Find answers from the community

Updated 2 months ago

Hey all, i have a general question. I am

Hey all, i have a general question. I am looking at creating a RAG application using all our issues on slack. Each issues usually has a stack trace and general error message, the script that was in error and a bunch of other data. I was thinking of developing a streamlit application to process each message. It would display the message and then i can add a resolution to it. in a stepwise fashion. I thne want to store the issue and resolution in a vector database but not sure whether i should attach the resolution as metadata or just chunk the document with the issue and resolution before storing it in the database. Should I chunk at all or store the stack trace as a single document. Would appreciate any insights
W
j
4 comments
I think if you have the data in the particular format like

Issue/Query:
Resolution:

Then you should create Node for each issue separately. This will in turn will help embedding model to identify nodes more correctly.
And you can keep tags related to each issue in the Metadata like maybe keywords, flags etc
Add a reply
Sign up and join the conversation on Discord