Module cache 

Result caching with LRU eviction and TTL

Design Philosophy

This cache implementation follows enterprise-grade patterns:

• Thread-Safe: Uses Arc + RwLock for concurrent access
• LRU Eviction: Least Recently Used items are evicted first
• TTL Support: Entries expire after configured time-to-live
• Statistics: Tracks hits, misses, and hit rates
• Lazy Cleanup: Expired items cleaned on access (no background threads)

Usage Example

use llm_shield_models::cache::{ResultCache, CacheConfig};
use llm_shield_core::ScanResult;
use std::time::Duration;

let cache = ResultCache::new(CacheConfig {
    max_size: 1000,
    ttl: Duration::from_secs(300),
});

// Insert a result
let result = ScanResult::pass("safe text".to_string());
cache.insert("key1".to_string(), result);

// Retrieve it
if let Some(cached_result) = cache.get("key1") {
    println!("Cache hit!");
}

// Check statistics
let stats = cache.stats();
println!("Hit rate: {:.2}%", stats.hit_rate() * 100.0);

Structs

CacheConfig

Configuration for the result cache

CacheStats

Cache performance statistics

ResultCache

Thread-safe result cache with LRU eviction and TTL