Module cache 

Tiered caching infrastructure.

Reflex uses:

• L1: exact-match lookup (in-memory)
• L2: semantic search (vector DB)

Start at TieredCache and TieredLookupResult.

Re-exports

pub use l1::L1Cache;

pub use l1::L1CacheHandle;

pub use l1::L1LookupResult;

pub use l2::BqSearchBackend;

pub use l2::DEFAULT_TOP_K_BQ;

pub use l2::DEFAULT_TOP_K_FINAL;

pub use l2::L2_COLLECTION_NAME;

pub use l2::L2_VECTOR_SIZE;

pub use l2::L2CacheError;

pub use l2::L2CacheResult;

pub use l2::L2Config;

pub use l2::L2LookupResult;

pub use l2::L2SemanticCache;

pub use l2::L2SemanticCacheHandle;

pub use l2::NvmeStorageLoader;

pub use l2::StorageLoader;

pub use l2::MockL2SemanticCache;mock

pub use l2::MockStorageLoader;mock

pub use tiered::MockTieredCache;mock

pub use tiered::TieredCache;

pub use tiered::TieredCacheHandle;

pub use tiered::TieredLookupResult;

pub use types::REFLEX_STATUS_ERROR;

pub use types::REFLEX_STATUS_HEADER;

pub use types::REFLEX_STATUS_HEALTHY;

pub use types::REFLEX_STATUS_NOT_READY;

pub use types::REFLEX_STATUS_READY;

pub use types::REFLEX_STATUS_STORED;

pub use types::ReflexStatus;

Modules

l1

L1 exact-match cache. L1 exact-match cache (in-memory).

l2

L2 semantic cache. L2 semantic cache (embedding + vector search + rescoring).

tiered

L1+L2 tiered cache wrapper. Tiered cache: L1 exact + L2 semantic.

types

Status/header types shared across the cache pipeline.