Trait Embedder

Source

pub trait Embedder: Send + Sync {
    // Required methods
    fn dimension(&self) -> usize;
    fn embed<'life0, 'life1, 'async_trait>(
        &'life0 self,
        texts: &'life1 [String],
    ) -> Pin<Box<dyn Future<Output = Result<Vec<Vec<f32>>, MemoryError>> + Send + 'async_trait>>
       where 'life0: 'async_trait,
             'life1: 'async_trait,
             Self: 'async_trait;

    // Provided method
    fn is_ranking(&self) -> bool { ... }
}

Expand description

Compute embeddings for one or more texts.

Output vectors must each be of length Embedder::dimension. Implementations must be deterministic at the type level — the dimensionality cannot vary across calls on the same instance.

Required Methods§

Source

fn dimension(&self) -> usize

Embedding dimensionality. Must be constant for a given Embedder instance — long-term memory backends reject vectors of the wrong length at runtime.

Source

fn embed<'life0, 'life1, 'async_trait>( &'life0 self, texts: &'life1 [String], ) -> Pin<Box<dyn Future<Output = Result<Vec<Vec<f32>>, MemoryError>> + Send + 'async_trait>>
where 'life0: 'async_trait, 'life1: 'async_trait, Self: 'async_trait,

Compute one embedding per input text.

Provided Methods§

Source

fn is_ranking(&self) -> bool

Whether distinct inputs produce distinct directions, i.e. whether a vector store wired with this embedder actually ranks by relevance.

Defaults to true — a real embedding model needs no ceremony. Only NonRankingEmbedder and its like override it, which is what lets a store report klieo_core::memory::RecallSemantics::NonRanking instead of claiming vector recall it cannot deliver.