Challenge 1: An Offline-First Repository — Solution

class NoteRepository @Inject constructor(
    private val dao: NoteDao,
    private val api: NoteApiService
) {
    val notes: Flow<List<Note>> = dao.getAllNotes()

    suspend fun refresh() {
        try {
            val remoteNotes = api.getNotes()
            remoteNotes.forEach { dao.insert(it) }
        } catch (e: IOException) {
            // Network failed — do nothing; the UI keeps observing whatever
            // is already cached in Room via the "notes" Flow above.
        }
    }
}

Notes:
- notes exposes dao.getAllNotes() DIRECTLY — the UI (via a ViewModel)
  only ever observes the local database, never the network response
  itself. This is the defining trait of offline-first: local storage is
  the single source of truth the UI actually watches.
- refresh() is a one-way sync: fetch from the network, then write each
  result into Room via dao.insert(...). Because notes is a Flow backed
  by Room (per Chapter 3), those writes automatically cause "notes" to
  re-emit with the fresh data — no separate manual "update the UI" step
  is needed after a successful refresh.
- The empty catch block is intentional, not an oversight — if the
  network call fails, the UI simply continues showing whatever was
  already cached from the last successful refresh (or from earlier app
  usage), which is a perfectly reasonable degraded experience rather
  than an error state that needs special UI handling.