Controlled vocabularies are defined in several places

[mortenwh]

• chapter 4 in the documentation
• files in the thesauri folder
• under xs:simpleType in the xsd/mmd.xsd file

I suggest to write some tests to check that the definitions are internally consistent.

[steingod]

Sounds like a good idea. Not sure how to do this across these sources though?

[mortenwh]

That should be possible. Maybe not easy and may require some coding but I think it would be useful to integrate it in the CI. I'll add it to the S-ENDA backlog, then we can see. Probably not highest priority at the moment, though..

[steingod]

Agree

[mortenwh]

@TAlonglong - I took this one out of the May sprint. It is too complicated to solve in one go now.

[ferrighi]

I suggest to begin with keeping only the thesauri/mmd-vocabulary.ttl as reference and use a git action to generate mmd-vocabulary.xml (as we do for adoc->html/pdf). This can be done with rdflib. I can test this a bit. For the schema, we should probably remove the vocab list from that and import valid vocabularies from the ttl/xml. So we keep the schema formally validating the structure and the vocabulary to validate the content.