Trait Embedder 

pub trait Embedder: Send + Sync {
    // Required method
    fn embed<'life0, 'life1, 'async_trait>(
        &'life0 self,
        text: &'life1 str,
    ) -> Pin<Box<dyn Future<Output = Result<Vec<f32>, MemoryError>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait;
}

Turns text into a vector.

Implement over whatever embedder is available. The engine never calls this on the document side — PrecomputedSemanticIndex::build does that once — so the latency that matters is the single query embedding per recall.

Required Methods

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

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

Embed one string. Vectors must be L2-normalised and all the same width.

Implementors

impl Embedder for StaticEmbedder