docling-project / docling-project/docling

[Bee] Incorrect reading order for Key Value regions

Open
#3,773 2 comments 0 reactions 0 assignees View on GitHub
enhancement
Dominant language
Python
Stars
66.4k
Forks
4.8k
Avg merge
3d 4h
Merged PRs (30d)
95

Description

### Requested feature
For some content in the PDF document, it includes multiple columns or blocks with key and value regions, and the structure is not recognized as tables; the reading order should be key first, value second, one by one, but in the markdown results, they are mixed together.

For example:
file content
Image

result:

Image

There are several testing documents:

1. Summary section in [test_kvp_reading_order.pdf](https://github.com/user-attachments/files/29789718/test_kvp_reading_order.pdf), programmatic PDF
2. form section in [AB.FILES.038670001-02-PDF_A_MockUp-page-5.pdf](https://github.com/user-attachments/files/29789801/AB.FILES.038670001-02-PDF_A_MockUp-page-5.pdf), Form data, scanned PDF, but the section is not recognized as table.
Image

Contributor guide

Open the contributing guide

Research direction

Start by reproducing the markdown output with test_kvp_reading_order.pdf and AB.FILES.038670001-02-PDF_A_MockUp-page-5.pdf, comparing the reported key/value regions with their expected reading order. Trace the PDF-to-markdown reading-order entry point and add or update regression coverage so each key is followed by its corresponding value in both examples.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
data
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.