arkindex/training/migrations/0008_dataset_unique_elements.py

+# Generated by Django 4.1.7 on 2024-03-28 14:54
+
+from django.db import migrations, models
+
+
+def update_unique_elements(apps, schema_editor):
+    """Update unique_elements to False when some elements are already duplicated"""
+    Dataset = apps.get_model("training", "Dataset")
+    DatasetElement = apps.get_model("training", "DatasetElement")
+    Dataset.objects.filter(
+        models.Exists(
+            DatasetElement.objects
+            .filter(set__dataset_id=models.OuterRef("pk"))
+            .values("element")
+            .annotate(dups=models.Count("element"))
+            .filter(dups__gte=2)
+        )
+    ).update(unique_elements=False)
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("training", "0007_datasetset_model"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="dataset",
+            name="unique_elements",
+            field=models.BooleanField(default=True),
+        ),
+        migrations.RunPython(
+            update_unique_elements,
+            reverse_code=migrations.RunPython.noop,
+            elidable=True,
+        ),
+    ]

arkindex/training/models.py

@@ -266,6 +266,7 @@ class Dataset(models.Model):
     name = models.CharField(max_length=100, validators=[MinLengthValidator(1)])
     description = models.TextField(validators=[MinLengthValidator(1)])
     state = EnumField(DatasetState, default=DatasetState.Open, max_length=50)
+    unique_elements = models.BooleanField(default=True)
 
     class Meta:
         constraints = [

arkindex/training/serializers.py

@@ -4,7 +4,7 @@ from collections import defaultdict
 from textwrap import dedent
 
 from django.db import transaction
-from django.db.models import Count, Q
+from django.db.models import Count, Exists, OuterRef, Q
 from drf_spectacular.utils import extend_schema_field
 from rest_framework import permissions, serializers
 from rest_framework.exceptions import PermissionDenied, ValidationError
@@ -575,6 +575,17 @@ class DatasetSerializer(serializers.ModelSerializer):
             raise ValidationError("Either do not specify set names to use the default values, or specify a non-empty list of names.")
         return set_names
 
+    def validate_unique_elements(self, unique):
+        if unique is True and self.instance and Exists(
+            DatasetElement.objects
+            .filter(set__dataset_id=OuterRef(self.instance.pk))
+            .values("element")
+            .annotate(dups=Count("element"))
+            .filter(dups__gte=2)
+        ):
+            raise ValidationError("Elements are currently contained by multiple sets.")
+        return unique
+
     def validate(self, data):
         data = super().validate(data)
 
@@ -636,6 +647,7 @@ class DatasetSerializer(serializers.ModelSerializer):
             # Hidden field to set the creator as the authenticated user
             "default_creator",
             "task_id",
+            "unique_elements",
             "created",
             "updated",
         )
@@ -702,7 +714,18 @@ class DatasetElementSerializer(serializers.ModelSerializer):
 
     def validate(self, data):
         data = super().validate(data)
-        data.pop("dataset")
+        dataset = data.pop("dataset")
+        if dataset.unique_elements and (
+            set := (
+                dataset.sets
+                .filter(set_elements__element_id=data["element_id"])
+                .values_list("name", flat=True)
+                .first()
+            )
+        ):
+            raise ValidationError({"element_id": [
+                f"The dataset prevent duplication and this element is already present in set {set}."
+            ]})
         return data