fix: add logic for sameAs

resolves #401

apis_ontology/importers.py

@@ -1,6 +1,9 @@
 from django.apps import apps
+from django.core.exceptions import ImproperlyConfigured
+from django.db.utils import IntegrityError
 from apis_core.generic.importers import GenericModelImporter
 from apis_core.utils.helpers import create_object_from_uri
+from apis_core.apis_metainfo.models import Uri
 
 
 class BaseEntityImporter(GenericModelImporter):
@@ -9,19 +12,38 @@ class BaseEntityImporter(GenericModelImporter):
 
     def create_instance(self):
         data = self.get_data(drop_unknown_fields=False)
+        if "sameas" in data:
+            data["sameas"] = data["sameas"].split("|")
+            sa = Uri.objects.filter(uri__in=data["sameas"])
+            if sa.count() == 1:
+                return sa.first().root_object
+            elif sa.count() > 1:
+                raise IntegrityError(
+                    f"Multiple objects found for sameAs URIs {data['sames']}. "
+                    f"This indicates a data integrity problem as these URIs should be unique."
+                )
         modelfields = [field.name for field in self.model._meta.fields]
         data_croped = {key: data[key] for key in data if key in modelfields}
         subj = self.model.objects.create(**data_croped)
+        if "sameas" in data:
+            for uri in data["sameas"]:
+                Uri.objects.create(uri=uri, root_object_id=subj.id)
         related_keys = [
             (x, x.split("__")[1], x.split("__")[2]) for x in data.keys() if "__" in x
         ]
-        for rk in related_keys:
-            key, obj, rel = rk
-            RelatedModel = apps.get_model("apis_ontology", obj)
-            RelationType = apps.get_model("apis_ontology", rel)
-            if key in data:
-                related_obj = create_object_from_uri(data[key], RelatedModel)
-                RelationType.objects.create(subj=subj, obj=related_obj)
+        try:
+            for rk in related_keys:
+                key, obj, rel = rk
+                RelatedModel = apps.get_model("apis_ontology", obj)
+                RelationType = apps.get_model("apis_ontology", rel)
+                if key in data:
+                    related_obj = create_object_from_uri(data[key], RelatedModel)
+                    RelationType.objects.create(subj=subj, obj=related_obj)
+        except:  # noqa: E722
+            subj.delete()
+            raise ImproperlyConfigured(
+                f"Error in creating related Objects for {self.model.__class__.__name__}"
+            )
 
         return subj
 
@@ -39,3 +61,7 @@ def mangle_data(self, data):
 
 class InstitutionImporter(BaseEntityImporter):
     pass
+
+
+class PlaceImporter(BaseEntityImporter):
+    pass

apis_ontology/rdfimport/InstitutionFromDNB.toml

@@ -30,3 +30,13 @@ WHERE {
   ?subject gndo:dateOfTermination ?end_date_written
 }
 """
+[[attributes]]
+# sameAs
+sparql = '''
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+SELECT (GROUP_CONCAT(?sameas_pre; separator='|') as ?sameas)
+WHERE {
+  ?subject owl:sameAs ?sameas_pre
+}
+GROUP BY ?subject
+'''

apis_ontology/rdfimport/PersonFromDNB.toml

@@ -90,3 +90,13 @@ WHERE {
   ?subject gndo:placeOfDeath ?place_of_death__Place__StarbIn
 }
 """
+[[attributes]]
+# sameAs
+sparql = '''
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+SELECT (GROUP_CONCAT(?sameas_pre; separator='|') as ?sameas)
+WHERE {
+  ?subject owl:sameAs ?sameas_pre
+}
+GROUP BY ?subject
+'''

apis_ontology/rdfimport/PlaceFromDNB.toml

@@ -0,0 +1,53 @@
+#####################################################
+# Create an E53_Place from a d-nb.info RDF endpoint #
+# the second regex is for testing
+regex = "https://d-nb.info.*|/.*wien.rdf"
+superclass = "apis_ontology.models.Place"
+sameas = """
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+SELECT ?sameas
+WHERE {
+?subject owl:sameAs ?sameas
+}
+"""
+[[attributes]]
+# label
+sparql = """
+PREFIX gndo: <https://d-nb.info/standards/elementset/gnd#>
+SELECT ?label
+WHERE {
+  ?subject gndo:preferredNameForThePlaceOrGeographicName ?label
+}
+"""
+[[attributes]]
+# longitude
+sparql = '''
+PREFIX geo: <http://www.opengis.net/ont/geosparql#>
+SELECT ?longitude
+WHERE {
+  ?subject geo:hasGeometry ?geo1 .
+  ?geo1 geo:asWKT ?point .
+  BIND(REPLACE(str(?point), "Point \\( \\+?(-?\\d+.\\d+).*", "$1") as ?longitude)
+  }
+'''
+[[attributes]]
+# latitude
+sparql = '''
+PREFIX geo: <http://www.opengis.net/ont/geosparql#>
+SELECT ?latitude
+WHERE {
+  ?subject geo:hasGeometry ?geo1 .
+  ?geo1 geo:asWKT ?point .
+  BIND(REPLACE(str(?point), "^Point\\s*\\(\\s*[+-]?\\d+\\.\\d+\\s+([+-]?\\d+\\.\\d+)\\s*\\)$", "$1") as ?latitude)
+  }
+'''
+[[attributes]]
+# sameAs
+sparql = '''
+PREFIX owl: <http://www.w3.org/2002/07/owl#>
+SELECT (GROUP_CONCAT(?sameas_pre; separator='|') as ?sameas)
+WHERE {
+  ?subject owl:sameAs ?sameas_pre
+}
+GROUP BY ?subject
+'''