issue #1053 - cohort genotype versions - make common cohort version m…

…atch cohort.version

snpdb/migrations/0132_alter_cohortgenotypecollection_unique_together_and_more.py

@@ -0,0 +1,26 @@
+# Generated by Django 4.2.10 on 2024-05-20 01:39
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('snpdb', '0131_download_liftover_chain_files'),
+    ]
+
+    operations = [
+        migrations.AlterUniqueTogether(
+            name='cohortgenotypecollection',
+            unique_together=set(),
+        ),
+        migrations.AddField(
+            model_name='cohortgenotypecollection',
+            name='collection_type',
+            field=models.CharField(choices=[('C', 'Common'), ('U', 'Uncommon')], default='U', max_length=1),
+        ),
+        migrations.AlterUniqueTogether(
+            name='cohortgenotypecollection',
+            unique_together={('cohort', 'cohort_version', 'collection_type')},
+        ),
+    ]

snpdb/migrations/0133_one_off_set_cgc_common.py

@@ -0,0 +1,27 @@
+# Generated by Django 4.2.10 on 2024-05-20 01:39
+
+from django.db import migrations
+from django.db.models import Subquery, OuterRef
+
+
+def _one_off_set_cgc_common(apps, _schema_editor):
+    CohortGenotypeCollection = apps.get_model("snpdb", "CohortGenotypeCollection")
+    Cohort = apps.get_model("snpdb", "Cohort")
+
+    CGC_COMMON = "C"
+
+    # To avoid the unique_together (cohort/version) the common ones were left as 0
+    cgc_qs = CohortGenotypeCollection.objects.filter(common_filter__isnull=False, cohort_version=0)
+    cohort_version_subquery = Subquery(Cohort.objects.filter(id=OuterRef('cohort_id')).values('version')[:1])
+    cgc_qs.update(collection_type=CGC_COMMON, cohort_version=cohort_version_subquery)
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('snpdb', '0132_alter_cohortgenotypecollection_unique_together_and_more'),
+    ]
+
+    operations = [
+        migrations.RunPython(_one_off_set_cgc_common, reverse_code=lambda _, __: None),
+    ]

snpdb/models/models_cohort.py

@@ -26,7 +26,7 @@
 from library.preview_request import PreviewModelMixin, PreviewKeyValue
 from library.utils import invert_dict
 from patients.models_enums import Zygosity
-from snpdb.models.models_enums import ImportStatus
+from snpdb.models.models_enums import ImportStatus, CohortGenotypeCollectionType
 from snpdb.models.models_genome import GenomeBuild
 from snpdb.models.models_variant import Variant
 from snpdb.models.models_vcf import VCF, Sample
@@ -376,13 +376,23 @@ class CohortGenotypeCollection(RelatedModelsPartitionModel):
     celery_task = models.CharField(max_length=36, null=True)
     task_version = models.ForeignKey(CohortGenotypeTaskVersion, null=True, on_delete=CASCADE)
     marked_for_deletion = models.BooleanField(null=False, default=False)
-    # common_collection will be set on the 'interesting/rare' CGC
+    collection_type = models.CharField(max_length=1, choices=CohortGenotypeCollectionType.choices,
+                                       default=CohortGenotypeCollectionType.UNCOMMON)
+    # common_collection will be set on the 'uncommon' (interesting/rare) CGC
     common_collection = models.OneToOneField('self', null=True, related_name="uncommon", on_delete=CASCADE)
     # common filter will be set on the 'common' CGC
     common_filter = models.ForeignKey(CohortGenotypeCommonFilterVersion, null=True, on_delete=PROTECT)
 
     class Meta:
-        unique_together = ('cohort', 'cohort_version')
+        unique_together = ('cohort', 'cohort_version', 'collection_type')
+
+    def __str__(self) -> str:
+        parts = [f"CohortGenotypeCollection: {self.cohort}"]
+        if self.cohort_version != self.cohort.version:
+            parts.append(f"(v.{self.cohort.version} != {self.cohort_version})")
+        if self.common_filter:
+            parts.append(str(self.common_filter))
+        return " ".join(parts)
 
     def percent_common(self) -> float:
         common = self.common_collection.cohortgenotype_set.count()

snpdb/models/models_enums.py

@@ -227,3 +227,8 @@ class DataState(models.TextChoices):
     @staticmethod
     def should_create_new_record(data_state):
         return data_state not in [DataState.DELETED, DataState.SKIPPED]
+
+
+class CohortGenotypeCollectionType(models.TextChoices):
+    COMMON = "C", "Common"
+    UNCOMMON = "U", "Uncommon"

snpdb/tasks/cohort_genotype_tasks.py

@@ -75,19 +75,23 @@ def create_cohort_genotype_collection(cohort):
     name = f"{cohort.name} ({cohort.pk}:{cohort.version})"
     num_samples = cohort.cohortsample_set.count()
     common_collection = None
+    kwargs = {
+        "cohort": cohort,
+        "cohort_version": cohort.version,
+        "num_samples": num_samples,
+    }
+
     if common_filter := get_common_filter(cohort.genome_build):
         common_collection = CohortGenotypeCollection.objects.create(name=f"{name} common",
-                                                                    cohort=cohort,
-                                                                    cohort_version=0,  # so it isn't retrieved
                                                                     common_filter=common_filter,
-                                                                    num_samples=num_samples)
+                                                                    collection_type=CohortGenotypeCollectionType.COMMON,
+                                                                    **kwargs)
         logging.info(f"Created common collection: {common_collection}")
 
     collection = CohortGenotypeCollection.objects.create(name=name,
-                                                         cohort=cohort,
-                                                         cohort_version=cohort.version,
                                                          common_collection=common_collection,
-                                                         num_samples=num_samples)
+                                                         collection_type=CohortGenotypeCollectionType.UNCOMMON,
+                                                         **kwargs)
 
     logging.info(f"Created {collection}")
     return collection
@@ -96,6 +100,7 @@ def create_cohort_genotype_collection(cohort):
 def _get_sample_zygosity_count_sql(sample_value, zygosity):
     return f'CASE WHEN (({sample_value}) = \'{zygosity}\') THEN 1 ELSE 0 END'
 
+
 def _get_left_outer_join_on_variant(partition_table):
     return f'LEFT OUTER JOIN "{partition_table}" ON ("snpdb_variant"."id" = "{partition_table}"."variant_id")'