0.0.3 (Sep 2, 2016)
Jul 29, 2016
Apr 10, 2017
Víctor Albertos (VictorAlbertos)
Miguel Garcia (miguelbcr)
Víctor Albertos (VictorAlbertos)
Source code
APK file



The act of caching data with ReactiveCache is just another transformation in the reactive chain. ReactiveCache's API exposes both Single, SingleTransformer and Completable reactive types to gracefully merge the caching actions with the data stream.


  • A dual cache based on both memory and disk layers.
  • Automatic deserialization-serialization for custom Types, List, Map and Array.
  • Pagination
  • A lifetime system to expire data on specific time lapses.
  • Data encryption.
  • Customizable disk cache size limit.
  • Migrations to evict data by Type between releases.
  • A complete set of built-in functions to perform write operations easily using List, such as addFirst, evictLast, addAll and so on.


Add to top level file

allprojects {
    repositories {
        maven { url "" }

Add to app module file

dependencies {
    compile 'com.github.VictorAlbertos:ReactiveCache:1.1.1-2.x'
    compile 'com.github.VictorAlbertos.Jolyglot:gson:0.0.3'
    compile 'io.reactivex.rxjava2:rxjava:2.0.4'



Create a single instace of ReactiveCache for your entire application. The builder offers some additional configurations.

ReactiveCache reactiveCache = new ReactiveCache.Builder()
        .using(application.getFilesDir(), new GsonSpeaker());

evictAll() returns a Completable which evicts the cached data for every provider:



Call reactiveCache#provider() to create a Provider to manage cache operations. The builder offers some additional configurations.

Provider<List<Model>> cacheProvider =

replace() returns a SingleTransformer which replaces the provider data with the item emitted from the Single source. If the source throws an exception, calling replace() doesn't evict the provider data.


read() returns an Single which emits the provider data. If there isn't any data available, throws an exception.

readWithLoader() returns a SingleTransformer which emits the provider data. If there isn't any data available, it subscribes to the Single source to cache and emit its item.


evict() returns a Completable which evicts the provider data:



Call reactiveCache#providerGroup() to create a ProviderGroup to manage cache operations with pagination support. The builder offers some additional configurations.

ProviderGroup<List<Model>> cacheProvider =

ProviderGroup exposes the same methods as Provider but requesting a key as an argument. That way the scope of the provider data in every operation is constrained to the data associated with the key.




evict() is an overloaded method to evict the provider data for the entire collection of groups.


Built-in functions for writing operations

When the data is encoded as type List<Model>, you may use ProviderList and ProviderGroupList. Both clases inherit from their base clase ( Provider and ProviderGroup respectively), so - besides exposing all their base funcionality- they offer a supletory api to perform write operations.

Call reactiveCache#providerList() to create a ProviderList.

ProviderList<Model> cacheProvider =

Or call reactiveCache#providerGroupList() to create a ProviderGroupList.

ProviderGroupList<Model> cacheProviderGroup =

Both cacheProvider.entries() and cacheProviderGroup.entries(group) return an ActionsList<Model> instance which allows to easily operate with the cached data thought a whole set of functions.

ActionsList<Model> actions = cacheProvider.entries();
ActionsList<Model> actions = cacheProviderGroup.entries(group);

Every function exposed through actions return a Completable which must be subscribed to in order to consume the action. Follow some examples:

actions.addFirst(new Model())

//Add a new element at 5 position
actions.add((position, count) -> position == 5, new Model())

//Evict first element if the cache has already 300 records
actions.evictFirst(count -> count > 300)

//Update the element with id 5
actions.update(model -> model.getId() == 5, model -> {
    return mock;

//Update all inactive modelds
actions.updateIterable(model -> model.isInactive(), model -> {
    return mock;

This table summarizes the available functions.

Use cases

Next examples illustrate how to use ReactiveCache on the data layer for client Android applications. They follow the well-known repository pattern in order to deal with data coming from a remote repository (server) and a local one (ReactiveCache).

