Add support for DCAT-AP v3

[jahow]

Implements #15

Remaining to do:

• Update templates
• Try it out in the editor
• Maybe do some kind of automatic upgrading to v3 on update-fixed-info?

[jahow]

Note: I could not find an actual record describing a DatasetSeries so this is purely theoretical for now; open to any suggestions/help, thanks!

[CMath04]

Is there a need for keeping DCAT-ap 2 compatibility in the editor ?
Since DCAT 2 and 3 are pretty similar, maybe the update-fixed-info.xsl should automatically migrate any DCAT-ap 2 records being edited.

[CMath04]

I believe these elements should be in the dcat namespace.
As far as I understood, the dct:next should be at the dcat:Dataset level, alongs with a dcat:prev elements.

[CMath04]

Reading a bit through the vocabulary specification (https://www.w3.org/TR/vocab-dcat-3/#dataset-series-specification) and the application profile implementation (https://semiceu.github.io/DCAT-AP/releases/3.0.0/),

From my understanding, the idea is to have this type of RDF :

<rdf:RDF ...>
    <dcat:Catalog>
        <dcat:record>
            <dcat:CatalogRecord>
                <!-- some record -->
            </dcat:CatalogRecord>
        </dcat:record>
        <dcat:dataset>
            <dcat:DatasetSeries rdf:about="https://example.com/dataset/some-data-collected-annually">
                <dct:title xml:lang="en">Some data collected annually</dct:title>
                <dct:description xml:lang="en">With a description</dct:description>
                <!-- Any other field from dcat:Dataset class -->

                <dcat:first rdf:resource="https://example.com/dataset/some-data-collected-for-year-2020"/>
                <dcat:last rdf:resource="https://example.com/dataset/some-data-collected-for-year-2022"/>
            </dcat:DatasetSeries>
        </dcat:dataset>
        <dcat:dataset>
            <dcat:Dataset rdf:about="https://example.com/dataset/some-data-collected-for-year-2020">
                <dct:title xml:lang="en">Some data collected for year 2020</dct:title>
                <!-- Any other field from dcat:Dataset class -->

                <dcat:inSeries rdf:resource="https://example.com/dataset/some-data-collected-annually"/>
                <dcat:next rdf:resource="https://example.com/dataset/some-data-collected-for-year-2021"/>
            </dcat:Dataset>
        </dcat:dataset>
        <dcat:dataset>
            <dcat:Dataset rdf:about="https://example.com/dataset/some-data-collected-for-year-2021">
                <dct:title xml:lang="en">Some data collected for year 2021</dct:title>
                <!-- Any other field from dcat:Dataset class -->

                <dcat:inSeries rdf:resource="https://example.com/dataset/some-data-collected-annually"/>
                <dcat:prev rdf:resource="https://example.com/dataset/some-data-collected-for-year-2020"/>
                <dcat:next rdf:resource="https://example.com/dataset/some-data-collected-for-year-2022"/>
            </dcat:Dataset>
        </dcat:dataset>
        <dcat:dataset>
            <dcat:Dataset rdf:about="https://example.com/dataset/some-data-collected-for-year-2022">
                <dct:title xml:lang="en">Some data collected for year 2022</dct:title>
                <!-- Any other field from dcat:Dataset class -->

                <dcat:inSeries rdf:resource="https://example.com/dataset/some-data-collected-annually"/>
                <dcat:prev rdf:resource="https://example.com/dataset/some-data-collected-for-year-2021"/>
            </dcat:Dataset>
        </dcat:dataset>
    </dcat:Catalog>
</rdf:RDF>

Now, we don't necessarily need for this to be a single metadata record in GeoNetwork.
Hence, I see two big approaches for implementing this:

• Option 1 is to have single metadata records containing everything at once. The md editor would list all the datasets in a single tab (or many tabs) and the user would be able to edit all of them at once.

• Option 2 is to have a separated standalone dcat:DatasetSeries metadata record and links between all datasets by using.

I feel option 2 offer the more flexibility, especially for indexing, but can get quite complex. For every linking operations, the links between series and datasets (dcat:inSeries <=> dcat:first & dcat;last) need to be updated. As well as maintain all the links between related datasets (dcat:prev & dcat:next).
On the other hand, option 1 seems easier to use from user perspective but might result in giant metadata records and it not easy to opt in from existing dataset record which need to be fully migrated. It also get hard to correctly index.