[Fixes #14381] Relax Remote Service URL uniqueness

geonode/base/api/tests.py

@@ -516,7 +516,7 @@ def test_base_resources(self):
         response = self.client.get(url, format="json")
         self.assertEqual(response.status_code, 200)
         self.assertEqual(len(response.data), 5)
-        self.assertEqual(response.data["total"], 28)
+        self.assertEqual(response.data["total"], 29)
 
         url = f"{reverse('base-resources-list')}?filter{{metadata_only}}=false"
         # Anonymous
@@ -984,7 +984,7 @@ def test_sort_resources(self):
         response = self.client.get(f"{url}?sort[]=title", format="json")
         self.assertEqual(response.status_code, 200)
         self.assertEqual(len(response.data), 5)
-        self.assertEqual(response.data["total"], 28)
+        self.assertEqual(response.data["total"], 29)
         # Pagination
         self.assertEqual(len(response.data["resources"]), 10)
 
@@ -997,7 +997,7 @@ def test_sort_resources(self):
         response = self.client.get(f"{url}?sort[]=-title", format="json")
         self.assertEqual(response.status_code, 200)
         self.assertEqual(len(response.data), 5)
-        self.assertEqual(response.data["total"], 28)
+        self.assertEqual(response.data["total"], 29)
         # Pagination
         self.assertEqual(len(response.data["resources"]), 10)
 

geonode/geoserver/signals.py

@@ -34,8 +34,8 @@
 from geonode.layers.models import Dataset
 from geonode.services.enumerations import CASCADED
 
-from . import BACKEND_PACKAGE
-from .tasks import geoserver_cascading_delete, geoserver_post_save_datasets
+from geonode.geoserver import BACKEND_PACKAGE
+from geonode.geoserver.tasks import geoserver_cascading_delete, geoserver_post_save_datasets
 
 logger = logging.getLogger("geonode.geoserver.signals")
 
@@ -104,14 +104,23 @@ def geoserver_pre_save_maplayer(instance, sender, **kwargs):
     # Set dataset
     if instance.dataset is None:
         dataset_queryset = Dataset.objects.filter(Q(alternate=instance.name) | Q(name=instance.name))
-        if instance.local and instance.store:
+        if instance.store:
             dataset_queryset = dataset_queryset.filter(store=instance.store)
         elif instance.ows_url:
             dataset_queryset = dataset_queryset.filter(remote_service__base_url=instance.ows_url)
         try:
             instance.dataset = dataset_queryset.get()
-        except (Dataset.DoesNotExist, Dataset.MultipleObjectsReturned):
+        except Dataset.DoesNotExist:
             pass
+        except Dataset.MultipleObjectsReturned:
+            # Expected since #14381: the same endpoint can be registered by several services, so a
+            # name can match one dataset per service. Nothing on the map layer tells them apart.
+            logger.warning(
+                "Ambiguous dataset lookup for map layer '%s' (store=%s, ows_url=%s), leaving it unresolved",
+                instance.name,
+                instance.store,
+                instance.ows_url,
+            )
 
 
 @deprecated(version="3.2.1", reason="Use direct calls to the ReourceManager.")

geonode/harvesting/harvesters/arcgis.py

@@ -34,7 +34,6 @@
 import arcrest
 import requests
 from django.contrib.gis import geos
-from django.template.defaultfilters import slugify
 
 from geonode.layers.enumerations import GXP_PTYPES
 from geonode.layers.models import Dataset
@@ -204,7 +203,6 @@ def _get_resource_descriptor(
         _, service_name, service_type = parse_remote_url(harvestable_resource.unique_identifier)
         epsg_code, spatial_extent = _parse_spatial_extent(layer_representation["extent"])
         ows_url = harvestable_resource.harvester.remote_url
-        store = slugify(ows_url)
         name = layer_representation.get("id", layer_representation.get("name", "Undefined"))
         title = layer_representation.get("name", layer_representation.get("title", "Undefined"))
         workspace = "remoteWorkspace"
@@ -228,7 +226,8 @@ def _get_resource_descriptor(
             ),
             reference_systems=[epsg_code],
             additional_parameters={
-                "store": store,
+                # store is a GeoServer grouping: remote datasets have none
+                "store": None,
                 "workspace": workspace,
                 "alternate": alternate,
                 "ows_url": ows_url,
@@ -306,7 +305,6 @@ def _get_resource_descriptor(
         _, service_name, service_type = parse_remote_url(harvestable_resource.unique_identifier)
         epsg_code, spatial_extent = _parse_spatial_extent(layer_representation["extent"])
         ows_url = harvestable_resource.harvester.remote_url
-        store = slugify(ows_url)
         name = layer_representation.get("id", layer_representation.get("name", "Undefined"))
         title = layer_representation.get("name", layer_representation.get("title", "Undefined"))
         workspace = "remoteWorkspace"
@@ -330,7 +328,8 @@ def _get_resource_descriptor(
             ),
             reference_systems=[epsg_code],
             additional_parameters={
-                "store": store,
+                # store is a GeoServer grouping: remote datasets have none
+                "store": None,
                 "workspace": workspace,
                 "alternate": alternate,
                 "ows_url": ows_url,

