Dynamic document analysis with generated vector and embedding model

[themac13]

New to this world - I am able to get ingest to load about 3k files. Querying through the CLI and the web UI produces great results.

I wanted to see if I can upload a new document(s) on the fly and run a query to the current vector+embedding model?

I am thinking I could always write in python to read a file, put the whole file's text in the question and then add in the query at the end of the text content to summaries the document and call it via the API, and get it to produce an answer that is not only based on the document but also the vector+embedding model?

So for example

query: Summarise this document: "text from harry potter"

Is this the right tool? Is there any suggestion for what I can use?