Improve serialization of statements

Rather than just include a Literal, add it as a node
(with a class if provided) with a RDFS.label property, eg:

    dct:accessRights [ a dct:RightsStatement ;
	rdfs:label "Statement about access rights" ] ;

An URI can also be used:

    dct:accessRights <https://example.org/vocab/access-right/TODO/PUBLIC> ;

    [...]

    <https://example.org/vocab/access-right/TODO/PUBLIC> a dct:RightsStatement .

This is in line with the recommendation in the DCAT 3 spec:

https://www.w3.org/TR/vocab-dcat-3/#ex-license-and-access-rights

ckanext/dcat/profiles/base.py

@@ -757,6 +757,38 @@ def _add_spatial_to_dict(self, dataset_dict, key, spatial):
                 }
             )
 
+    def _add_statement_to_graph(self, data_dict, key, subject, predicate, _class=None):
+        """
+        Adds a statement property to the graph.
+        If it is a Literal value, it is added as a node (with a class if provided)
+        with a RDFS.label property, eg:
+
+            dct:accessRights [ a dct:RightsStatement ;
+                rdfs:label "Statement about access rights" ] ;
+
+        An URI can also be used:
+
+            dct:accessRights <https://example.org/vocab/access-right/TODO/PUBLIC> ;
+
+            [...]
+
+            <https://example.org/vocab/access-right/TODO/PUBLIC> a dct:RightsStatement .
+        """
+        value = self._get_dict_value(data_dict, key)
+        if value:
+            _object = URIRefOrLiteral(value)
+            if isinstance(_object, Literal):
+                statement_ref = BNode()
+                self.g.add((subject, predicate, statement_ref))
+                if _class:
+                    self.g.add((statement_ref, RDF.type, _class))
+                self.g.add((statement_ref, RDFS.label, _object))
+
+            else:
+                self.g.add((subject, predicate, _object))
+                if _class:
+                    self.g.add((_object, RDF.type, _class))
+
     def _schema_field(self, key):
         """
         Returns the schema field information if the provided key exists as a field in

ckanext/dcat/profiles/euro_dcat_ap_base.py

@@ -282,9 +282,7 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
             ("version", OWL.versionInfo, ["dcat_version"], Literal),
             ("version_notes", ADMS.versionNotes, None, Literal),
             ("frequency", DCT.accrualPeriodicity, None, URIRefOrLiteral, DCT.Frequency),
-            ("access_rights", DCT.accessRights, None, URIRefOrLiteral, DCT.AccessRights),
             ("dcat_type", DCT.type, None, URIRefOrLiteral),
-            ("provenance", DCT.provenance, None, URIRefOrLiteral, DCT.ProvenanceStatement),
         ]
         self._add_triples_from_dict(dataset_dict, dataset_ref, items)
 
@@ -500,6 +498,23 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
             ):
                 resource_license_fallback = dataset_dict["license_url"]
 
+        # Statetements
+        self._add_statement_to_graph(
+            dataset_dict,
+            "access_rights",
+            dataset_ref,
+            DCT.accessRights,
+            DCT.RightsStatement
+        )
+
+        self._add_statement_to_graph(
+            dataset_dict,
+            "provenance",
+            dataset_ref,
+            DCT.provenance,
+            DCT.ProvenanceStatement
+        )
+
         # Resources
         for resource_dict in dataset_dict.get("resources", []):
 
@@ -514,7 +529,6 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
                 ("name", DCT.title, None, Literal),
                 ("description", DCT.description, None, Literal),
                 ("status", ADMS.status, None, URIRefOrLiteral),
-                ("rights", DCT.rights, None, URIRefOrLiteral, DCT.RightsStatement),
                 ("license", DCT.license, None, URIRefOrLiteral, DCT.LicenseDocument),
                 ("access_url", DCAT.accessURL, None, URIRef, RDFS.Resource),
                 ("download_url", DCAT.downloadURL, None, URIRef, RDFS.Resource),
@@ -530,6 +544,15 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
             ]
             self._add_list_triples_from_dict(resource_dict, distribution, items)
 
