Maintaining Header Hierarchy in Markdown Output for Enhanced Section Splitting

[naufalso]

Hello,

First, I want to express my gratitude to the team for creating such an impressive tool! Docling has been incredibly useful in converting documents to Markdown format with ease and precision. Your efforts in building this robust solution are greatly appreciated.

While using Docling, I noticed that the Markdown output consistently uses second-level headings (##) for splitting sections. This approach works well for many scenarios, but I wonder if it is intended behavior.

For my use case, preserving the original header structure of a PDF (e.g., chapters, sections, and subsections) in the Markdown output is essential. Maintaining this hierarchy would allow for more nuanced data splitting by headers while keeping the context intact. This feature would be particularly useful when leveraging tools like LangChain's MarkdownHeaderTextSplitter.

Is there a way to configure Docling to maintain the original document's header hierarchy in the Markdown output? If this feature isn’t currently available, are there plans to support it in future updates?

Thank you again for your excellent work and for considering this feature request!

[PeterStaar-IBM]

@naufalso This is high on our list as well. We have the hierarchy for docx and html, and are now working on adding it to the pdf. The problem with the latter is that section-headers are detected via object detection, and we have a-priori no information what the level is.

We are trying to first use the table-of-contents in pdf, but hopefully soon, we will have a more dedicated model for this.

[PeterStaar-IBM]

That would be great!

[naufalso]

Hello,

I wanted to check if there are any updates regarding this issue. I’ve noticed that other PDF parser libraries, like Marker or MinerU, may have addressed this functionality.

Thank you.

[PeterStaar-IBM]

Yes, there is a lot of progress on the back. We are working heavily to do first good eval sets and methods. Expect updates in the next weeks

[naufalso]

Thank you for the update! It's great to hear about the progress and the focus on creating quality evaluation sets and methods. I look forward to hearing more in the coming weeks.

[rahepler2]

Any movement on this? I'm still having issues where the headers from children are the same ## level as their parent.

[PeterStaar-IBM]

@rahepler2 yes, we believe that our new VLM models (follow up on SmolDocling) should start addressing this problem holistically.

[rahepler2]

@PeterStaar-IBM Awesome, thanks Peter! Will this be a roadmap item in the future, or something that would be rolled out in the near future? No rush on response, just curious.

[ArashJavan]

@PeterStaar-IBM thank you for the excellent work on Docling. We're very interested in understanding the current progress and timeline for this feature, as Docling serves as the backbone of our document processing pipeline.

Regarding the ToC (Table of Contents) extraction you mentioned: we're currently planning to implement this functionality ourselves, but if I understand correctly, this is something on your roadmap? If so, could you share:

1. What methodology you're planning to use for ToC extraction
2. An estimated timeline for when this feature might be released

This would help us decide whether to wait for the official implementation or proceed with our own solution. Thank you!

[BEN-50763]

Hey team, love the library.

Also tracking for this solution, levelled hierarchy would be incredible, #1 feature request for us too.

VLM solution sounds interesting, could you please elaborate?