Support Multimodal Input Datasets in Carnot Chat and AI File Search #33
Description
Updated Text
I would like to expand the scope of this issue to not only include support for PDF file(s) but image and audio files as well. Essentially, users should be able to select any file type(s) (that we can read) in their Dataset(s) and Carnot should be able to process that dataset.
Original Text
Enable the chat workflow to ingest and analyze PDF files alongside existing text files. The goal is for users to include PDFs in datasets, reference them from the chat UI, and have Carnot run combined queries across mixed file types.
Same for AI File search in Data Management Page