Struct SelfQueryRetriever 

pub struct SelfQueryRetriever<M: BaseChatModel> { /* private fields */ }

Self-querying retriever: LLM splits the query -> whitelist validation -> filtered similarity search.

Implements RetrieverTrait and can be wrapped into an LCEL chain by crate::RetrieverRunnable.

Implementations

impl<M: BaseChatModel> SelfQueryRetriever<M>

pub fn new( llm: impl Into<Arc<M>>, store: Arc<dyn VectorStore>, embeddings: Arc<dyn Embeddings>, allowed_attributes: Vec<String>, ) -> Self

Creates a self-querying retriever.

• llm: the model responsible for splitting the natural-language query (supports structured output, or falls back to text).
• store / embeddings: the vector store and embedding model used for the filtered similarity search.
• allowed_attributes: the whitelist of attributes allowed in filter fields; an empty whitelist disables filtering entirely (LLM-returned filters are always ignored with a warning); with a non-empty whitelist, a filter referencing a field outside the whitelist explicitly returns RetrieverError::InvalidFilter rather than silently dropping it and falling back to an unfiltered search.

Trait Implementations

impl<M: BaseChatModel> RetrieverTrait for SelfQueryRetriever<M>

fn retrieve<'life0, 'life1, 'async_trait>( &'life0 self, query: &'life1 str, k: usize, ) -> Pin<Box<dyn Future<Output = Result<Vec<Document>, RetrieverError>> + Send + 'async_trait>>

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

Retrieves relevant documents

fn retrieve_with_scores<'life0, 'life1, 'async_trait>( &'life0 self, query: &'life1 str, k: usize, ) -> Pin<Box<dyn Future<Output = Result<Vec<SearchResult>, RetrieverError>> + Send + 'async_trait>>

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

Retrieves relevant documents (with scores)

fn add_documents<'life0, 'async_trait>( &'life0 self, documents: Vec<Document>, ) -> Pin<Box<dyn Future<Output = Result<(), RetrieverError>> + Send + 'async_trait>>

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

Adds documents