RAG (Retrieval Augmented Generation)

Here we want to build a RAG system addressing the real-life problem of working with data in stylised PDFs. These PDFs often contain graphs and figures which are crucial to understanding, and our RAG ingestion system should be able to handle them.

Here is an example of a stylised PDF, we chose the Nedbank annual report 2024 to have a go at looking at financial data and how to ingest it into a RAG system.