Add rights statement handling and multilingual support in DCAT profile

Catalog: - Rights statement Dataset: - Fix temporal resolution, only using ISO 8601 Distribution: - Add always accessUrl DataService: - Add dcat:theme
mjanez · Jan 23, 2025 · daf6adb · daf6adb
1 parent 94d5d65
commit daf6adb
Show file tree

Hide file tree

Showing 4 changed files with 69 additions and 20 deletions.
diff --git a/ckanext/schemingdcat/profiles/base.py b/ckanext/schemingdcat/profiles/base.py
@@ -45,6 +45,7 @@
     SPDX,
     CNT,
     ORG,
+    ODRS,
     # Default values
     eu_dcat_ap_default_values,
     )
@@ -80,6 +81,7 @@
     "owl": OWL,
     "org": ORG,
     "spdx": SPDX,
+    "odrs": ODRS,
 }
 
 default_lang = config.get("ckan.locale_default", "en")
@@ -748,6 +750,26 @@ def _clean_publisher(self, dataset_ref):
             self.g.remove((dataset_ref, DCT.publisher, publisher))
             self.g.remove((publisher, None, None))
 
+    def _is_valid_temporal_resolution(self, value: str) -> bool:
+        """
+        Validate ISO-8601 duration format.
+        
+        Format: P[nY][nM][nD][T[nH][nM][nS]]
+        Examples:
+            PT1H          - 1 hour
+            P1Y           - 1 year
+            P3M           - 3 months
+            P1DT12H      - 1 day, 12 hours
+            PT30M        - 30 minutes
+        """
+        if not value or not isinstance(value, str):
+            return False
+
+        # Simplified pattern that allows single units
+        pattern = r'^P(?:\d+Y)?(?:\d+M)?(?:\d+D)?(?:T(?:\d+H)?(?:\d+M)?(?:\d+S)?)?$|^PT(?:\d+H)?(?:\d+M)?(?:\d+S)?$'
+
+        return bool(re.match(pattern, value))
+
     # Graph enhancements. Fix Graph literals
     def _process_batch(self, graph: Graph, updates: List[Tuple]) -> None:
         """

diff --git a/ckanext/schemingdcat/profiles/dcat_ap/eu_dcat_ap_2.py b/ckanext/schemingdcat/profiles/dcat_ap/eu_dcat_ap_2.py
@@ -237,15 +237,6 @@ def _graph_from_dataset_v2(self, dataset_dict, dataset_ref):
         # Catalog URI
         catalog_ref = catalog_uri()
 
-        # Standard values
-        self._add_triple_from_dict(
-            dataset_dict,
-            dataset_ref,
-            DCAT.temporalResolution,
-            "temporal_resolution",
-            _datatype=XSD.duration,
-        )
-
         # Lists
         for key, predicate, fallbacks, type, datatype, _class in (
             (
@@ -364,13 +355,14 @@ def _graph_from_dataset_v2(self, dataset_dict, dataset_ref):
                 self.g.add((distribution_ref, DCATAP.availability, URIRef(distribution_availability)))
 
             # Temporal resolution
-            self._add_triple_from_dict(
-                resource_dict,
-                distribution_ref,
-                DCAT.temporalResolution,
-                "temporal_resolution",
-                _datatype=XSD.duration,
-            )
+            temporal_resolution = resource_dict.get("temporal_resolution")
+            log.debug('temporal_resolution:%s', temporal_resolution)
+            if temporal_resolution and self._is_valid_temporal_resolution(temporal_resolution):
+                self.g.add((
+                    distribution_ref, 
+                    DCAT.temporalResolution, 
+                    Literal(temporal_resolution, datatype=XSD.duration)
+                ))
 
