wip for better csw-dcat output for data.gouv.fr

[landryb]

cf georchestra/georchestra#4182, PR mostly for diff readability of https://github.com/georchestra/georchestra/files/14127940/tpl-rdf.xsl.txt, all work from @jeanpommier

geOrchestra/geonetwork checklist

• PR only involves cherry-picked commits from upstream.
• PR contains custom code which will soon be available in an upstream release and can be overriden => mention core-geonetwork version if possible.
• PR contains custom geOrchestra code, which need to be verified during future migrations.
  • I have properly filled the migration-helper-changelog.md file.

[landryb]

note for $self: this is to solve Description des données non renseignée on harvested data, which comes from the attached resource description

[landryb]

/datasets doesnt map to anything in geonetwork under resources?

[landryb]

after discussion with @jeanpommier this section might not be needed ?

[landryb]

the complete section about legalConstraints should be checked with @MaelREBOUX to know if we use the right method to fill the licensing info

[landryb]

so regarding licences mapping, i've did a bunch more tests, and with what we have now in the PR, if we have this for a md:

<gmd:MD_LegalConstraints>
  <gmd:accessConstraints>
    <gmd:MD_RestrictionCode codeList="...#MD_RestrictionCode" codeListValue="otherRestrictions"/>
  </gmd:accessConstraints>
  <gmd:otherConstraints>
    <gco:CharacterString>LO/OL</gco:CharacterString>
  </gmd:otherConstraints>

we do end up with <dct:license>LO/OL</dct:license> in the dcat output, and if we have

<gmd:MD_LegalConstraints>
  <gmd:useConstraints>
    <gmd:MD_RestrictionCode codeList="...#MD_RestrictionCode" codeListValue="otherRestrictions"/>
  </gmd:useConstraints>
  <gmd:otherConstraints>
    <gco:CharacterString>LO/OL</gco:CharacterString>
  </gmd:otherConstraints>

we end up with <dct:accessRights>LO/OL</dct:accessRights> in the dcat output. We can of course use both accessConstraints & useConstraints to get dct:accessRights and dct:licence but from https://doc.data.gouv.fr/moissonnage/dcat/#jeu-de-donn%C3%A9es i think dct:licence matters first for udata.

im not the specialist to judge which one makes sense, but what matters is from my reading of https://guides.data.gouv.fr/guide-data.gouv.fr/moissonnage/les-differents-types-de-moissonneurs#detection-des-licences-par-le-moissonnage if we have a value (LO/OL here) that matches one of the id/title/url & alternates from https://www.data.gouv.fr/api/1/datasets/licenses/, then udata will recognize it.

see for example: https://demo.georchestra.org/datahub/dataset/faa240f5-90a8-4e1f-b8da-39c913bed9df
as dcat: https://demo.georchestra.org/geonetwork/csw-dcat-singlemd/eng/csw?SERVICE=CSW&VERSION=2.0.2&REQUEST=GetRecordById&outputSchema=http://www.w3.org/ns/dcat%23&id=faa240f5-90a8-4e1f-b8da-39c913bed9df
and finally, as harvested by udata: https://demo.data.gouv.fr/fr/datasets/fiche-parfaite-pour-data-gouv/ - note that the licence is properly recognized as Licence Ouverte / Open Licence

as for the xsl voodoo, i don't really understand the indirection done through otherRestrictions/otherConstraints, it's only to be able to have the value coming from a free-form field instead of a select box in geonetwork ?

eg if i select licence for access constraints i have this in the ISO:

<gmd:accessConstraints>
  <gmd:MD_RestrictionCode codeList="...#MD_RestrictionCode" codeListValue="license"/>

which traduces to an empty <dct:accessRights/> as the MD_RestrictionCode has no child in the XML.

[landryb]

to note for the license mapping, ecolab has this: ecolabdata/ecospheres-core-geonetwork@6724d09