- Add CKAN -> DCAT support publisher identifier

ckanext/dcat/converters.py

@@ -27,8 +27,7 @@ def dcat_to_ckan(dcat_dict):
     if isinstance(dcat_publisher, basestring):
         package_dict['extras'].append({'key': 'dcat_publisher_name', 'value': dcat_publisher})
     elif isinstance(dcat_publisher, dict) and dcat_publisher.get('name'):
-        if dcat_publisher.get('name'):
-            package_dict['extras'].append({'key': 'dcat_publisher_name', 'value': dcat_publisher.get('name')})
+        package_dict['extras'].append({'key': 'dcat_publisher_name', 'value': dcat_publisher.get('name')})
 
         if dcat_publisher.get('mbox'):
             package_dict['extras'].append({'key': 'dcat_publisher_email', 'value': dcat_publisher.get('mbox')})

ckanext/dcat/profiles/euro_dcat_ap_base.py

@@ -372,6 +372,7 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
                 "email": self._get_dataset_value(dataset_dict, "publisher_email"),
                 "url": self._get_dataset_value(dataset_dict, "publisher_url"),
                 "type": self._get_dataset_value(dataset_dict, "publisher_type"),
+                "identifier": self._get_dataset_value(dataset_dict, "publisher_identifier"),
             }
         elif dataset_dict.get("organization"):
             # Fall back to dataset org
@@ -396,6 +397,7 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
                     "email": org_dict.get("email"),
                     "url": org_dict.get("url"),
                     "type": org_dict.get("dcat_type"),
+                    "identifier": org_dict.get("identifier"),
                 }
         # Add to graph
         if publisher_ref:
@@ -406,6 +408,7 @@ def _graph_from_dataset_base(self, dataset_dict, dataset_ref):
                 ("email", FOAF.mbox, None, Literal),
                 ("url", FOAF.homepage, None, URIRef),
                 ("type", DCT.type, None, URIRefOrLiteral),
+                ("identifier", DCT.identifier, None, URIRefOrLiteral),
             ]
             self._add_triples_from_dict(publisher_details, publisher_ref, items)
 

ckanext/dcat/profiles/euro_dcat_ap_scheming.py

@@ -187,6 +187,13 @@ def _not_empty_dict(data_dict):
                 _type=URIRef,
                 value_modifier=self._add_mailto,
             )
+            self._add_triple_from_dict(
+                publisher,
+                publisher_ref,
+                DCT.identifier,
+                "identifier",
+                _type=URIRefOrLiteral
+            )
 
         temporal = dataset_dict.get("temporal_coverage")
         if (

ckanext/dcat/schemas/dcat_ap_full.yaml

@@ -66,6 +66,10 @@ dataset_fields:
 
     - field_name: type
       label: Type
+
+    - field_name: identifier
+      label: Identifier
+      help_text: Unique identifier for the publisher, such as a ROR ID.
   help_text: Entity responsible for making the dataset available.
 
 - field_name: license_id

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

@@ -94,6 +94,7 @@ def test_e2e_ckan_to_dcat(self):
                     "email": "[email protected]",
                     "url": "https://example.org",
                     "type": "public_body",
+                    "identifier": "http://example.org/publisher-id",
                 },
             ],
             "temporal_coverage": [
@@ -301,6 +302,12 @@ def test_e2e_ckan_to_dcat(self):
             DCT.type,
             dataset_dict["publisher"][0]["type"],
         )
+        assert self._triple(
+            g,
+            publisher[0][2],
+            DCT.identifier,
+            dataset_dict["publisher"][0]["identifier"]
+        )
 
         temporal = [t for t in g.triples((dataset_ref, DCT.temporal, None))]