Closed oldschoola closed 2 months ago
I'm working on this. Typically, this seems to be a standard RAG use case, and there are several frameworks available to accomplish this. The hottest topic regarding RAG at the moment is GraphRAG, which involves RAG in conjunction with knowledge graphs. This can be done, for example, using LlamaIndex/LangChain in combination with Neo4j.
I implemented a first version to summarize a PDF-file and genrate a mindmap from the summary.
The PDF is converted to markdown first using the pymupdf4llm
library.
I implemented another way using native multimodal capabilities of (VertexAI / Gemini) models to use PDF files.
Is it possible to use this with a pdf?
I see it says "Create a map based on external text data" How would i covert a pdf to the format this can use?