geonode/harvesting/harvesters/geonodeharvester.py

@@ -392,6 +392,8 @@ def _get_resource_descriptor(
                 {
                     "alternate": resource["alternate"],
                     "workspace": resource["workspace"],
+                    # store is a GeoServer grouping: remote datasets have none
+                    "store": None,
                 }
             )
         return descriptor
@@ -782,6 +784,8 @@ def _get_dataset_additional_parameters(
             "resource_type": "dataset",
             "alternate": api_record.get("alternate", descriptor.identification.name),
             "workspace": api_record.get("workspace"),
+            # store is a GeoServer grouping: remote datasets have none
+            "store": None,
         }
         if descriptor.identification.native_format.lower() == RemoteDatasetType.VECTOR.value:
             result["subtype"] = GeoNodeDatasetType.VECTOR.value

geonode/harvesting/harvesters/wms.py

@@ -32,7 +32,6 @@
 
 from django.conf import settings
 from django.contrib.gis import geos
-from django.template.defaultfilters import slugify
 from geonode.layers.models import Dataset
 from geonode.base.models import Link, ResourceBase
 from geonode.layers.enumerations import GXP_PTYPES
@@ -309,7 +308,6 @@ def get_resource(
             # WMS does not provide the date of the resource.
             # Use current time for the date stamp and resource time.
             time = datetime.now()
-            service_name = slugify(self.remote_url)[:255]
             contact = resourcedescriptor.RecordDescriptionContact(**data["contact"])
             result = base.HarvestedResourceInfo(
                 resource_descriptor=resourcedescriptor.RecordDescription(
@@ -335,7 +333,7 @@ def get_resource(
                     reference_systems=[relevant_layer["crs"]],
                     additional_parameters={
                         "alternate": relevant_layer["name"],
-                        "store": service_name,
+                        "store": None,
                         "workspace": "remoteWorkspace",
                         "ows_url": relevant_layer["wms_url"],
                         "ptype": GXP_PTYPES["WMS"],

geonode/harvesting/tests/test_harvester_worker_wms.py

@@ -16,6 +16,8 @@
 # along with this program. If not, see <http://www.gnu.org/licenses/>.
 #
 #########################################################################
+from unittest import mock
+
 from geonode.harvesting.harvesters import wms
 from geonode.tests.base import GeoNodeBaseSimpleTestSupport
 
@@ -29,3 +31,25 @@ def test_get_nsmap(self):
         for original, expected in fixtures:
             result = wms._get_nsmap(original)
             self.assertEqual(result, expected)
+
+    def test_harvested_resource_has_no_store(self):
+        # store is a GeoServer grouping, so harvested datasets must not claim one (#14381)
+        data = {
+            "contact": {"role": "", "name": "someone"},
+            "layers": [
+                {
+                    "name": "bahra",
+                    "title": "Bahra",
+                    "abstract": "",
+                    "keywords": [],
+                    "spatial_extent": None,
+                    "crs": "EPSG:4326",
+                    "wms_url": "https://wms.example.org/geoserver/wms?layers=bahra",
+                }
+            ],
+        }
+        worker = wms.OgcWmsHarvester("https://wms.example.org/geoserver/wms", 1)
+        harvestable_resource = mock.MagicMock(unique_identifier="bahra", geonode_resource=None)
+        with mock.patch.object(worker, "_get_data", return_value=data):
+            harvested = worker.get_resource(harvestable_resource)
+        self.assertIsNone(harvested.resource_descriptor.additional_parameters["store"])

geonode/layers/api/tests.py

@@ -141,8 +141,8 @@ def test_filter_dirty_state(self):
         self.assertEqual(response.status_code, 200)
         dataset_list = response.data["datasets"]
 
