Cache Manager / Data Storage

1. Requirements Document

Overview

A centralized system to manage client-side persistence. It supports multiple storage backends (Memory, LocalStorage, SecureStorage) and provides unified APIs for caching APIs, images, and user preferences.

User Stories

• As a Developer, I want to cache API responses to make the app work offline.
• As a Developer, I want to store sensitive tokens securely.
• As a User, I want my preferences (theme, language) to be remembered.

Functional Requirements

1. Storage Engines: support for In-Memory, Shared Preferences, and Secure Storage (Keychain/Keystore).
2. Expiration Policies: Time-to-Live (TTL) support for cached items.
3. eviction: LRU (Least Recently Used) eviction strategy for size-limited caches.
4. Types: Support for String, JSON, and primitive types.

Non-Functional Requirements

• Speed: Sync reads for memory cache; async for disk.
• Security: Encryption at rest for secure storage bucket.

2. Technical Document

Architecture

API Design

Cache Policy Configuration

enum EvictionPolicy {
  lru, // Least Recently Used
  fifo, // First In, First Out
  none
}
 
class CacheConfig {
    final Duration ttl;
    final int maxEntries;
    final EvictionPolicy eviction;
 
    const CacheConfig({
        this.ttl = const Duration(hours: 1),
        this.maxEntries = 100,
        this.eviction = EvictionPolicy.lru
    });
}

Usage Example

// Initialize
final cache = CacheManager(
    engine: DiskStorageEngine(),
    config: CacheConfig(ttl: Duration(days: 7))
);
 
// Write
await cache.put('user_profile', userJson);
 
// Read
final profile = await cache.get('user_profile');

Internal Logic

• Memory Layer: Uses a LinkedHashMap for O(1) access and inherent insertion ordering for LRU.
• Disk Layer: Uses Hive boxes or shared_preferences for persistence.
• Locking: Uses Mutex (mutual exclusion) to prevent race conditions during read/write on the same key.