From a8a3f255ec18d822712e0b0d900e90e7d5ed6949 Mon Sep 17 00:00:00 2001
From: amercader <amercadero@gmail.com>
Date: Wed, 19 Jun 2024 12:23:16 +0200
Subject: [PATCH] [#56] Don't serialize empty repeating subfields

Scheming adds a dict with empty keys when empty
repeating subfields are submitted from the form. Check that there's an
actual value before creating the triples when serializing
---
 .../dcat/profiles/euro_dcat_ap_scheming.py    | 11 ++++++----
 ckanext/dcat/tests/test_scheming_support.py   | 22 +++++++++++++++++++
 2 files changed, 29 insertions(+), 4 deletions(-)

diff --git a/ckanext/dcat/profiles/euro_dcat_ap_scheming.py b/ckanext/dcat/profiles/euro_dcat_ap_scheming.py
index c945deff..12eb540e 100644
--- a/ckanext/dcat/profiles/euro_dcat_ap_scheming.py
+++ b/ckanext/dcat/profiles/euro_dcat_ap_scheming.py
@@ -120,8 +120,11 @@ def graph_from_dataset(self, dataset_dict, dataset_ref):
         Add triples to the graph from new repeating subfields
         """
 
+        def _not_empty_dict(data_dict):
+            return any(data_dict.values())
+
         contact = dataset_dict.get("contact")
-        if isinstance(contact, list) and len(contact):
+        if isinstance(contact, list) and len(contact) and _not_empty_dict(contact[0]):
             for item in contact:
                 contact_uri = item.get("uri")
                 if contact_uri:
@@ -144,7 +147,7 @@ def graph_from_dataset(self, dataset_dict, dataset_ref):
                 )
 
         publisher = dataset_dict.get("publisher")
-        if isinstance(publisher, list) and len(publisher):
+        if isinstance(publisher, list) and len(publisher) and _not_empty_dict(publisher[0]):
             publisher = publisher[0]
             publisher_uri = publisher.get("uri")
             if publisher_uri:
@@ -172,7 +175,7 @@ def graph_from_dataset(self, dataset_dict, dataset_ref):
             )
 
         temporal = dataset_dict.get("temporal_coverage")
-        if isinstance(temporal, list) and len(temporal):
+        if isinstance(temporal, list) and len(temporal) and _not_empty_dict(temporal[0]):
             for item in temporal:
                 temporal_ref = BNode()
                 self.g.add((temporal_ref, RDF.type, DCT.PeriodOfTime))
@@ -183,7 +186,7 @@ def graph_from_dataset(self, dataset_dict, dataset_ref):
                 self.g.add((dataset_ref, DCT.temporal, temporal_ref))
 
         spatial = dataset_dict.get("spatial_coverage")
-        if isinstance(spatial, list) and len(spatial):
+        if isinstance(spatial, list) and len(spatial) and _not_empty_dict(spatial[0]):
             for item in spatial:
                 if item.get("uri"):
                     spatial_ref = CleanedURIRef(item["uri"])
diff --git a/ckanext/dcat/tests/test_scheming_support.py b/ckanext/dcat/tests/test_scheming_support.py
index d8770e2f..3779780d 100644
--- a/ckanext/dcat/tests/test_scheming_support.py
+++ b/ckanext/dcat/tests/test_scheming_support.py
@@ -514,6 +514,28 @@ def test_publisher_fallback_org_ignored_if_publisher_field_present(self):
             g, publisher[0][2], FOAF.name, dataset_dict["publisher"][0]["name"]
         )
 
+    def test_empty_repeating_subfields_not_serialized(self):
+
+        dataset_dict = {
+            "name": "test-dataset-3",
+            "title": "Test DCAT dataset 3",
+            "notes": "Lorem ipsum",
+            "spatial_coverage": [
+                {
+                    "uri": "",
+                    "geom": "",
+                },
+            ],
+        }
+
+        dataset = call_action("package_create", **dataset_dict)
+
+        s = RDFSerializer()
+        g = s.g
+
+        dataset_ref = s.graph_from_dataset(dataset)
+        assert not [t for t in g.triples((dataset_ref, DCT.spatial, None))]
+
     def test_legacy_fields(self):
 
         dataset_dict = {