Struct EmbeddingsFilter 

pub struct EmbeddingsFilter { /* private fields */ }

Filters documents based on cosine similarity between the query embedding and document content embeddings.

Implementations

impl EmbeddingsFilter

pub fn new(embeddings: Arc<dyn Embeddings>, threshold: f32) -> Self

Create a new EmbeddingsFilter with the given embeddings provider and similarity threshold.

pub fn with_default_threshold(embeddings: Arc<dyn Embeddings>) -> Self

Create a new EmbeddingsFilter with the default threshold of 0.75.

Trait Implementations

impl DocumentCompressor for EmbeddingsFilter

fn compress_documents<'life0, 'life1, 'async_trait>( &'life0 self, documents: Vec<Document>, query: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Vec<Document>, SynapseError>> + Send + 'async_trait>>

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

Compress or filter documents based on relevance to the query.