android-data-layer

An offline-first data layer coordinates a Room local cache with a Retrofit remote source via the repository pattern. It serves observable data from the local database using Flow while synchronizing remote updates through suspendable network calls.

The single source of truth pattern centralizes data access in a Repository that always serves local Room entities via Flow. The Repository fetches fresh data from the remote source and updates the local database, ensuring the UI observes a single, consistent local cache.

Background synchronization uses WorkManager to implement strategies like Stale-While-Revalidate and the Outbox Pattern. WorkManager orchestrates fetching remote data via Retrofit and persisting updates into the local Room database when the app is backgrounded.

Yes, this approach is designed for architectures using Clean Architecture with Hilt. Hilt binds the repository interfaces to their concrete implementations, managing the dependencies for Room, Retrofit, and WorkManager synchronization components.

Network errors are handled gracefully using error handling wrappers around suspendable Retrofit calls. The repository catches remote fetch exceptions and relies on the local Room cache to continue serving data, maintaining an offline-first experience.

The best way is exposing Room queries as Flow<List<Entities>> through a Repository. This allows the UI to reactively observe local database changes, automatically updating whenever the remote synchronization process writes fresh data into the local cache.