Learn how Graph RAG can improve the accuracy and reliability of RAG applications! Store information as nodes and edges in a Neo4j database, and give your LLM the ability to query it so it can retrieve entity and relational information to provide informed answers.

Graphs and RAG

Creating nodes

Creating relationships

Saving graph documents

Querying a knowledge graph

Writing Cypher statements

Running Cypher statements

The MERGE clause

Text-to-Cypher Graph RAG with Neo4j

Building a text-to-Cypher chain

Text-to-Cypher retrieval chain

Getting Started with Graph RAG and Neo4j

Text-to-Cypher applications work well in many circumstances, but we can do better than that! Discover how to construct graph databases using different graph models including lexical and domain graphs. Create Neo4j vector indexes so that you can have the best of both worlds and run graph and vector retrieval simultaneously!

Lexical graphs

Elements of a lexical graph

Splitting the play into Acts

Creating a hierarchical lexical graph

Combining lexical graphs with vector search

Creating text chunks

Creating a vector index

Domain graphs

Creating a structured output

Requesting a structured output

Providing few-shot examples

Building a hybrid retrieval chain

Runnable lambdas

Assigning additional values to an input

The final link in the chain

Graph Models and Hybrid RAG

Although Graph RAG applications are generally more reliable than vector RAG, they aren't totally infallible. In this chapter, you'll learn to evaluate your Graph RAG applications, spot incorrect or duplicate graph nodes, and integrate long-term memory so user preferences can be learned over time.

Entity resolution

Using extracted entities

Graph-based entity resolution

Evaluating Graph RAG with RAGAS

Creating a Ragas evaluation

Evaluating context retrieval with Ragas

Memory graphs

Saving conversation memory in the graph

Extracting facts from conversation histories

Using extracted conversation facts

Congratulations!

Improving Retrieval Quality

Are you bored of vectors, embeddings, and vector RAG applications yet? Look no further! In this course, you'll discover how Graph RAG can greatly improve the accuracy and reliability of RAG applications by storing and querying information in the form of nodes and relationships. Combine Neo4j graph databases with LangChain and you get a truly awesome way of retrieving and integrating external data with LLMs.

<h2>Level-Up Your RAG Applications with Graphs</h2>Are you bored of vectors, embeddings, and vector RAG applications yet? Look no further! In this course, you'll discover how Graph RAG can greatly improve the accuracy and reliability of RAG applications by storing and querying information in the form of nodes and relationships. Combine Neo4j graph databases with LangChain and you get a truly awesome way of retrieving and integrating external data with LLMs.<br><br><h2>Create Neo4j Graph Databases from Unstructured Text</h2>What if my dataset is messy unstructured text rather than a graph? Don't panic&mdash;you'll learn how to use LLMs with structured outputs to extract entities and relationships from text, and create new nodes and relationships for your graph database. You'll utilize the Pydantic library to define strict data structures for your LLM to populate with extracted text data.<br><br><h2>Combine Vectors and Graphs for Hybrid RAG</h2>You don't have to choose between vectors vs. graphs&mdash;you can have the best of both worlds! Discover how you can retrieve from both data sources in a single workflow and carefully construct prompts to integrate them into a hybrid RAG application.<br><br><h2>Integrate Long-Term Chatbot Memory</h2>Graph databases like Neo4j aren't only useful as knowledge bases for retrieval, you can also store long-term information like user facts and preferences as graphs! This long-term memory can then be queried just like any other graph database to integrate these preferences and personalize your applications.

Retrieval Augmented Generation (RAG) with LangChain

Create more accurate and reliable RAG systems with Graph RAG and hybrid RAG.

Graph RAG with LangChain and Neo4j

Manager, Developer Education at Neo4j

Using relationships in semantic search

Text properties to embeddings

Neo4j as a vector store

Chunking existing node properties

Creating a vector index on an existing graph

Vector retrieval with LCEL

Let's practice!

https://s3.amazonaws.com/assets.datacamp.com/production/course_39706/subtitles/course_39706_b06d1b73838bc883a68ab336840d2659.vtt

Graph Models and Hybrid RAG - Combining lexical graphs with vector search

Combining lexical graphs with vector search

Create Your Free Account