+            # Statetements
+            self._add_statement_to_graph(
+                resource_dict,
+                "rights",
+                distribution,
+                DCT.rights,
+                DCT.RightsStatement
+            )
+
             # Set default license for distribution if needed and available
 
             if resource_license_fallback and not (distribution, DCT.license, None) in g:

ckanext/dcat/tests/profiles/dcat_ap/test_euro_dcatap_profile_serialize.py

@@ -21,6 +21,7 @@
 from ckanext.dcat.profiles import (
     DCAT, DCT, ADMS, XSD, VCARD, FOAF, SCHEMA,
     SKOS, LOCN, GSP, OWL, SPDX, GEOJSON_IMT,
+    RDFS,
 )
 from ckanext.dcat.profiles.euro_dcat_ap_base import DISTRIBUTION_LICENSE_FALLBACK_CONFIG
 from ckanext.dcat.utils import DCAT_EXPOSE_SUBCATALOGS
@@ -121,8 +122,6 @@ def test_graph_from_dataset(self):
         assert self._triple(g, dataset_ref, OWL.versionInfo, dataset['version'])
         assert self._triple(g, dataset_ref, ADMS.versionNotes, extras['version_notes'])
         assert self._triple(g, dataset_ref, DCT.accrualPeriodicity, extras['frequency'])
-        assert self._triple(g, dataset_ref, DCT.accessRights, extras['access_rights'])
-        assert self._triple(g, dataset_ref, DCT.provenance, extras['provenance'])
         assert self._triple(g, dataset_ref, DCT.type, extras['dcat_type'])
 
         # Tags
@@ -156,6 +155,14 @@ def test_graph_from_dataset(self):
                     _type = item[2][num]
                 assert self._triple(g, dataset_ref, item[1], _type(value))
 
+        # Statements
+        for item in [
+            ('access_rights', DCT.accessRights),
+            ('provenance', DCT.provenance),
+        ]:
+            statement = [s for s in g.objects(dataset_ref, item[1])][0]
+            assert self._triple(g, statement, RDFS.label, extras[item[0]])
+
     def test_identifier_extra(self):
         dataset = {
             'id': '4b6fe9ca-dc77-4cec-92a4-55c6624a5bd6',
@@ -670,7 +677,6 @@ def test_distribution_fields(self):
         assert self._triple(g, distribution, RDF.type, DCAT.Distribution)
         assert self._triple(g, distribution, DCT.title, resource['name'])
         assert self._triple(g, distribution, DCT.description, resource['description'])
-        assert self._triple(g, distribution, DCT.rights, resource['rights'])
         assert self._triple(g, distribution, DCT.license, URIRef(resource['license']))
         assert self._triple(g, distribution, ADMS.status, URIRef(resource['status']))
 
@@ -703,6 +709,10 @@ def test_distribution_fields(self):
         assert self._triple(g, checksum, SPDX.checksumValue, resource['hash'], data_type='http://www.w3.org/2001/XMLSchema#hexBinary')
         assert self._triple(g, checksum, SPDX.algorithm, URIRef(resource['hash_algorithm']))
 
+        # Statements
+        statement = [s for s in g.objects(distribution, DCT.rights)][0]
+        assert self._triple(g, statement, RDFS.label, resource['rights'])
+
     def test_distribution_size_not_number(self):
 
         resource = {

ckanext/dcat/tests/profiles/dcat_ap_2/test_scheming_support.py

@@ -25,6 +25,7 @@
     GSP,
     OWL,
     SPDX,
+    RDFS,
 )
 from ckanext.dcat.tests.utils import BaseSerializeTest, BaseParseTest
 
