ParentDocumentRetriever
ParentDocumentRetriever is a retrieval component that matches or indexes smaller document segments while returning their larger enclosing documents as context.
Essential GraphRAG
Parent document retrieval is a strategy that embeds or matches smaller child segments while returning their enclosing parent document for broader context.
- How it is used
- Improves RAG retrieval when child-level semantic matching finds relevant passages but the language model needs broader parent context.
- How deep it goes
- Defines the hierarchy and implements splitting, embedding, graph import, and child indexing, but does not show a complete query-time retrieval call.
- Examples
- Uses Einstein’s Patents and Inventions, section splitting, 2,000-character parents, and 500-character children.
- What you can do after
- Builds a parent-child retrieval index with Python, embeddings, Cypher, and a Neo4j vector index.
Read how Essential GraphRAG explains ParentDocumentRetriever →
AI Agents and Applications
ParentDocumentRetriever is a LangChain retriever that indexes smaller child chunks while returning their larger parent chunks as retrieval context.
- How it is used
- Improves retrieval-augmented generation by matching granular child embeddings while supplying context-rich parent chunks.
- How deep it goes
- Explains splitters, stores, ingestion, verification, and querying, but defers systematic performance evaluation to dataset benchmarking.
- Examples
- Demonstrates Chroma, InMemoryStore, RecursiveCharacterTextSplitter, HTML ingestion, and a Cornwall Ranger search.
- What you can do after
- Build it from scratch with LangChain components, ingest documents, query child embeddings, and inspect returned parent context.
Read how AI Agents and Applications explains ParentDocumentRetriever →
Related concepts
- text chunkuses · A text chunk is a bounded segment of a larger text treated as a unit for processing, storage, or retrieval.
- MultiVectorRetrieveralternative to · MultiVectorRetriever is a LangChain retriever that indexes multiple vector representations associated with source documents and returns the corresponding original documents.
- retrieval-augmented generationused by · Retrieval-augmented generation is a technique that retrieves relevant information from external sources and supplies it as context to a generative model when producing an output.
- JSONuses · JSON is a standardized, language-independent text format for representing structured data using objects, arrays, strings, numbers, Boolean values, and null.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.