DCAT profiles in GeoNetwork

[maudetes]

Hello!

We've seen latest works on DCAT mapping refactoring and we're very interested by these developments, in particular for GeoDCAT-AP. We've tested applying the SEMICeu XSLT directly in udata as well.

We would have some questions and are curious on the following points:

• Will the mapping be exactly based on SEMICeu XSLT for GeoDCAT-AP or are you going to maintain a version adapted for GeoNetwork?
• As an example, on the distribution aspect, the SEMICeu XSLT tends to specify distribution as dcat:landingPage (instead of creating a dcat:Distribution) if gmd:CI_OnlineFunctionCode / gmd:function is not specified (as described in GeoDCAT-AP). This value is recommended in INSPIRE but may be missing today from current records, leading to distribution being incorrectly represented. Is this something you've observed and may want to take action on (either at the GeoDCAT-AP mapping level or directly on the ISO model)?
• Have you seen other potential issues during your implementation with this XSLT?
• If any issue is found with the mapping and you can't do without modifying it, what would be the strategy of action on GeoNetwork's side? Would you rather ask the SEMICeu for a new version and/or patch in GeoNetwork (temporary or not) ?

Thank you in advance for your feedback on this!

[fxprunayre]

First note that there is a revision of GeoDCAT-AP ongoing with DCAT3 (https://joinup.ec.europa.eu/collection/semic-support-centre/event/introductory-webinar-revision-geodcat-ap). The XSLT will probably be updated in that context.

Will the mapping be exactly based on SEMICeu XSLT for GeoDCAT-AP

The SEMICeu XSLT conversion is included in #7600 with minor improvements (https://github.com/SEMICeu/iso-19139-to-dcat-ap/pulls). This conversion is from ISO19139 to RDF and if needed a conversion from ISO19115-3 is applied.

or are you going to maintain a version adapted for GeoNetwork?

#7600 propose to have a base conversion for DCAT and then extend it for the various profiles. It also focus first on ISO19115-3. So it does not only focus on INSPIRE domain. It is quite inspired by SEMICeu mapping

As an example, on the distribution aspect, the SEMICeu XSLT tends to specify distribution as dcat:landingPage

About distribution, there is room for improvement but current proposal creates a distribution if there is none of the following function 'information', 'search', 'completeMetadata', 'browseGraphic', 'upload', 'emailService'
(https://github.com/geonetwork/core-geonetwork/pull/7600/files#diff-5fa6856eb0e0615450b025fa6ac293be136ee98f4b82b908997ebf6667438ef8R128)

Have you seen other potential issues during your implementation with this XSLT?

Yes, see https://github.com/SEMICeu/iso-19139-to-dcat-ap/pulls

If any issue is found with the mapping and you can't do without modifying it, what would be the strategy of action on GeoNetwork's side? Would you rather ask the SEMICeu for a new version and/or patch in GeoNetwork (temporary or not) ?

Depends probably if you only focus on INSPIRE records or not and ISO19139 and/or ISO19115-3.
At least we can propose PR to the XSLT and follows the ongoing revision work.