Congratulations!
1. Congratulations!
Congratulations on making it to the end of the course! What a journey!2. Chapter 1
In Chapter 1, you learned about the advantages of graph RAG over vector RAG solutions, and3. Chapter 1
used the langchain_neo4j library to create and store information as nodes, relationships, and graph documents.4. Chapter 1
You used this understanding to build a text-to-Cypher chain so that user inputs could be translated into Cypher queries to retrieve relevant information from a Neo4j graph database.5. Chapter 2
In Chapter 2, you learned about how to build graphs from unstructured text sources, starting with lexical graphs. You created splitting mechanisms to split on lexical elements like Acts and scene in Shakespeare's Romeo and Juliet. Then you stored the resulting nodes and relationships in a graph database.6. Chapter 2
You discovered that you don't have to choose between vector and graph RAG, and that you can have the best of both worlds! You created a Neo4j vector index and stored vectors as node properties, so that both retrieval mechanisms could be used at once.7. Chapter 2
You also created domain graphs using structured outputs defined through pydantic objects.8. Chapter 3
Finally, in Chapter 3, you used entity resolution to extract entities from unstructured text. You used the resulting graph to calculate similarity and identify duplicate nodes.9. Chapter 3
You evaluated the RAG workflow using the ragas framework and RAG-specific metrics like context precision and noise sensitivity.10. Chapter 3
You even added conversation history to your RAG workflow, extracted facts from the conversation, and stored them in your knowledge graph.11. What next?
So where from here? The resources linked will help you continue on your RAG journey.12. Neo4j GraphAcademy
If you'd like to learn more about Neo4j, graph databases, graph theory, or take applied Graph RAG courses, check out Neo4j's GraphAcademy.13. Let's practice!
Congratulations again and thank you for taking the course!Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.