You could try ingesting your files with open-parse. It analyzes the layout of your documents, converts to markdown and extract tables. Then you could feed this

You could try ingesting your files with open-parse. It analyzes the layout of your documents, converts to markdown and extract tables. Then you could feed this output to a small model like gpt-3.5 and it should be able to classify the topic quite easily. Alternatively you could convert the pdf to images and feed that to a model. Or use naive text splitting implemented by something like LlamaIndex. Ton of options! Disclaimer: (I’m the author) https://github.com/Filimoa/open-parse https://github.com/Filimoa/open-parse