-        # dirty resource is now included in count
-        resource_count_clean = Dataset.objects.count()
+        # dirty resource is now included in count; the listing excludes metadata-only datasets
+        resource_count_clean = Dataset.objects.filter(metadata_only=False).count()
         self.assertEqual(len(dataset_list), resource_count_clean)
         # ensure that the updated dirty dataset is included in the response
         self.assertTrue(dirty_dataset.pk in [int(dataset["pk"]) for dataset in dataset_list])
@@ -157,10 +157,11 @@ def test_filter_dirty_state_include_dirty(self):
         dirty_dataset.dirty_state = True
         dirty_dataset.save()
 
+        # the listing excludes metadata-only datasets
         # clean resources
-        resource_count_clean = Dataset.objects.filter(dirty_state=False).count()
+        resource_count_clean = Dataset.objects.filter(dirty_state=False, metadata_only=False).count()
         # dirty resources
-        resource_count_dirty = Dataset.objects.filter(dirty_state=True).count()
+        resource_count_dirty = Dataset.objects.filter(dirty_state=True, metadata_only=False).count()
 
         resource_count_all = resource_count_clean + resource_count_dirty
 
@@ -187,7 +188,7 @@ def test_dataset_listing_advertised(self):
 
         prev_count = payload.json().get("total")
         # the user can see only the advertised resources
-        self.assertEqual(Dataset.objects.filter(advertised=True).count(), prev_count)
+        self.assertEqual(Dataset.objects.filter(advertised=True, metadata_only=False).count(), prev_count)
 
         payload = self.client.get(f"{url}?advertised=True")
         # so if advertised is True, we dont see the advertised=False resource

geonode/layers/migrations/0047_alter_dataset_store.py

@@ -0,0 +1,17 @@
+# Generated by Django 5.2.13 on 2026-08-14 10:00
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("layers", "0046_remove_modeltranslation"),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="dataset",
+            name="store",
+            field=models.CharField(blank=True, max_length=255, null=True, verbose_name="Store"),
+        ),
+    ]

geonode/layers/models.py

@@ -118,7 +118,8 @@ class Dataset(ResourceBase):
     # internal fields
     objects = DatasetManager()
     workspace = models.CharField(_("Workspace"), max_length=255)
-    store = models.CharField(_("Store"), max_length=255)
+    # Store is nullable for remote layers
+    store = models.CharField(_("Store"), max_length=255, null=True, blank=True)
     name = models.CharField(_("Name"), max_length=255)
     typename = models.CharField(_("Typename"), max_length=255, null=True, blank=True)
     ows_url = models.URLField(

geonode/layers/tests.py

@@ -284,6 +284,22 @@ def test_dataset_attribute_config(self):
         attribute_config = lyr.attribute_config()
         self.assertTrue("ftInfoTemplate" not in attribute_config)
 
+    def test_remote_datasets_are_exempt_from_store_workspace_name_uniqueness(self):
+        # store is NULL for remote datasets, and PostgreSQL does not consider two NULLs equal,
+        # so the (store, workspace, name) uniqueness no longer applies to them: the same remote
+        # endpoint can be registered by more than one service (#14381).
+        common = dict(
+            owner=get_user_model().objects.get(username="admin"),
+            workspace="remoteWorkspace",
+            name="bahra",
+            subtype="remote",
+            store=None,
+        )
+        first = Dataset.objects.create(uuid=str(uuid4()), title="bahra", alternate="bahra", **common)
+        second = Dataset.objects.create(uuid=str(uuid4()), title="bahra", alternate="bahra", **common)
+        self.assertNotEqual(first.pk, second.pk)
+        self.assertEqual(Dataset.objects.filter(workspace="remoteWorkspace", name="bahra").count(), 2)
+
     def test_dataset_styles(self):
         lyr = Dataset.objects.all().first()
         # There should be a total of 3 styles
@@ -296,7 +312,7 @@ def test_dataset_styles(self):
         except UnicodeEncodeError:
             self.fail("str of the Style model throws a UnicodeEncodeError with special characters.")
 
