docling-project / docling-project/docling
from docling v2.73.1 to v2.92.0: “date” removed
- Dominant language
- Python
- Stars
- 66.4k
- Forks
- 4.8k
- Avg merge
- 2d 21h
- Merged PRs (30d)
- 84
Description
### Bug
...
I’ve switched from v2.73.1 to v2.92.0 and compared the results (MD files). Overall, file sizes and the number of chunks have decreased noticeably (by around 20%) without any relevant content being lost, as far as I can tell. It seems to generate less junk/noise, which is great!
There is, however, one important exception:
In v2.73.0, the date from the first pages of reports is correctly converted to MD, whereas in v2.92.0 the date is missing entirely. Unfortunately, this is quite a problem, as the date is one of the most important pieces of metadata (filter) and, generally speaking, is very relevant information that should not be removed.
### Steps to reproduce
...
### Docling version
...
2.92.0
### Python version
...
3.12.12
Contributor guide
Assessment
This issue has not been assessed yet.