SCKAN-274 feat: Add constraint unique together to synonym

backend/composer/management/commands/ingest_anatomical_entities.py

@@ -12,15 +12,15 @@
 
 
 class Command(BaseCommand):
-    help = "Ingests Anatomical Entities CSV file(s), with robust error handling."
+    help = "Ingests Anatomical Entities CSV file(s)."
 
     def add_arguments(self, parser):
         parser.add_argument("csv_files", nargs="+", type=str)
         parser.add_argument("--show_complete_logs", action='store_true',
                             help="Show detailed logs during processing")
 
     def _process_anatomical_entity(self, name, ontology_uri, synonym, show_complete_logs, processed_uris,
-                                   synonym_accumulator):
+                                   unique_synonyms):
         try:
             is_first_occurrence = ontology_uri not in processed_uris
 
@@ -34,26 +34,29 @@ def _process_anatomical_entity(self, name, ontology_uri, synonym, show_complete_
                     anatomical_entity.save()
                     if show_complete_logs:
                         self.stdout.write(
-                            self.style.SUCCESS(f"Updated {anatomical_entity.ontology_uri} name to {name}."))
+                            self.style.SUCCESS(f"Updated {anatomical_entity.ontology_uri} name to {name}.")
+                        )
 
             processed_uris.add(ontology_uri)
 
-            if synonym and synonym.lower() not in [s.alias.lower() for s in anatomical_entity.synonyms.all()]:
-                synonym_accumulator.append(Synonym(anatomical_entity=anatomical_entity, alias=synonym))
-                if show_complete_logs:
-                    self.stdout.write(
-                        self.style.SUCCESS(f"Synonym '{synonym}' added for {anatomical_entity.ontology_uri}."))
+            synonym_key = (ontology_uri, synonym.lower()) if synonym else None
+            if synonym and synonym_key not in unique_synonyms:
+                if not Synonym.objects.filter(anatomical_entity=anatomical_entity, name__iexact=synonym).exists():
+                    unique_synonyms[synonym_key] = Synonym(anatomical_entity=anatomical_entity, name=synonym)
+                    if show_complete_logs:
+                        self.stdout.write(
+                            self.style.SUCCESS(f"Synonym '{synonym}' added for {anatomical_entity.ontology_uri}."))
         except IntegrityError as e:
             self.stdout.write(self.style.ERROR(f"Error processing {ontology_uri}: {e}"))
 
     @transaction.atomic
     def handle(self, *args, **options):
         start_time = time.time()
         show_complete_logs = options['show_complete_logs']
-        synonym_accumulator = []
+        unique_synonyms = {}
+        processed_uris = set()
 
         for csv_file in options["csv_files"]:
-            processed_uris = set()
             try:
                 with open(csv_file, newline="", encoding="utf-8", errors="ignore") as csvfile:
                     reader = csv.DictReader(csvfile, delimiter=",", quotechar='"')
@@ -65,21 +68,24 @@ def handle(self, *args, **options):
                         name = row[NAME].strip()
                         synonym = row[SYNONYM].strip() if row[SYNONYM] else None
 
-                        self._process_anatomical_entity(name, ontology_uri, synonym, show_complete_logs,
-                                                        processed_uris, synonym_accumulator)
+                        self._process_anatomical_entity(name, ontology_uri, synonym, show_complete_logs, processed_uris,
+                                                        unique_synonyms)
 
-                        if len(synonym_accumulator) >= BULK_LIMIT:
-                            Synonym.objects.bulk_create(synonym_accumulator)
-                            synonym_accumulator = []
+                        if len(unique_synonyms) >= BULK_LIMIT:
+                            Synonym.objects.bulk_create(unique_synonyms.values(), ignore_conflicts=True)
+                            unique_synonyms.clear()
 
             except FileNotFoundError:
                 self.stdout.write(self.style.ERROR(f"File {csv_file} does not exist."))
             except Exception as e:
                 self.stdout.write(self.style.ERROR(f"An error occurred while processing {csv_file}: {e}"))
 
             # Ensure any remaining synonyms are created
-            if synonym_accumulator:
-                Synonym.objects.bulk_create(synonym_accumulator)
+            if unique_synonyms:
+                try:
+                    Synonym.objects.bulk_create(unique_synonyms.values(), ignore_conflicts=True)
+                except Exception as e:
+                    self.stdout.write(self.style.ERROR(f"An error occurred during bulk creation: {e}"))
 
         end_time = time.time()
         self.stdout.write(self.style.SUCCESS(f"Operation completed in {end_time - start_time:.2f} seconds."))

backend/composer/migrations/0041_synonym.py

@@ -22,7 +22,7 @@ class Migration(migrations.Migration):
                         verbose_name="ID",
                     ),
                 ),
-                ("alias", models.CharField(db_index=True, max_length=200)),
+                ("name", models.CharField(db_index=True, max_length=200)),
                 (
                     "anatomical_entity",
                     models.ForeignKey(

backend/composer/migrations/0042_auto_20240313_1718.py

@@ -17,7 +17,7 @@ def deduplicate_anatomical_entities(apps, schema_editor):
             # If the ontology_uri is a duplicate, move this entity to Synonym
             primary_entity = AnatomicalEntity.objects.get(id=primary_anatomical_entities[entity.ontology_uri])
             # Create a synonym for the duplicate entity
-            Synonym.objects.create(anatomical_entity=primary_entity, alias=entity.name)
+            Synonym.objects.create(anatomical_entity=primary_entity, name=entity.name)
 
             # Update ConnectivityStatement origins to point to the primary entity
             for cs in ConnectivityStatement.objects.filter(origins=entity):

backend/composer/migrations/0045_alter_synonym_unique_together.py

@@ -0,0 +1,16 @@
+# Generated by Django 4.1.4 on 2024-03-18 17:57
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("composer", "0044_remove_anatomicalentity_ae_unique_upper_name"),
+    ]
+
+    operations = [
+        migrations.AlterUniqueTogether(
+            name="synonym",
+            unique_together={("anatomical_entity", "name")},
+        ),
+    ]

backend/composer/models.py

@@ -241,7 +241,10 @@ class Meta:
 
 class Synonym(models.Model):
     anatomical_entity = models.ForeignKey(AnatomicalEntity, on_delete=models.CASCADE, related_name="synonyms")
-    alias = models.CharField(max_length=200, db_index=True)
+    name = models.CharField(max_length=200, db_index=True)
+
+    class Meta:
+        unique_together = ('anatomical_entity', 'name',)
 
 
 class Tag(models.Model):