Fix CI

[mweidling]

The mets.xml of the grenzboten-test directory uses a prefix without corresponding dir in fileGrp/@USE. This PR fixes the typo.

[mweidling]

After having altered the OCR-D- prefix I get the following error for the gutachten dir:

<report valid="false">
  <error>PAGE_TEMP1: Line 123: Element '{http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}UnorderedGroupIndexed': Missing child element(s). Expected is one of ( {http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}UserDefined, {http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}Labels, {http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}RegionRef, {http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}OrderedGroup, {http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15}UnorderedGroup ).</error>
  <notice>fileGrp USE does not begin with 'OCR-D-': IMG</notice>
  <notice>fileGrp USE does not begin with 'OCR-D-': TEMP1</notice>
  <notice>fileGrp USE does not begin with 'OCR-D-': TEMP2</notice>
</report>

[mweidling]

The reason for the error mentioned above seems to be https://github.com/OCR-D/assets/blob/master/data/gutachten/data/TEMP1/PAGE_TEMP1.xml#L123. Removing it results in a successful report.

@kba Is this real data that shouldn't be altered or can we just add the missing child attribute?

[kba]

The reason for the error mentioned above seems to be https://github.com/OCR-D/assets/blob/master/data/gutachten/data/TEMP1/PAGE_TEMP1.xml#L123. Removing it results in a successful report.

@kba Is this real data that shouldn't be altered or can we just add the missing child attribute?

No, we actually need that in the test_empty_groups_to_regionrefindexed test in tests/model/test_ocrd_page.py. We might want to exclude the gutachten folder from validation in the Makefile, i.e. adding -not -name "gutachten" to https://github.com/OCR-D/assets/blob/master/Makefile#L51

But the OCRD vs OCR-D typo is spot on, can be merged.

Unfortunately, I haven't updated the assets in core in a while and there are some issues now unrelated to this PR.