Struct RagPipeline 

pub struct RagPipeline { /* private fields */ }

Retrieval-Augmented Generation pipeline.

Orchestrates document ingestion (chunking, embedding, storage) and context-aware retrieval for LLM injection.

Implementations

impl RagPipeline

pub fn new( vector_store: Arc<dyn VectorStore>, embedder: Arc<dyn EmbeddingProvider>, config: RagConfig, ) -> Self

Create a new RAG pipeline.

pub async fn ingest_document( &self, doc: &Document, ) -> ArgentorResult<Vec<DocumentChunk>>

Ingest a single document: chunk it, embed each chunk, and store.

pub async fn ingest_batch( &self, docs: &[Document], ) -> ArgentorResult<Vec<Vec<DocumentChunk>>>

Batch-ingest multiple documents.

pub async fn query( &self, question: &str, top_k: Option<usize>, ) -> ArgentorResult<RagResult>

Query the knowledge base and return scored, filtered chunks.

pub async fn query_with_context( &self, question: &str, top_k: Option<usize>, context_window: usize, ) -> ArgentorResult<RagResult>

Query and return a formatted context string sized to fit a given context window (in estimated tokens).

pub fn config(&self) -> &RagConfig

Return a reference to the pipeline configuration.