A persistent graph index? DRIFT Search?

[S3tLat3ral]

This is a great implementation using langchain, thanks for sharing it! I have some questions:

• how can we index on a continuous basis, i.e. indexing takes new text docs and merges it into the same reusable graph?
• how/will you be implementing the ms version's DRIFT Search?

[ksachdeva]

Thanks for the questions and feedback. I haven't looked at DRIFT yet but may do it in coming weeks.