Load a .md markdown file with the Default Data Loader

Questions
1

Hi.
I upload a file and pipe it as input to the Pinecone Vector Database. This uses the Default Data Loader for “Document”. It works great for PDF and TXT documents. But how do I input MD files? I always get a MIME error with MD.
It works when I rename them to TXT.

This is important because properly structured MD files are better split into sensible chunks by the “Recursive Character Splitter” with selected option “Markdown”.

2

Hi @Ralf011, welcome to community.
Can you share your workflow here please? Just to get the correct reproducible usage

3
4

Trying to optimize my RAG by including the reranker, which seems to work best when the chunks were split according to markdown structure. I use Docling to convert my PDF into structured markdown and then le the “Recursive Character Text Splitter” consider the MD structure. Simple example attached.

5

shared below. any help?

6

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.