feat: Make ThreadedDoclingParseDocumentBackend the default

[PeterStaar-IBM]

Checklist:

• Documentation has been updated, if necessary.
• Examples have been added, if necessary.
• Tests have been added, if necessary.

[github-actions]

✅ DCO Check Passed

Thanks @PeterStaar-IBM, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🔴 1 of 2 protections blocking · waiting on 👀 reviews

	Protection	Waiting on
🔴	Require two reviewer for test updates	👀 reviews
🟢	Enforce conventional commit	—

🔴 Require two reviewer for test updates

Waiting for

• #approved-reviews-by >= 2

This rule is failing.

When test data is updated, we require two reviewers

• #approved-reviews-by >= 2

---

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[codecov]

Codecov Report

❌ Patch coverage is 90.90909% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
docling/backend/docling_parse_backend.py	87.50%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[cau-git]

Testing with docling-parse 7.8.0 on docling 2.112.0 (2026-07-15)

Below the results of an evaluation run on the first 1000 doclaynet test set samples.

Non-threaded docling-parse backend (using pypdfium renderer)

================================================================================
CONVERSION THROUGHPUT
================================================================================
Pages: 1000
Wall time: 842.12s
Rate: 1.19 pages/s

================================================================================
TEXT METRICS
================================================================================
+-----------+------+-----------+----------+-------------+------------+---------------+--------------+------------------+-----------------+-------------+------------+---------------+--------------+
| Dataset   |    n |   f1 mean |   f1 med |   prec mean |   prec med |   recall mean |   recall med |   edit_dist mean |   edit_dist med |   bleu mean |   bleu med |   meteor mean |   meteor med |
+===========+======+===========+==========+=============+============+===============+==============+==================+=================+=============+============+===============+==============+
| dla/test  | 1000 |     0.876 |    0.963 |        0.89 |      0.965 |         0.888 |         0.97 |            0.235 |           0.122 |       0.736 |      0.864 |         0.826 |        0.941 |
| TOTAL     | 1000 |     0.876 |    0.963 |        0.89 |      0.965 |         0.888 |         0.97 |            0.235 |           0.122 |       0.736 |      0.864 |         0.826 |        0.941 |
+-----------+------+-----------+----------+-------------+------------+---------------+--------------+------------------+-----------------+-------------+------------+---------------+--------------+

================================================================================
TABLE METRICS (TEDS)
================================================================================
+-----------+-----+-------------+------------+---------------------+--------------------+
| Dataset   |   n |   teds mean |   teds med |   teds_content mean |   teds_content med |
+===========+=====+=============+============+=====================+====================+
| dla/test  | 508 |       0.716 |      0.894 |               0.591 |               0.72 |
| TOTAL     | 508 |       0.716 |      0.894 |               0.591 |               0.72 |
+-----------+-----+-------------+------------+---------------------+--------------------+

================================================================================
TABLE METRICS (TEDS, MATCHED ONLY)
================================================================================
+-----------+-----+-------------+------------+---------------------+--------------------+
| Dataset   |   n |   teds mean |   teds med |   teds_content mean |   teds_content med |
+===========+=====+=============+============+=====================+====================+
| dla/test  | 428 |        0.85 |      0.926 |               0.701 |              0.763 |
| TOTAL     | 428 |        0.85 |      0.926 |               0.701 |              0.763 |
+-----------+-----+-------------+------------+---------------------+--------------------+

Threaded docling-parse backend (using native renderer)

================================================================================
CONVERSION THROUGHPUT
================================================================================
Pages: 1000
Wall time: 805.47s
Rate: 1.24 pages/s

================================================================================
TEXT METRICS
================================================================================
+-----------+------+-----------+----------+-------------+------------+---------------+--------------+------------------+-----------------+-------------+------------+---------------+--------------+
| Dataset   |    n |   f1 mean |   f1 med |   prec mean |   prec med |   recall mean |   recall med |   edit_dist mean |   edit_dist med |   bleu mean |   bleu med |   meteor mean |   meteor med |
+===========+======+===========+==========+=============+============+===============+==============+==================+=================+=============+============+===============+==============+
| dla/test  | 1000 |     0.857 |    0.953 |        0.86 |      0.954 |         0.886 |         0.97 |            0.271 |           0.154 |       0.701 |      0.838 |         0.809 |        0.921 |
| TOTAL     | 1000 |     0.857 |    0.953 |        0.86 |      0.954 |         0.886 |         0.97 |            0.271 |           0.154 |       0.701 |      0.838 |         0.809 |        0.921 |
+-----------+------+-----------+----------+-------------+------------+---------------+--------------+------------------+-----------------+-------------+------------+---------------+--------------+

================================================================================
TABLE METRICS (TEDS)
================================================================================
+-----------+-----+-------------+------------+---------------------+--------------------+
| Dataset   |   n |   teds mean |   teds med |   teds_content mean |   teds_content med |
+===========+=====+=============+============+=====================+====================+
| dla/test  | 508 |       0.542 |       0.76 |               0.437 |               0.54 |
| TOTAL     | 508 |       0.542 |       0.76 |               0.437 |               0.54 |
+-----------+-----+-------------+------------+---------------------+--------------------+

================================================================================
TABLE METRICS (TEDS, MATCHED ONLY)
================================================================================
+-----------+-----+-------------+------------+---------------------+--------------------+
| Dataset   |   n |   teds mean |   teds med |   teds_content mean |   teds_content med |
+===========+=====+=============+============+=====================+====================+
| dla/test  | 330 |       0.835 |      0.901 |               0.673 |              0.729 |
| TOTAL     | 330 |       0.835 |      0.901 |               0.673 |              0.729 |
+-----------+-----+-------------+------------+---------------------+--------------------+

Conclusions

• Text metrics with threaded docling-parse look slightly worse but not by much
• Table metrics with threaded docling-parse look slightly worse for those tables that could be matched
• Critical: By far less tables could be matched with threaded docling-parse than with classic serial version (330 vs 428 out of 508 toal). This implies that the threaded docling-parse renderer creates output in which the current layout model often misclassifies tables. Needs further investigation.