graphrag-patterns - GraphRAG Implementation Patterns for Neo4j
Provides GraphRAG implementation patterns including hybrid search, Text2Cypher, agentic retrieval, and RAGAS evaluation for Neo4j.
Tags
Updated: 2026-09-09Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Create hybrid retrievers
- Generate Cypher queries with few-shot
- Route queries to retrieval strategies
- Enrich vector search with graph
- Evaluate retrieval quality with RAGAS
- Execute stepback prompting retrieval
Inputs
- Neo4j database connection driver
- Embedding models
- LLM models
- User search queries
- RAG evaluation datasets
Outputs
- Enriched search result lists
- Generated Cypher query chains
- Query strategy classification labels
- RAGAS quality metric scores
- Deduplicated retrieval documents
Requirements
- Neo4j graph database
- Python runtime environment
- langchain_neo4j package
- ragas package