list index out of range on non-XML fileGrp about browse-ocrd HOT 5 CLOSED

309b83c

thanks!

Hmm, it does work for image-only fileGrps now, but still happens if some fileGrp misses a pageId (as soon as I step to that page). Re-open or new issue?

from browse-ocrd.

Ah, ok. So the fileGrp has no PAGE-XML entry for a particular page_id, right?

from browse-ocrd.

So the fileGrp has no PAGE-XML entry for a particular page_id, right?

Exactly! (Happens when using -g selectively during processing, or adding pages to a workspace later-on.)

from browse-ocrd.

Can you check with the latest commit fd02515 ?
I found it difficult to reproduce a crash, I tried it here

Browsing to a fileGrp without PAGE-XML should now give you a warning:
11:04:37.785 WARNING ocrd_browser.model.document - No PAGE-XML and no image for page 'PHYS_0020' in fileGrp 'OCR-D-GT-PAGE'
and that should be it.

If the problem still persists, an example workspace would be nice.

from browse-ocrd.

Can you check with the latest commit fd02515 ?

Yes, this works perfectly – thanks a lot!

I found it difficult to reproduce a crash, I tried it here

browse-ocrd/tests/model/test_document.py

Line 92 in fd02515

def test_page_for_id_with_nothing_for_page_and_fileGrp(self):

Indeed, this is a correct way to simulate missing pages. (You also could have created that workspace dynamically, by removing the page ad-hoc in the test, but never mind.)

Browsing to a fileGrp without PAGE-XML should now give you a warning:

It does exactly that.

from browse-ocrd.