docling-project / docling-project/docling
[Bee] Improve table extraction with multi-level row spanning
Open
enhancement
table structure
triage/close-stale
- Dominant language
- Python
- Stars
- 66.4k
- Forks
- 4.8k
- Avg merge
- 2d 21h
- Merged PRs (30d)
- 84
Description
### Requested feature
The table on this scanned PDF page has three kinds of rows with different sizes; the table extraction result should be improved. Especially for the first column, the single block is segmented into multiple rows, which is wrong.
### Extraction result in md:
...
test file:
[yanbaopptmerge_0c79d327060dbf9f1582d03c235dadb039533a19091d2c0d24f2ad95d267f79b.pdf_2.pdf](https://github.com/user-attachments/files/26109091/yanbaopptmerge_0c79d327060dbf9f1582d03c235dadb039533a19091d2c0d24f2ad95d267f79b.pdf_2.pdf)
...
Contributor guide
Assessment
This issue has not been assessed yet.