Module cache 

Query result caching with LRU eviction

This module provides an LRU (Least Recently Used) cache for query results, significantly improving performance for repeated queries.

Performance Impact

• Cache hit: ~1-10μs (instant)
• Cache miss: Full search time (~1-10ms)
• Typical hit rate: 20-40% for production workloads

Memory Usage

• Each cached entry: ~(key_size + result_size) bytes
• Default capacity: 1000 entries
• Configurable max size

Modules

early_termination

Early termination optimization for k-NN search

Structs

CacheStats

Cache statistics

QueryCache

LRU cache for query results with TTL support

SharedQueryCache

Thread-safe LRU cache wrapper