Struct CacheBackedEmbeddings 

pub struct CacheBackedEmbeddings { /* private fields */ }

An embeddings wrapper that caches results in a Store backend.

Previously computed embeddings are stored in the provided Store keyed by the SHA-256 hash of the input text. On subsequent calls, cached embeddings are returned directly, and only uncached texts are sent to the inner embeddings provider.

This aligns with LangChain Python’s CacheBackedEmbeddings architecture, allowing any Store implementation (in-memory, SQLite, PostgreSQL, Redis, etc.) to serve as the caching backend.

Implementations

impl CacheBackedEmbeddings

pub fn new( inner: Arc<dyn Embeddings>, store: Arc<dyn Store>, namespace: impl Into<String>, ) -> Self

Create a new cached embeddings wrapper.

• inner — the underlying embeddings provider to delegate to on cache misses.
• store — the Store backend for persisting cached embeddings.
• namespace — a logical namespace within the store (combined with "embedding_cache" as the prefix).

Trait Implementations

impl Embeddings for CacheBackedEmbeddings

fn embed_documents<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, texts: &'life1 [&'life2 str], ) -> Pin<Box<dyn Future<Output = Result<Vec<Vec<f32>>, SynapticError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Embed multiple texts (for batch document embedding).

fn embed_query<'life0, 'life1, 'async_trait>( &'life0 self, text: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Vec<f32>, SynapticError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Embed a single query text.