This repository reads the Neo4j documentation and GraphAcademy content and stores the information as (:Page) and (:Section) nodes in a Neo4j Database.
The chunks.py file searches for (:Section) nodes in the database that do not have embedding, calls the OpenAPI API to generate an embedding.
These embeddings can be used to find the sections in the documentation that are the most similar a user’s input and passed to an LLM to generate an appropriate response.
This data is used to power the Chatbot on Neo4j GraphAcademy.