-    def test_dataset_links(self):
+    def test_vector_links(self):
         lyr = Dataset.objects.filter(subtype="vector").first()
         self.assertEqual(lyr.subtype, "vector")
 
@@ -324,6 +340,7 @@ def test_dataset_links(self):
             links = Link.objects.filter(resource=lyr.resourcebase_ptr, link_type="image")
             self.assertIsNotNone(links)
 
+    def test_raster_links(self):
         lyr = Dataset.objects.filter(subtype="raster").first()
         self.assertEqual(lyr.subtype, "raster")
         if check_ogc_backend(geoserver.BACKEND_PACKAGE):
@@ -350,6 +367,42 @@ def test_dataset_links(self):
             links = Link.objects.filter(resource=lyr.resourcebase_ptr, link_type="image")
             self.assertIsNotNone(links)
 
+    def test_set_resource_default_links_passes_auth_config_to_service_handler(self):
+        remote_auth_config = object()
+        instance = MagicMock()
+        instance.resourcebase_ptr = MagicMock()
+        instance.srid = "EPSG:4326"
+        instance.bbox_polygon = True
+        instance.bbox_string = "0,0,1,1"
+        instance.ows_url = "http://example.com/ows"
+        instance.alternate = "remote:layer"
+        instance.can_have_wfs_links = False
+        instance.subtype = "vector"
+        instance.can_have_style = False
+        instance.prepare_wms_links.return_value = []
+        instance.get_thumbnail_url.return_value = None
+        instance.get_real_instance.return_value = MagicMock(ptype="NOT_WMS")
+        instance.remote_service = MagicMock(
+            service_url="http://example.com/ows?service=WMS",
+            type="WMS",
+            id=42,
+            auth_config=remote_auth_config,
+        )
+
+        with (
+            patch("geonode.utils.check_ogc_backend", return_value=True),
+            patch("geonode.resource.utils.is_remote_resource", return_value=False),
+            patch("geonode.base.models.Link.objects") as mock_link_objects,
+            patch("geonode.services.serviceprocessors.get_service_handler") as mock_get_service_handler,
+        ):
+            mock_link_objects.filter.return_value.count.return_value = 0
+            mock_get_service_handler.return_value = MagicMock(_create_dataset_legend_link=MagicMock())
+
+            set_resource_default_links(instance, instance)
+
+            _, kwargs = mock_get_service_handler.call_args
+            self.assertIs(kwargs.get("auth_config"), remote_auth_config)
+
     def test_get_valid_user(self):
         # Verify it accepts an admin user
         adminuser = get_user_model().objects.get(is_superuser=True)

geonode/security/tests.py

@@ -1438,7 +1438,8 @@ def test_get_visible_resources_should_return_resource_with_metadata_only_true(se
             actual = get_visible_resources(
                 queryset=layers, metadata_only=True, user=get_user_model().objects.get(username=self.user)
             )
-            self.assertEqual(1, actual.count())
+            # plus "dataset metadata true" from the fixtures
+            self.assertEqual(2, actual.count())
         finally:
             if dataset:
                 dataset.delete()
@@ -1470,7 +1471,8 @@ def test_get_visible_resources_advanced_workflow(self):
         self.assertIsNotNone(standard_user)
         admin_user.is_superuser = True
         admin_user.save()
-        layers = Dataset.objects.all()
+        # get_visible_resources only returns metadata_only=False resources
+        layers = Dataset.objects.filter(metadata_only=False)
 
         actual = get_visible_resources(
             queryset=Dataset.objects.all(),

geonode/services/forms.py

@@ -22,15 +22,15 @@
 from django import forms
 from django.core.exceptions import ValidationError
 from django.utils.translation import gettext_lazy as _
-import taggit
+from taggit.forms import TagField
 
 from geonode.security.auth_handlers import BasicAuthHandler
 from geonode.security.auth_registry import auth_handler_registry
 from geonode.security.models import AuthConfig
 
-from . import enumerations
-from .models import Service, get_service_type_choices
-from .serviceprocessors import get_service_handler
+from geonode.services import enumerations
+from geonode.services.models import Service, get_service_type_choices
+from geonode.services.serviceprocessors import get_service_handler
 from geonode.utils import is_safe_url
 
 logger = logging.getLogger(__name__)
@@ -72,10 +72,6 @@ def clean_url(self):
 
         if not is_safe_url(proposed_url):
             raise ValidationError(_("Invalid URL provided"))
-
-        existing = Service.objects.filter(base_url=proposed_url).exists()
-        if existing:
-            raise ValidationError(_("Service %(url)s is already registered"), params={"url": proposed_url})
         return proposed_url
 
     def clean(self):
@@ -129,7 +125,7 @@ class ServiceForm(forms.ModelForm):
     )
     description = forms.CharField(label=_("Description"), widget=forms.Textarea(attrs={"cols": 60}))
     abstract = forms.CharField(label=_("Abstract"), widget=forms.Textarea(attrs={"cols": 60}))
-    keywords = taggit.forms.TagField(required=False)
+    keywords = TagField(required=False)
 
     class Meta:
         model = Service