Dicoding Developer Coaching #32: Android | Reactive Programming dengan RxJava & Flow
- 1. Reactive Programming dengan RXJava dan Kotlin Flow Sidiq Permana Co founder nbs.dev & Google Developer Expert
- 2. Apa yang akan kita pelajari? ● Reactive vs Imperative Programming ● Konsep Utama Reactive Programming ● Asynchronous dan Stream ● Observable - Operator - Observer ● Use Case Umum di Android ● Perbandingan Library ● Konklusi
- 3. Reactive vs Imperative Programming Dengan pendekatan imperatif, pengembang menulis kode yang menentukan langkah-langkah yang harus diambil komputer untuk mencapai tujuan. Ini kadang-kadang disebut sebagai pemrograman algoritmik. Sebaliknya, pendekatan fungsional melibatkan penyusunan masalah sebagai seperangkat aliran fungsi yang akan dijalankan untuk memproses data.
- 5. Reactive Programming ● Reactive Programming adalah istilah pemrograman umum yang berfokus pada reaksi terhadap perubahan, seperti nilai data atau peristiwa (event). ● Dengan kata lain, pemrograman dengan aliran data asinkron yang mengirimkan data ke konsumen (consumer) saat tersedia. ● Spreadsheet adalah contoh bagus dari pemrograman reaktif: sel yang bergantung pada sel lain secara otomatis "bereaksi" saat sel lain itu berubah.Resources.
- 7. Asynchronous dan Stream Data ● Asynchronous adalah proses di mana Anda dapat menjalankan perintah secara bersamaan. ● Stream Data adalah urutan data yang diurutkan dalam waktu
- 8. Observable - Operator - Observer
- 9. Contoh
- 10. Urutan Pemrosesan Dengan Chaining Operator
- 11. Urutan Pemrosesan Dengan Chaining Operator
- 12. Hot dan Cold Stream Cold Hot Data transmission time Lazy stream - new emitting process will start when someone requests (subscribe) Eager stream - the emitting process is always broadcasted even no one asks Number of observer Unicast - only has one observer Multicast - can have multiple observers Location of data created Data is generated outside the stream Data is created in the stream Simple analogy CDs that need to be played (request) first by someone before listened A radio that continues to broadcast even though no one is listening. Many people can also listening
- 17. Faedah Utama ● Aplikasi lebih reaktif terhadap event! Seperti memproses event pengguna seperti ketuk dua kali, pencarian, tombol validasi, dan UX kompleks lainnya. ● Mengakses data dari sumber data secara asinkron tanpa callback hell. ● Memproses data yang diperoleh dari API, seperti mengurutkan dan memfilter data. ● Multi threading yang lebih baik.
- 18. Aplikasi Berasa Lebih Reaktif
- 19. Lebih Fokus Pada Hasil Akhir Ketimbang Prosesnya fun register(newUser: NewUser) { val username: String = newUser.getUsername() val password: String = newUser.getPassword() api.register(newUser, object : Callback<Boolean>() { fun onResponse(success: Boolean) { if (success) { api.login(username, password), object : Callback<Token>() { fun onResponse(token: Token) { api.getUser(token, object : Callback<UserDetails>() { fun onResponse(userDetails: UserDetails) { // continue with app flow } }) } }) } } }) } fun register(newUser: NewUser) { val username: String = newUser.getUsername() val password: String = newUser.getPassword() api.register(newUser) .filter { success -> success } .flatMap { success -> api.login(username, password) } .flatMap { token -> api.getUser(token) } .subscribe { userDetails -> // continue with app flow } }
- 20. Lebih Fokus Pada Hasil Akhir Ketimbang Prosesnya
- 21. Lebih Fokus Pada Hasil Akhir Ketimbang Prosesnya
- 22. Multithreading yang lebih cakep! observeSomething() .subscribeOn(Schedulers.io()) .observeOn(AndroidSchedulers.mainThread()) .subscribe { name -> println(name)} CoroutineScope(Dispatchers.Main).launch { observeSomething() .flowOn(Dispatchers.IO) .collect { name -> println(name) } }
- 23. Reactive Library
- 24. Membuat Data Stream Observable.just("1", "2", "3", "4", "5", "6") .map { string -> string.toInt() } .filter { number -> number%2 == 1 } .subscribe { result -> println("$result is odd number") } Dengan Observable RXJava CoroutineScope.launch { flowOf("1", "2", "3", "4", "5", "6") .map { string -> string.toInt() } .filter { number -> number % 2 == 1 } .collect { result -> println("$result is odd number") } } Dengan Flow
- 25. Tipe Stream RxJava ● Observable: to emits unlimited data streams. ● Flowable: to emits unlimited data streams with back pressure strategy. ● Single: to emits one data only. ● Maybe: same as Single, but it can be no data. ● Completable: stream which can only complete without emitting values. Flow ● Flow: There is no separate implementations without and with back pressure support in Kotlin Flow.
- 26. Backpressure Backpressure is a situation where the Publisher provides excessive data so that the Subscriber cannot accommodate it. There are several back pressure strategies to choose : ● BUFFER: Processes received data and waits for other data in the queue. ● DROP: Processes received data and ignores other data. ● LATEST: Processes received data, but stores the last element in cache. ● MISSING: Processes without using Backpressure Strategy, so you have to make your own custom. ● ERROR: Notifies MissingBackpressureException error if process is not robust.
- 27. Multithreading observeSomething() .subscribeOn(Schedulers.io()) .observeOn(AndroidSchedulers.mainThread()) .subscribe { name -> println(name)} CoroutineScope(Dispatchers.Main).launch { observeSomething() .flowOn(Dispatchers.IO) .collect { name -> println(name) } }
- 28. Common Schedulers and Dispatcher ● io: to performs processes on the I/O Thread such as calling API and writing databases. ● computation: to performs processes with high CPU requirements such as converting bitmaps or compressing. ● main: to receives the result of the run on the Android Main/UI Thread. ● single: useful if you have multiple processes originating from different places and want to merge them. ● custom: creates own custom scheduler with the help of Executor.
- 29. Perbandingan RxJava Flow Observable Types Observable, Flowable, Single, Maybe, Completable Flow (Only one) Create Observable create, just, dll. flowOf Get data subscribe collect Context preservation subscribeOn, observeOn flowOn Thread type Schedulers Dispatchers Backpressure Manual By default handled by Flow Request with Retrofit Need RxJava Adapter Can be used directly Database with Room Need library Room RxJava Need library Room ktx Lifecycle Management Using Disposable Using Scope
- 30. Kesimpulan ● RxJava dan Coroutine Flow memiliki kelebihan dan kekurangan dalam implementasi terlebih dalam learning curve ketika adopsinya. ● Penggunaan tools tidak terlepas dari kebutuhan dan pemahaman yang dimiliki kadang juga karena selera! ● Mempelajari dan memahami keduanya akan memperbanyak pilihan opsi untuk penyelesaian masalah dengan tools yang tepat! ● Give it try and Enjoy!