ensure valid coordinates by intersection with parent…

[bertsky]

• Border: intersect with page frame
• text regions: intersect with (new) Border
• text lines: intersect with (new) text region
  (and back-transform at all)

This fixes the negative coordinates, and also the line-region inconsistency (line bboxes way beyond tight region outlines).

I'll fix removing the existing derived images with cropped in a PR to core, so this will be transparent and automatic here.

[bertsky]

Oh, and could you please try introducing some versioning (so users get to see what version produced their PAGE/METS annotations)?

[bertsky]

I'll fix removing the existing derived images with cropped in a PR to core, so this will be transparent and automatic here.

OCR-D/core#639

[jbarth-ubhd]

I've tried it again (with make NO_UPDATE=1 all after git remote add bertsky https://github.com/bertsky/sbb_textline_detection) , but get no text simple-layout pages.

Just to be sure I've grepped a string from master...bertsky:fix-coords :

(venv) [hd_xxxxx@login5 sbb_textline_detector]$ pwd
/beegfs/home/hd/hd_hd/hd_wu120/ocrd_all/venv/local/sub-venv/headless-tf1/lib/python3.7/site-packages/qurator/sbb_textline_detector
(venv) [hd_xxxxx@login5 sbb_textline_detector]$ grep -r 'new border would be empty, skipping' .
./ocrd_cli.py:                log.warning("new border would be empty, skipping")
Binary file ./__pycache__/ocrd_cli.cpython-37.pyc matches

So the my venv seems up to date(?)

[jbarth-ubhd]

example: https://digi.ub.uni-heidelberg.de/diglitData/v/ocrd/dobbert1861-empty.tgz

[bertsky]

with make NO_UPDATE=1 all

Did you git -C core pull origin before that? (You need an up-to-date core also in the sub-venv.)

but get no text simple-layout pages.

You get what?

[jbarth-ubhd]

No Text from ... Von meinem iPhone gesendet Am 05.11.2020 um 15:04 schrieb Robert Sachunsky <notifications@github.com>:  with make NO_UPDATE=1 all Did you git -C core pull origin before that? (You need an up-to-date core also in the sub-venv.) but get no text simple-layout pages. You get what? — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#48 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AHJ32U5HZTA4WL2MIOZDZIDSOKWFNANCNFSM4TFBA2FA>.

[jbarth-ubhd]

hmmm...

[hd_xxx@login5 ocrd_all]$ ~/local/bin/git -C core pull origin
warning: Es wird davon abgeraten zu Pullen, ohne anzugeben, wie mit abweichenden
Branches umgegangen werden soll. Sie können diese Nachricht unterdrücken,
...
remote: Enumerating objects: 45, done.
remote: Counting objects: 100% (45/45), done.
remote: Compressing objects: 100% (26/26), done.
remote: Total 45 (delta 24), reused 34 (delta 19), pack-reused 0
Entpacke Objekte: 100% (45/45), 129.57 KiB | 93.00 KiB/s, Fertig.
Von https://github.com/OCR-D/core
   c8136bf5..ffe4e545  master     -> origin/master
 * [neues Tag]         v2.20.0    -> v2.20.0
Im Moment auf keinem Branch.
Bitte geben Sie den Branch an, welchen Sie zusammenführen möchten.
Siehe git-pull(1) für weitere Details.

    git pull <Remote-Repository> <Branch>

[hd_xxx@login5 ocrd_all]$ make NO_UPDATE=1 all
make: Nothing to be done for `all'.

ok I'll wait till your changes are pulled into ocrd_all.

[jbarth-ubhd]

oops. ~/local/bin/git -C core pull origin master seems to help.

[bertsky]

This fixes the negative coordinates, and also the line-region inconsistency (line bboxes way beyond tight region outlines).

I don't understand why this essential fix has not been merged yet.

[cneud]

Thank you for the contribution! For internal reasons, we have not had the time to properly review this yet. I will put it on the agenda for our next internal meeting.