Expand description
Embedding model integration for MenteDB.
Provides a trait-based architecture for embedding providers with built-in caching and a hash-based provider for testing.
Re-exports§
pub use cache::CacheStats;pub use cache::CachedEmbedding;pub use cache::EmbeddingCache;pub use candle_provider::CandleEmbeddingProvider;pub use hash_provider::HashEmbeddingProvider;pub use http_provider::HttpEmbeddingConfig;pub use http_provider::HttpEmbeddingProvider;pub use manager::EmbeddingManager;pub use manager::EmbeddingStats;pub use provider::AsyncEmbeddingProvider;pub use provider::EmbeddingProvider;
Modules§
- cache
- LRU embedding cache with hit/miss tracking. LRU cache for computed embeddings to avoid recomputation.
- candle_
provider - Local embedding provider using Candle (pure Rust ML framework). Local embedding provider using Candle (pure Rust ML framework).
- hash_
provider - Deterministic hash based embedding provider for testing. A deterministic, zero-dependency embedding provider for testing and development.
- http_
provider - HTTP based embedding provider for remote model APIs. Generic HTTP-based embedding provider for OpenAI, Cohere, Voyage, and other APIs.
- manager
- Embedding manager that wraps providers with caching. Embedding manager that wraps a provider with caching and statistics.
- provider
- Trait definitions for sync and async embedding providers. Embedding provider traits for synchronous and asynchronous embedding generation.