Module output_cache 

Semantic output deduplication cache with TTL and pluggable eviction policies.

Structs

CacheKey

Key identifying a cached prompt/model/temperature combination.

CacheStats

Statistics snapshot returned by OutputCache::stats.

CachedOutput

A single cached LLM response entry.

OutputCache

In-memory output deduplication cache.

Enums

EvictionPolicy

Policy used to select which entry to evict when the cache is full.

Functions

cache_key

Build a CacheKey from a prompt string, model name, and temperature.

fnv1a_hash

FNV-1a 64-bit hash of data.