Support dct:license on DCAT harvester

CHANGELOG.md

@@ -2,7 +2,7 @@
 
 ## Current (in progress)
 
-- Nothing yet
+- Support dct:license on DCAT harvester [#2589](https://github.com/opendatateam/udata/pull/2589)
 
 ## 2.6.1 (2021-01-26)
 

udata/core/dataset/rdf.py

@@ -418,6 +418,7 @@ def dataset_from_rdf(graph, dataset=None, node=None):
                 licenses.add(value.identifier.toPython())
 
     default_license = dataset.license or License.default()
-    dataset.license = License.guess(*licenses, default=default_license)
+    dataset_license = rdf_value(d, DCT.license)
+    dataset.license = License.guess(dataset_license, *licenses, default=default_license)
 
     return dataset

udata/harvest/tests/dcat/catalog.xml

@@ -4,6 +4,7 @@
   xmlns:owl="http://www.w3.org/2002/07/owl#"
   xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
   xmlns:dcat="http://www.w3.org/ns/dcat#"
+  xmlns:dct="http://purl.org/dc/terms/"
   xmlns:dcterms="http://purl.org/dc/terms/"
   xmlns:vcard="http://www.w3.org/2006/vcard/ns#"
   xmlns:schema="http://schema.org/"
@@ -22,6 +23,7 @@
         <dcterms:description>Dataset 3 description</dcterms:description>
         <dcat:keyword>Tag 1</dcat:keyword>
         <dcat:distribution rdf:resource="http://data.test.org/datasets/3/resources/1"/>
+        <dct:license>Licence Ouverte Version 2.0</dct:license>
       </dcat:Dataset>
     </dcat:dataset>
     <dcterms:title>Sample DCAT Catalog</dcterms:title>

udata/harvest/tests/test_dcat_backend.py

@@ -7,6 +7,7 @@
 
 from udata.models import Dataset, License
 from udata.core.organization.factories import OrganizationFactory
+from udata.core.dataset.factories import LicenseFactory
 
 from .factories import HarvestSourceFactory
 from .. import actions
@@ -44,11 +45,6 @@ def callback(request, context):
 @pytest.mark.usefixtures('clean_db')
 @pytest.mark.options(PLUGINS=['dcat'])
 class DcatBackendTest:
-    @pytest.fixture(autouse=True)
-    def inject_licenses(self):
-        # Create fake licenses
-        for license_id in 'lool', 'fr-lo':
-            License.objects.create(id=license_id, title=license_id)
 
     def test_simple_flat(self, rmock):
         filename = 'flat.jsonld'
@@ -218,6 +214,23 @@ def test_supported_mime_type(self, rmock):
         assert job.errors == []
         assert len(job.items) == 3
 
+    def test_xml_catalog(self, rmock):
+        LicenseFactory(id='lov2', title='Licence Ouverte Version 2.0')
+
+        url = mock_dcat(rmock, 'catalog.xml', path='catalog.xml')
+        org = OrganizationFactory()
+        source = HarvestSourceFactory(backend='dcat',
+                                      url=url,
+                                      organization=org)
+
+        actions.run(source.slug)
+
+        # test dct:license support
+        extras = {'extras__dct:identifier': '3'}
+        dataset = Dataset.objects.get(**extras)
+        assert dataset.license.id == 'lov2'
+
+
     def test_unsupported_mime_type(self, rmock):
         url = DCAT_URL_PATTERN.format(path='', domain=TEST_DOMAIN)
         rmock.head(url, headers={'Content-Type': 'text/html; charset=utf-8'})