I want to build a RAG workflow in n8n that can extract answers from documents and return the document itself, while highlighting the relevant part or providing the page number, paragraph, or excerpt where the information was found. Does anyone have a tutorial, example, or workflow for this?