             # Spatial resolution in meters
             spatial_resolution_in_meters = self._read_list_value(
@@ -484,7 +476,7 @@ def _graph_from_dataset_v2(self, dataset_dict, dataset_ref):
                 data_service_hvd_properties = [
                     ('hvd_category', DCATAP.hvdCategory, lambda x: x),
                     (None, DCT.license, self.g.value, dataset_ref),
-                    (None, DCT.accessRights, self.g.value, dataset_ref)
+                    (None, DCT.accessRights, self.g.value, dataset_ref),
                 ]
 
                 # Process mappings
@@ -501,6 +493,10 @@ def _graph_from_dataset_v2(self, dataset_dict, dataset_ref):
                             URIRef(value)
                         ))
 
+                # Add all DCAT.theme from dataset_ref to access_service_node
+                for theme in self.g.objects(dataset_ref, DCAT.theme):
+                    self.g.add((access_service_node, DCAT.theme, theme))
+
                 # Add DCAT.contactPoint from dataset_ref to access_service_node
                 contact_point = self.g.value(dataset_ref, DCAT.contactPoint)
                 if contact_point:

diff --git a/ckanext/schemingdcat/profiles/dcat_config.py b/ckanext/schemingdcat/profiles/dcat_config.py
@@ -49,6 +49,7 @@
 OWL = Namespace("http://www.w3.org/2002/07/owl#")
 SPDX = Namespace("http://spdx.org/rdf/terms#")
 CNT = Namespace("http://www.w3.org/2011/content#")
+ODRS = Namespace("http://schema.theodi.org/odrs#")
 
 CODELISTS_DIR = Path(__file__).resolve().parent.parent / "codelists"
 EU_VOCABS_DIR = CODELISTS_DIR / "dcat"
@@ -125,6 +126,8 @@
     'theme_es': 'http://datos.gob.es/kos/sector-publico/sector/sector-publico',
     'theme_eu': 'http://publications.europa.eu/resource/authority/data-theme/GOVE',
     'theme_taxonomy': 'http://datos.gob.es/kos/sector-publico/sector/',
+    'rights_uri_label': 'Derechos relativos a la reutilización del Catálogo de Datos Abiertos',
+    'rights_attribution_text': 'Atribución de la autoría al organismo.',
     'spatial_uri': 'http://datos.gob.es/recurso/sector-publico/territorio/Pais/España',
 }
 
@@ -149,6 +152,8 @@
     'theme_es': 'http://datos.gob.es/kos/sector-publico/sector/sector-publico',
     'theme_eu': 'http://publications.europa.eu/resource/authority/data-theme/GOVE',
     'theme_taxonomy': 'http://datos.gob.es/kos/sector-publico/sector/',
+    'rights_uri_label': 'Derechos relativos a la reutilización del Catálogo de Datos Abiertos',
+    'rights_attribution_text': 'Atribución de la autoría al organismo.',
     'spatial_uri': 'http://datos.gob.es/recurso/sector-publico/territorio/Pais/España',
 }
 
@@ -179,6 +184,8 @@
     'theme_taxonomy': 'http://inspire.ec.europa.eu/theme',
     'theme_es_taxonomy': 'http://datos.gob.es/kos/sector-publico/sector',
     'theme_eu_taxonomy': 'http://publications.europa.eu/resource/authority/data-theme',
+    'rights_uri_label': 'Rights related to the re-use of the Open Data Catalogue',
+    'rights_attribution_text': 'Attribution of authorship to the organisation.',
     'spatial_uri': 'http://publications.europa.eu/resource/authority/country/ESP',
 }
 
