Struct EmbeddingCache 

pub struct EmbeddingCache { /* private fields */ }

An embedding cache that avoids re-computing embeddings for unchanged content.

Keys are (id, content_hash) pairs. When the content hash changes, the old embedding is invalidated and a new one is computed.

Implementations

impl EmbeddingCache

pub fn new() -> Self

Create an empty cache.

pub fn len(&self) -> usize

Number of cached entries.

pub fn is_empty(&self) -> bool

Whether the cache is empty.

pub fn hits(&self) -> u64

Cache hit count.

pub fn misses(&self) -> u64

Cache miss count.

pub fn get(&mut self, id: &str, content_hash: &str) -> Option<&Vec<f32>>

Look up an embedding by ID and content hash.

Returns Some(vector) if the cache has a valid entry (matching hash). Returns None if the entry is missing or the hash has changed.

pub fn insert(&mut self, id: String, content_hash: String, vector: Vec<f32>)

Insert or update a cache entry.

pub fn embed_cached( &mut self, items: &[(String, String, String)], provider: &dyn EmbeddingProvider, ) -> Result<Vec<EmbeddedItem>>

Embed items using the cache, only computing embeddings for cache misses.

items is a list of (id, content_hash, text) tuples. Items with unchanged content hashes use cached vectors. Changed items are embedded in batch and cached.

pub fn save(&self, path: &Path) -> Result<(), Box<dyn Error>>

Save the cache to a Parquet file.

pub fn load(path: &Path) -> Result<Self, Box<dyn Error>>

Load a cache from a Parquet file.

Trait Implementations

impl Default for EmbeddingCache

fn default() -> Self

Returns the “default value” for a type.