@@ -76,10 +77,8 @@ def test_e2e_ckan_to_dcat(self):
         assert self._triple(
             g, dataset_ref, DCT.accrualPeriodicity, dataset["frequency"]
         )
-        assert self._triple(g, dataset_ref, DCT.provenance, dataset["provenance"])
         assert self._triple(g, dataset_ref, DCT.type, dataset["dcat_type"])
         assert self._triple(g, dataset_ref, ADMS.versionNotes, dataset["version_notes"])
-        assert self._triple(g, dataset_ref, DCT.accessRights, dataset["access_rights"])
         assert self._triple(
             g,
             dataset_ref,
@@ -274,6 +273,14 @@ def test_e2e_ckan_to_dcat(self):
         distribution_ref = self._triple(g, dataset_ref, DCAT.distribution, None)[2]
         resource = dataset_dict["resources"][0]
 
+        # Statements
+        for item in [
+            ('access_rights', DCT.accessRights),
+            ('provenance', DCT.provenance),
+        ]:
+            statement = [s for s in g.objects(dataset_ref, item[1])][0]
+            assert self._triple(g, statement, RDFS.label, dataset[item[0]])
+
         # Resources: core fields
 
         assert self._triple(g, distribution_ref, DCT.title, resource["name"])
@@ -286,7 +293,6 @@ def test_e2e_ckan_to_dcat(self):
 
         # Resources: standard fields
 
-        assert self._triple(g, distribution_ref, DCT.rights, resource["rights"])
         assert self._triple(
             g, distribution_ref, ADMS.status, URIRef(resource["status"])
         )
@@ -380,6 +386,10 @@ def test_e2e_ckan_to_dcat(self):
         ]
         assert endpoint_urls == resource["access_services"][0]["endpoint_url"]
 
+        # Resources: statements
+        statement = [s for s in g.objects(distribution_ref, DCT.rights)][0]
+        assert self._triple(g, statement, RDFS.label, resource['rights'])
+
     def test_publisher_fallback_org(self):
 
         org = factories.Organization(

ckanext/dcat/tests/profiles/dcat_ap_3/test_euro_dcatap_3_profile_serialize.py

@@ -24,6 +24,7 @@
     GSP,
     OWL,
     SPDX,
+    RDFS,
 )
 
 DCAT_AP_PROFILES = ["euro_dcat_ap_3"]
@@ -74,10 +75,8 @@ def test_e2e_ckan_to_dcat(self):
         assert self._triple(
             g, dataset_ref, DCT.accrualPeriodicity, dataset["frequency"]
         )
-        assert self._triple(g, dataset_ref, DCT.provenance, dataset["provenance"])
         assert self._triple(g, dataset_ref, DCT.type, dataset["dcat_type"])
         assert self._triple(g, dataset_ref, ADMS.versionNotes, dataset["version_notes"])
-        assert self._triple(g, dataset_ref, DCT.accessRights, dataset["access_rights"])
         assert self._triple(
             g,
             dataset_ref,
@@ -230,6 +229,14 @@ def test_e2e_ckan_to_dcat(self):
         distribution_ref = self._triple(g, dataset_ref, DCAT.distribution, None)[2]
         resource = dataset_dict["resources"][0]
 
+        # Statements
+        for item in [
+            ('access_rights', DCT.accessRights),
+            ('provenance', DCT.provenance),
+        ]:
+            statement = [s for s in g.objects(dataset_ref, item[1])][0]
+            assert self._triple(g, statement, RDFS.label, dataset[item[0]])
+
         # Alternate identifiers
         ids = []
         for subject in [t[2] for t in g.triples((dataset_ref, ADMS.identifier, None))]:
@@ -248,7 +255,6 @@ def test_e2e_ckan_to_dcat(self):
 
         # Resources: standard fields
 
-        assert self._triple(g, distribution_ref, DCT.rights, resource["rights"])
         assert self._triple(
             g, distribution_ref, ADMS.status, URIRef(resource["status"])
         )
@@ -343,6 +349,10 @@ def test_e2e_ckan_to_dcat(self):
         ]
         assert endpoint_urls == resource["access_services"][0]["endpoint_url"]
 
+        # Resources: statements
+        statement = [s for s in g.objects(distribution_ref, DCT.rights)][0]
+        assert self._triple(g, statement, RDFS.label, resource['rights'])
+
     def test_byte_size_non_negative_integer(self):
 
         dataset = {