Query and display PDFs with queries

Hey guys!

I have little experience with vector databases. I have an application that wants to make queries to documents to solve doubts using gpt API, I would like the answers to be accompanied with the main document, for example, a PDF. Is it possible to do this with weaviate? That is, do a query and return not only the corresponding text chunk, but also show the PDF where the text comes from, if possible, even the page where the answer was obtained.

What would be the best approach to achieve something like this?
Thank you!

Hi @Annz and welcome!

You could save the document source URL as a property or the whole document as a blob.

You can then choose to not vectorise certain fields (see this page). Which means that these fields like the URL won’t affect your vector, but can be retrieved.

This page shows you how (text) objects are converted into vectors (Retrievers & Vectorizers | Weaviate - vector database).

Cheers!
JP