Simple user session.

class UserRepository {
    private final Provider<User> cacheProvider;
    private final ApiUser api;

    UserRepository(ApiUser api, ReactiveCache reactiveCache) {
      this.api = api;
      this.cacheProvider = reactiveCache.<User>provider()

    Single<User> login(String email) {
      return api.loginUser(email)

    Single<Boolean> isLogged() {
          .map(user -> true)
          .onErrorReturn(observer -> false);

    Single<User> profile() {

    Completable updateUserName(String name) {
          .map(user -> {
            return user;

    Completable logout() {
      return api.logout().andThen(cacheProvider.evict());

Adding and removing tasks.

class TasksRepository {
    private final ProviderList<Task> cacheProvider;
    private final ApiTasks api;

    TasksRepository(ApiTasks api, ReactiveCache reactiveCache) {
      this.api = api;
      this.cacheProvider = reactiveCache.<Task>providerList()

    Single<Reply<List<Task>>> tasks(boolean refresh) {
      return refresh ? api.tasks().compose(cacheProvider.replaceAsReply())
          : api.tasks().compose(cacheProvider.readWithLoaderAsReply());

    Completable addTask(String name, String desc) {
      return api.addTask(1, name, desc)
              .addFirst(new Task(1, name, desc)));

    Completable removeTask(int id) {
      return api.removeTask(id)
              .evict((position, count, element) -> element.getId() == id));

Paginated feed of events.

class EventsRepository {
    private final ProviderGroup<List<Event>> cacheProvider;
    private final ApiEvents apiEvents;

    EventsRepository(ApiEvents apiEvents, ReactiveCache reactiveCache) {
      this.apiEvents = apiEvents;
      this.cacheProvider = reactiveCache.<List<Event>>providerGroup()

    Single<Reply<List<Event>>> events(boolean refresh, int page) {
      if (refresh) {




When building ReactiveCache the next global configurations are available thought the builder:

  • diskCacheSize(int) sets the max memory in megabytes for all the cached data on disk. Default value is 100.

  • encrypt(String) sets the key to be used for encrypting the data on those providers as such configured.

  • useExpiredDataWhenNoLoaderAvailable() if invoked, ReactiveCache dispatches records already expired instead of throwing.

  • migrations(List<MigrationCache>) every MigrationCache expects a version number and a Class[] to check what cached data matches with these classes to evict it from disk. Use MigrationCache for those Type which have added new fields between app releases.

ReactiveCache reactiveCache = new ReactiveCache.Builder()
            new MigrationCache(1, new Class[] {Model.class}),
            new MigrationCache(1, new Class[] {Model1.class})))
        .using(application.getFilesDir(), new GsonSpeaker());

Config provider

When building Provider, ProviderList, ProviderGroup or ProviderGroupList the next configuration is available thought the builder:

  • encrypt(boolean) when true, the data cached by this provider is encrypted using the key specified in ReactiveCache#encript(key). Default value is false.

  • expirable(boolean) when false, the data cached by this provider is not eligible to be expired if not enough space remains on disk. Default value is true.

  • lifeCache(long, TimeUnit) sets the amount of time before the data would be expired. By default the data has no life time.

 Provider<Model> cacheModel = reactiveCache.<Model>provider()
          .lifeCache(60, TimeUnit.MINUTES)


Víctor Albertos

Another author's libraries using RxJava:

  • Mockery: Android and Java library for mocking and testing networking layers with built-in support for Retrofit.
  • RxCache: Reactive caching library for Android and Java. (ReactiveCache uses internally the core from RxCache).
  • RxActivityResult: A reactive-tiny-badass-vindictive library to break with the OnActivityResult implementation as it breaks the observables chain.
  • RxFcm: RxJava extension for Android Firebase Cloud Messaging (aka fcm).
  • RxSocialConnect: OAuth RxJava extension for Android.