book

Essential GraphRAG

Name: Essential GraphRAG
ISBN: 9781633436268

by Tomaz Bratanic, Oskar Hane

July 2025

Intermediate to advanced

176 pages

4h 50m

English

Manning Publications

Audiobook available

Read now

Unlock full access

Essential GraphRAG
copyright
contents
foreword
preface
acknowledgments
about this book
about the authors
about the cover illustration
1 Improving LLM accuracy
1.1 Introduction to LLMs1.2 Limitations of LLMs1.2.1 Knowledge cutoff problem1.2.2 Outdated information1.2.3 Pure hallucinations1.2.4 Lack of private information1.3 Overcoming the limitations of LLMs1.3.1 Supervised finetuning1.3.2 Retrieval-augmented generation1.4 Knowledge graphs as the data storage for RAG applications

2 Vector similarity search and hybrid search
2.1 Components of a RAG architecture2.1.1 The retriever2.1.2 The generator2.2 RAG using vector similarity search2.2.1 Application data setup2.2.2 The text corpus2.2.3 Text chunking2.2.4 Embedding model2.2.5 Database with vector similarity search function2.2.6 Performing vector search2.2.7 Generating an answer using an LLM2.3 Adding full-text search to the RAG application to enable hybrid search2.3.1 Full-text search index2.3.2 Performing hybrid search2.4 Concluding thoughts
3 Advanced vector retrieval strategies
3.1 Step-back prompting3.2 Parent document retriever3.2.1 Retrieving parent document strategy data3.3 Complete RAG pipeline
4 Generating Cypher queries from natural language questions
4.1 The basics of query language generation4.2 Where query language generation fits in the RAG pipeline4.3 Useful practices for query language generation4.3.1 Using few-shot examples for in-context learning4.3.2 Using database schema in the prompt to show the LLM the structure of the knowledge graph4.3.3 Adding terminology mapping to semantically map the user question to the schema4.3.4 Format instructions4.4 Implementing a text2cypher generator using a base model4.5 Specialized (finetuned) LLMs for text2cypher4.6 What we’ve learned and what text2cypher enables
5 Agentic RAG
5.1 What is agentic RAG?5.1.1 Retriever agents5.1.2 The retriever router5.1.3 Answer critic5.2 Why do we need agentic RAG?5.3 How to implement agentic RAG5.3.1 Implementing retriever tools5.3.2 Implementing the retriever router5.3.3 Implementing the answer critic5.3.4 Tying it all together
6 Constructing knowledge graphs with LLMs
6.1 Extracting structured data from text6.1.1 Structured Outputs model definition6.1.2 Structured Outputs extraction request6.1.3 CUAD dataset6.2 Constructing the graph6.2.1 Data import6.2.2 Entity resolution6.2.3 Adding unstructured data to the graph
7 Microsoft’s GraphRAG implementation
7.1 Dataset selection7.2 Graph indexing7.2.1 Chunking7.2.2 Entity and relationship extraction7.2.3 Entity and relationship summarization7.2.4 Community detection and summarization7.3 Graph retrievers7.3.1 Global search7.3.2 Local search
8 RAG application evaluation
8.1 Designing the benchmark dataset8.1.1 Coming up with test examples8.2 Evaluation8.2.1 Context recall8.2.2 Faithfulness8.2.3 Answer correctness8.2.4 Loading the dataset8.2.5 Running evaluation8.2.6 Observations8.3 Next steps
appendix A The Neo4j environment
A.1 Cypher query languageA.2 Neo4j installationA.2.1 Neo4j Desktop installationA.2.2 Neo4j Docker installationA.2.3 Neo4j AuraA.3 Neo4j Browser configurationA.4 Movies datasetA.4.1 Loading via the Neo4j Query GuideA.4.2 Trying the online versionA.4.3 Loading via Cypher
references

Content preview from Essential GraphRAG

2 Vector similarity search and hybrid search

This chapter covers

Introduction to embeddings, embedding models, vector space, and vector similarity search
How vector similarity fits in RAG applications
A practical walkthrough of a RAG application using vector similarity search
Adding full-text search to the RAG application to see how enabling a hybrid search approach can improve results

Creating a knowledge graph can be an iterative process where you start with unstructured data and then add structure to it. This is often the case when you have a lot of unstructured data and you want to start using it to answer questions.

This chapter will look at how we can use RAG to answer questions using unstructured data. We’ll look at how to use ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9781633436268Publisher Support Publisher Website

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Essential GraphRAG

by Tomaz Bratanic, Oskar Hane