Graph RAG overview

Microsoft-style graph+vector hybrid retrieval pipeline. Extracts entities and relationships from documents into a knowledge graph, applies community detection (Leiden algorithm) to identify topic clusters, generates hierarchical summaries at multiple granularities, and combines graph traversal with vector similarity for retrieval. Supports both local search (entity-focused) and global search (theme-focused) query modes. Excels at answering holistic questions that require synthesizing information across many documents.