforms flows into an equivalent of Swift’s AsyncSequence. SKIE directly supports Swift's async/await, without thread restriction, and with automatic bi-directional cancellation (Combine and RxSwift require adapters). SKIE offers other features to produce a Swift-friendly API from Kotlin, including bridging various Kotlin types to Swift equivalents. It also doesn’t require adding additional dependencies in iOS projects.

The KMP-NativeCoroutines library helps you consume suspending functions and flows from iOS by generating necessary wrappers. KMP-NativeCoroutines supports Swift's async/await functionality as well as Combine and RxSwift. It has been available longer than SKIE, and thus you may encounter fewer edge cases with it today. Using KMP-NativeCoroutines requires adding a Cocoapod or SPM dependency in iOS projects.

Option 1. Configure SKIE

To set up the library, specify the SKIE plugin in shared/build.gradle.kts and click the Sync button.

plugins {
    id("co.touchlab.skie") version "0.6.1"
}

Consume the flow using SKIE

Return to Xcode and update the code using the library:

Use a loop and the await mechanism to iterate through the Greeting().greet() flow and update the greetings property every time the flow emits a value.
To support concurrency, wrap the asynchronous operation in a Task:
func startObserving() { Task { for await phrase in Greeting().greet() { self.greetings.append(phrase) } } }
Make sure ViewModel is marked with the @MainActor annotation. The annotation ensures that all asynchronous operations within ViewModel run on the main thread to comply with the Kotlin/Native requirement:
// ... extension ContentView { @MainActor class ViewModel: ObservableObject { @Published var greetings: [String] = [] func startObserving() { Task { for await phrase in Greeting().greet() { self.greetings.append(phrase) } } } } }
Re-run the iosApp configuration from Android Studio to make sure your app's logic is synced:

Option 2. Configure KMP-NativeCoroutines

Return to Android Studio. In the build.gradle.kts file of the whole project, add the KSP (Kotlin Symbol Processor) and KMP-NativeCoroutines plugins to the plugins{} block:
plugins { // ... id("com.google.devtools.ksp").version("1.9.21-1.0.16").apply(false) id("com.rickclephas.kmp.nativecoroutines").version("1.0.0-ALPHA-23").apply(false) }
In the shared build.gradle.kts file, configure the KMP-NativeCoroutines plugin:
plugins { // ... id("com.google.devtools.ksp") id("com.rickclephas.kmp.nativecoroutines") }
In the shared build.gradle.kts file, opt-in to the experimental @ObjCName annotation:
kotlin { // ... sourceSets{ all { languageSettings.optIn("kotlin.experimental.ExperimentalObjCName") } // ... } }
Synchronize the Gradle files by clicking Sync Now in the notification.

Mark the flow with KMP-NativeCoroutines

Open the Greeting.kt file in the shared/src/commonMain/kotlin directory.
Add the @NativeCoroutines annotation to the greet() function. This will ensure that the plugin generates the right code to support correct flow handling on iOS:
import com.rickclephas.kmp.nativecoroutines.NativeCoroutines class Greeting { // ... @NativeCoroutines fun greet(): Flow<String> = flow { // ... } }

Import the library using SPM in XCode

In Xcode, right-click the iosApp project in the left-hand menu and select Add Package Dependencies.
In the search bar, enter the package name:
https://github.com/rickclephas/KMP-NativeCoroutines.git
Keep the default options, "Branch" for Dependency Rule, and "master" for Version Rule, and then click the Add Package button.
In the next window, select "KMPNativeCoroutinesAsync" and "KMPNativeCoroutinesCore", and then click Add Package:

This should install the KMP-NativeCoroutines package necessary to work with the async/await mechanism.

Consume the flow using the KMP-NativeCoroutines library

In iosApp/ContentView.swift, update the startObserving() function to consume the flow using KMP-NativeCoroutine's asyncSequence() function for the Greeting().greet() function:
func startObserving() { do { let sequence = asyncSequence(for: Greeting().greet()) for try await phrase in sequence { self.greetings.append(phrase) } } catch { print("Failed with error: \(error)") } }
The loop and the await mechanism here are used here to iterate through the flow and update the greetings property every time the flow emits a value.
To support concurrency, wrap the asynchronous operation in Task:
func startObserving() { Task { do { let sequence = asyncSequence(for: Greeting().greet()) for try await phrase in sequence { self.greetings.append(phrase) } } catch { print("Failed with error: \(error)") } } }
Make sure ViewModel is marked with the @MainActor annotation. The annotation ensures that all asynchronous operations within ViewModel run on the main thread to comply with the Kotlin/Native requirement:
// ... import KMPNativeCoroutinesAsync import KMPNativeCoroutinesCore // ... extension ContentView { @MainActor class ViewModel: ObservableObject { @Published var greetings: Array<String> = [] func startObserving() { Task { do { let sequence = asyncSequence(for: Greeting().greet()) for try await phrase in sequence { self.greetings.append(phrase) } } catch { print("Failed with error: \(error)") } } } } }
Re-run the iosApp configuration from Android Studio to make sure your app's logic is synced:

Next step

In the final part of the tutorial, you'll wrap up your project and see what steps to take next.

Proceed to the next part

Get help

Kotlin Slack. Get an invite and join the #multiplatform channel.
Kotlin issue tracker. Report a new issue.

Last modified: 25 April 2024

Kotlin Multiplatform Development 1.6.2 Help