UI Data Mappers & Repositories

​

Data Strategy

​

Key Repositories & Mappers

• AppUiModelMapper: The central logic hub that converts DailyAppUsageRecord and AppScrollData into UI models. It performs complex joins in-memory to attach LimitInfo to every app item.
• AppMetadataRepository: Manages the lifecycle of app-specific information. It synchronizes the local database with the Android PackageManager.
• LimitsRepository: Handles the persistence and retrieval of user-defined time limits and app groupings.

​

Schema & Performance (Data Engineering)

• AppMetadata Table: Stores package names, version codes, and visibility flags.
  • Heuristic Visibility: The system determines if an app is “User Visible” by checking for a launcher intent.
  • Version Tracking: Uses versionCode to detect app updates and trigger metadata refreshes.
• LimitGroup & LimitedApp: Implements a One-to-Many relationship. A LimitGroup defines a shared time bucket, and multiple LimitedApp entries link specific packages to that bucket.
• Performance Optimization: The mapper provides “No-DB” variants (e.g., mapToAppOpenUiItemNoDb) to prevent N+1 query problems when rendering large lists.

​

Storage & Caching

• Icon Cache: App icons are not stored in the database. They are persisted as .png files in the internal app_icons directory. The database only stores an isIconCached boolean.
• Write-Through Metadata: When an app is installed or updated, the AppMetadataRepository immediately fetches the new info and updates the local cache.
• Cleanup Logic: The LimitsRepository includes a cleanupUninstalledApps routine to prevent orphaned limit records for apps no longer on the device.

​

Threading & Concurrency

• Dispatcher Policy: All mapping operations are wrapped in withContext(Dispatchers.IO) to ensure the Main Thread remains responsive during icon loading or database lookups.
• Asynchronous Streams: Repositories expose data using Flow, allowing the UI to reactively update when limits are changed or apps are uninstalled.
• Atomic Transactions: Complex operations, such as moving an app between limit groups, are wrapped in appDatabase.withTransaction to prevent Race Conditions.

​

UI Models (DTOs)

​

Usage Example

// In a ViewModel
val uiItems = records.map { record ->
    mapper.mapToAppUsageUiItem(record, todayUsageMap)
}

// Pre-fetch limits to avoid repeated DB hits
val limitsMap = limitsRepository.getLimitsForApps(packageNames)
val uiItems = records.map { record ->
    mapper.mapToAppOpenUiItemNoDb(record, limitsMap[record.packageName], ...)
}