update_021924.md

Update

• Implemented first version of Knowledge Graph RAG
  • LlamaIndex for graph framework
  • Nebula Graph for graph store
• Found possibility for integrating memory stream in LlamaIndex
  • We can modify ChatMemoryBuffer interface to store the number of requested words and dates.
• Experienced both graph stores; Neo4j and Nebula
  • Neo4j
    • It was easy to visualize knowledge graphs on web page
    • You can't visualize the graph on notebook jupyter directly.
  • Nebula
    • You can visulaize the graph on notebook jupyter.
    • It is incompatible with Knowledge Graph Retriever on LlamaIndex v0.10.6. I should try to update the version.
  • Tested integrating Chat engine with streamlit for displaying chats and newly generated subgraphs. But I found an undefined error in llama-index. I think LlamaIndex is unstable yet and they are working on that agilely.
  • Check the validity of RAG codes on Harry Potter book writings.

    %%ngql
    USE harry_potter_example;
    MATCH p=(n)-[*1..2]-()
    WHERE id(n) IN ['Dursley', 'Owls', 'Berkeley']
    RETURN p LIMIT 100;
    

    I asked about the relationship of Dursley, Owls and Berkeley.
    • Extract chapter 1 and 2 => N=10, E=7
    • Extract chatper 1, 2 and 3 => N=11, E=9

Todo

• Modify Nebula_graph_store_tutorial.ipynb to be easily produce Knowledge graphs
• Implement the memory stream structures on ChatMemoryBuffer
• Test Neo4j of compatible with Knowledge Graph Retriever