Skip to main content

Week 2

Milestones

  • Create cQube specific prompt to fetch details like related table and prompt to get sql query based
  • Create module to get subset of schema from the sql file
  • Integrate cQube datasets
  • A standalone application (chat interface) that connects to the cQube that allows for querying of data from cQube and returns the response as a table/big number/chart depending on user request.
  • An autocomplete to suggest users sample queries that they can do.

Screenshots / Videos

Query and data with graph Query and data without graph

Video

Contributions

Learnings

  • ChatGPT 3 and 3.5 model has a small token limit as compared to ChatGPT 4.0.
  • For chatbot the temperature of the models needs to be set in a way such that the results are deterministic.
  • The prompt engineering determines how good/bad the result produced by chatGpt will be.