docling-project / docling-project/docling
Markdown header levels are all the same
- Dominant language
- Python
- Stars
- 66.4k
- Forks
- 4.8k
- Avg merge
- 2d 21h
- Merged PRs (30d)
- 84
Description
### Bug
This is probably not a real bug but a feature request:
All headers in the generated markdown files are at the same level, i.e. ## .
For example, in the sample PDF to markdown conversion, section 3 has multiple sub-sections:
## 3 Processing pipeline
....
## 3.1 PDF backends
...
## 3.2 AI models
The correct header levels really should be something like this:
## 3 Processing pipeline
....
### 3.1 PDF backends
...
### 3.2 AI models
...
It would be great if the header level can be fixed, because this can keep the hierarchical structure of the original article.
### Steps to reproduce
Just run the sample code of this repo.
...
### Docling version
...
### Python version
...
Contributor guide
Assessment
This issue has not been assessed yet.