[Feature]: RAG from PDF File

[areytechai]

Background & Description

is there a way to use rag methodology with PDF or docx file read from filesystem

API & Usage

No response

How to implement

No response

[zsogitbe]

Yes, you need a library that can extract text from PDF and DOCX files (there are free libraries available), and then you can use the standard procedure for Retrieval-Augmented Generation (RAG) in LLamaSharp.

[areytechai]

@zsogitbe example of procecudure pleaes

[zsogitbe]

1. Extract text from a PDF using a free C# tool.
2. Chunk the extracted text for RAG.
3. Save the text chunks into Kernel memory with LLamaSharp.
4. Perform searches within the memory using LLamaSharp.

The quality will depend on the efficiency and cleverness of executing the former steps (it is not as easy as many novice AI 'experts' think).

I am willing to provide an example in exchange for payment (willing to give a quote for this based on your requirements). Otherwise, please look at the basic examples in the LLamaSharp repository.