I have a general question surrounding the use of Nebula as a Graph Rag. I’ve read the blogs which are awesome. I’ve. come to the realization that it takes documents chunks vectorizes them and stores them in a traditional vector database, additionally graphs themselves are then encoded and vectorized into the same vector database. Am I right to conclude that the first set of documents aren’t actually stored in the graphs? Just trying to wrap my head around what an integration with my KG and Causal Graphs might look like using Nebula. Today we use TypeDB from Vaticle which has served us well as our graph. We are currently working with Flowise which is based on LangChain and LLamaindex and it would be awesome if someone could write a Flowise component for Nebula