Generation of dcat:Catalog instance

[bfrichet3]

Dear Andrea,

I hope you are all fine. I have some questions concerning the way you create (or possibly you do not create yet) the Catalog Class by using INSPIRE records.

As far as I see the way we will implement the conversion tools from INSPIRE to DCAT, I guess there will have two ways to create some DCAT AP feeds from INSPIRE records:

1. by converting a single INSPIRE record into a DCAT Feed containing one Dataset Class;
2. by converting our whole CSW into a DCAT Feed contaning many datasets/services class.

In the first case, in order to be compliant with the DCAT (AP) specifications, you still need a Catalog class. I discovered by using your API that there is no Catalog in the created feed if you convert one INSPIRE record describing a dataset into a DCAT Feed.

So this is my first question: with wich semantical informations would recommand to feed the needed class? By using the informations providing from one CSW giving access to that record even though the CSW is much broader than the generated Catalog Class? By some hard-coded informations? Or you dont recommand to create a DCAT Feed from one single INSPIRE record?

In the second case, I read in the Mappings.md (https://github.com/SEMICeu/iso-19139-to-dcat-ap/blob/master/documentation/Mappings.md) that you create the Catalog Record from the INSPIRE record describing the CSW (or maybe I didn't understand that file good enough). I would have done on the same way. I still have a second specific question about the creation of the Catalog in that case: how would you feed the optional dct:identifier attribute of that class?

There are no resource identifiers in the INSPIRE service records and even if resource identifiers are possible for services following ISO 19115 I would not use that element to feed the dct:identifier of the Catalog Class. I could use that ISO 19115 element to feed dct:identifier of Data Service Classes created from INSPIRE records describing WMS/WFS/ATOM/.... because the resources are the same: a WMS described in a DCAT Feed or in an ISO record is still the the sameWMS so we can use the ISO identifier to feed the dct:identfier of the Data Service Class. But I wouldn't use that element to feed dct:identifier of a Catalog Class because a CSW is really different from a Catalog Class of a DCAT Feed. So in that case how would you identify a Catalog Class in the DCAT feed? With the access URL of the published DCAT Feed?

I hope my questions are specifi enough. We have been thinking about these points for a while and I wanted to get your opinion about these.

Regards,

Benoît

[bfrichet3]

Dear @andrea-perego,

Do you need extra informations or do you think that question is not really relevant?

Regards,

[andrea-perego]

You are right, @bfrichet3 .

The GeoDCAT-AP XSLT transforms only the information provided in the input records. So, it does not create a record for the catalogue itself since no information is provided about it via a GetRecords / GetRecordById request.

So, the definition of a standard solution is still an open issue. It would be interesting to know if anyone has addressed it and how.