Merge branch 'handle-duplicate-classes' into 'master'

Prevent IntegrityError on duplicate ML classes on CreateClassifications

Closes #462

See merge request !1005

arkindex/documents/serializers/ml.py

@@ -477,6 +477,15 @@ class ClassificationsSerializer(serializers.Serializer):
                 'worker_version': ['This field XOR classifier field must be set to create classifications']
             })
 
+        ml_class_names = [
+            classification['ml_class']
+            for classification in data['classifications']
+        ]
+        if len(ml_class_names) != len(set(ml_class_names)):
+            raise ValidationError({
+                'classifications': ['Duplicated ML classes are not allowed from the same source or worker version.']
+            })
+
         return data
 
     def create(self, validated_data):

arkindex/documents/tests/test_bulk_classification.py

@@ -248,3 +248,22 @@ class TestBulkClassification(FixtureAPITestCase):
                 ('catte', 0.85, True),
             ],
         )
+
+    def test_bulk_classification_no_duplicates(self):
+        """
+        Test the bulk classification API prevents creating classifications with duplicate ML classes
+        """
+        self.client.force_login(self.user)
+        with self.assertNumQueries(4):
+            response = self.client.post(
+                reverse('api:classification-bulk'),
+                format='json',
+                data=self.create_classifications_data([
+                    {"class_name": 'dog', "confidence": 0.99},
+                    {"class_name": 'dog', "confidence": 0.99},
+                ])
+            )
+        self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST)
+        self.assertDictEqual(response.json(), {
+            'classifications': ['Duplicated ML classes are not allowed from the same source or worker version.']
+        })