Struct MultiQueryRetriever 

pub struct MultiQueryRetriever { /* private fields */ }

MultiQueryRetriever

Uses an LLM to generate multiple query variants, retrieves with the base retriever for each one, then merges and dedups the results before returning.

Implementations

impl MultiQueryRetriever

pub fn new<L>(llm: L, base_retriever: Arc<dyn RetrieverTrait>) -> Self

where L: BaseChatModel + Send + Sync + 'static, L::Error: Into<ProviderError>,

Creates a MultiQueryRetriever (accepting any LLM implementing BaseChatModel)

pub fn new_arc( llm: Arc<dyn BaseChatModel<Error = ProviderError> + Send + Sync>, base_retriever: Arc<dyn RetrieverTrait>, ) -> Self

P0-3: Builds from an already-wrapped Arc<dyn BaseChatModel<Error = ProviderError>>

pub fn with_config(self, config: MultiQueryConfig) -> Self

Sets the MultiQuery configuration

pub fn with_num_queries(self, n: usize) -> Self

Sets the number of generated queries

pub fn with_k_per_query(self, k: usize) -> Self

Sets the number of documents returned per query

pub fn with_final_k(self, k: usize) -> Self

Sets the number of documents returned in the final result

pub async fn retrieve_multi( &self, query: &str, ) -> Result<Vec<Document>, MultiQueryError>

Generates multiple query variants, retrieves each one, and returns the merged deduped documents

pub async fn retrieve_multi_with_scores( &self, query: &str, ) -> Result<Vec<SearchResult>, MultiQueryError>

Generates multiple query variants, retrieves each one, and returns the merged deduped results with scores

pub async fn get_generated_queries( &self, query: &str, ) -> Result<Vec<String>, MultiQueryError>

Returns the LLM-generated query variants (without retrieving)