@@ -207,6 +214,8 @@
     'theme_taxonomy': 'http://inspire.ec.europa.eu/theme',
     'theme_es_taxonomy': 'http://datos.gob.es/kos/sector-publico/sector',
     'theme_eu_taxonomy': 'http://publications.europa.eu/resource/authority/data-theme',
+    'rights_uri_label': 'Rights related to the re-use of the Open Data Catalogue',
+    'rights_attribution_text': 'Attribution of authorship to the organisation.',
     'spatial_uri': 'http://publications.europa.eu/resource/authority/country/ESP',
 }
 

diff --git a/ckanext/schemingdcat/profiles/eu_dcat_ap_base.py b/ckanext/schemingdcat/profiles/eu_dcat_ap_base.py
@@ -41,6 +41,8 @@
     GEODCATAP,
     ELI,
     DCT,
+    DC,
+    ODRS,
     ADMS,
     VCARD,
     FOAF,
@@ -56,6 +58,7 @@
     metadata_field_names,
     default_translated_fields,
     eu_dcat_ap_default_values,
+    es_dcat_ap_default_values,
     dcat_ap_default_licenses,
     # URIS
     IANA_MEDIA_TYPES_BASE_URI,
@@ -829,7 +832,7 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
                     g.add((distribution, DCAT.mediaType, URIRef(mimetype_from_fmt)))      
 
             # URL fallback and old behavior
-            url = resource_dict.get("url")
+            url = resource_dict.get("url") or distribution
             download_url = resource_dict.get("download_url")
             access_url = resource_dict.get("access_url")
 
@@ -933,7 +936,6 @@ def _graph_from_catalog_base(self, catalog_dict, catalog_ref):
 
         # Mandatory elements by NTI-RISP/DCAT-AP-ES (datos.gob.es)
         items = [
-            ("identifier", DCT.identifier, catalog_uri(), URIRef),
             ("encoding", CNT.characterEncoding, "UTF-8", Literal),
             ("language", DCT.language, language, URIRefOrLiteral),
             ("spatial_uri", DCT.spatial, spatial_uri, URIRefOrLiteral),
@@ -943,7 +945,9 @@ def _graph_from_catalog_base(self, catalog_dict, catalog_ref):
             ("homepage", FOAF.homepage, config.get("ckan.site_url"), URIRef),
             ("license", DCT.license, license, URIRef),
             ("conforms_to", DCT.conformsTo, eu_dcat_ap_default_values["conformance"], URIRef),
-            ("access_rights", DCT.accessRights, access_rights, URIRefOrLiteral),
+            ("access_rights", DC.rights, f'{catalog_uri()}/rights', URIRefOrLiteral),
+            # Unnecesary properties for dcat:Catalog
+            #("identifier", DCT.identifier, catalog_uri(), URIRef),
             #("accessUrl", DCAT.accessURL, f'{catalog_uri()}/catalog.rdf', URIRef),
         ]
 
@@ -1008,6 +1012,24 @@ def _graph_from_catalog_base(self, catalog_dict, catalog_ref):
 
             self._add_triples_from_dict(publisher_details, publisher_ref, items)
 
+            # Create rights statement node
+            rights_uri = URIRef(f'{catalog_uri()}/rights')
+            g.add((catalog_ref, DC.rights, rights_uri))
+
+            # Add rights statement properties
+            g.add((rights_uri, RDF.type, ODRS.RightsStatement))
+
+            # Add multilingual labels
+            g.add((rights_uri, RDFS.label, Literal(eu_dcat_ap_default_values['rights_uri_label'], lang='en')))
+            g.add((rights_uri, RDFS.label, Literal(es_dcat_ap_default_values['rights_uri_label'], lang='es')))
+
+            # Add multilingual attribution text
+            g.add((rights_uri, ODRS.attributionText, Literal(eu_dcat_ap_default_values['rights_attribution_text'], lang='en')))
+            g.add((rights_uri, ODRS.attributionText, Literal(es_dcat_ap_default_values['rights_attribution_text'], lang='es')))
+
+            # Add non-language specific properties
+            g.add((rights_uri, ODRS.dataLicense, URIRef(eu_dcat_ap_default_values["license_url"])))
+            g.add((rights_uri, ODRS.attributionURL, URIRef(publisher_ref)))
 
     def _assign_theme_tags(self, dataset_dict, key, values):
         for value in values: