Create materialized view for variant publication dates

Author: bencap

alembic/versions/b85bc7b1bec7_materialized_view_for_variant_statistics.py

@@ -0,0 +1,75 @@
+"""materialized view for variant statistics
+Revision ID: b85bc7b1bec7
+Revises: c404b6719110
+Create Date: 2025-03-14 01:53:19.898198
+"""
+
+from alembic import op
+from alembic_utils.pg_materialized_view import PGMaterializedView
+from sqlalchemy.dialects import postgresql
+
+from mavedb.models.published_variant import signature, definition
+
+
+# revision identifiers, used by Alembic.
+revision = "b85bc7b1bec7"
+down_revision = "c404b6719110"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    op.create_entity(
+        PGMaterializedView(
+            schema="public",
+            signature=signature,
+            definition=definition.compile(dialect=postgresql.dialect()).string,
+            with_data=True,
+        )
+    )
+    op.create_index(
+        f"idx_{signature}_variant_id",
+        signature,
+        ["variant_id"],
+        unique=False,
+    )
+    op.create_index(
+        f"idx_{signature}_variant_urn",
+        signature,
+        ["variant_urn"],
+        unique=False,
+    )
+    op.create_index(
+        f"idx_{signature}_score_set_id",
+        signature,
+        ["score_set_id"],
+        unique=False,
+    )
+    op.create_index(
+        f"idx_{signature}_score_set_urn",
+        signature,
+        ["score_set_urn"],
+        unique=False,
+    )
+    op.create_index(
+        f"idx_{signature}_mapped_variant_id",
+        signature,
+        ["mapped_variant_id"],
+        unique=True,
+    )
+
+
+def downgrade():
+    op.drop_index(f"idx_{signature}_variant_id", signature)
+    op.drop_index(f"idx_{signature}_variant_urn", signature)
+    op.drop_index(f"idx_{signature}_mapped_variant_id", signature)
+    op.drop_index(f"idx_{signature}_score_set_id", signature)
+    op.drop_index(f"idx_{signature}_score_set_urn", signature)
+    op.drop_entity(
+        PGMaterializedView(
+            schema="public",
+            signature=signature,
+            definition=definition.compile(dialect=postgresql.dialect()).string,
+            with_data=True,
+        )
+    )

src/mavedb/models/__init__.py

@@ -12,6 +12,7 @@
     "license",
     "mapped_variant",
     "publication_identifier",
+    "published_variant",
     "raw_read_identifier",
     "refseq_identifier",
     "refseq_offset",

src/mavedb/models/published_variant.py

@@ -0,0 +1,45 @@
+from sqlalchemy import select, join
+
+from mavedb.db.view import MaterializedView, view
+
+from mavedb.models.score_set import ScoreSet
+from mavedb.models.variant import Variant
+from mavedb.models.mapped_variant import MappedVariant
+
+
+signature = "published_variants_materialized_view"
+definition = (
+    select(
+        Variant.id.label("variant_id"),
+        Variant.urn.label("variant_urn"),
+        MappedVariant.id.label("mapped_variant_id"),
+        ScoreSet.id.label("score_set_id"),
+        ScoreSet.urn.label("score_set_urn"),
+        ScoreSet.published_date.label("published_date"),
+        MappedVariant.current.label("current_mapped_variant"),
+    )
+    .select_from(
+        join(Variant, MappedVariant, Variant.id == MappedVariant.variant_id, isouter=True).join(
+            ScoreSet, ScoreSet.id == Variant.score_set_id
+        )
+    )
+    .where(
+        ScoreSet.published_date.is_not(None),
+    )
+)
+
+
+class PublishedVariantsMV(MaterializedView):
+    __table__ = view(
+        signature,
+        definition,
+        materialized=True,
+    )
+
+    variant_id = __table__.c.variant_id
+    variant_urn = __table__.c.variant_urn
+    mapped_variant_id = __table__.c.mapped_variant_id
+    score_set_id = __table__.c.score_set_id
+    score_set_urn = __table__.c.score_set_urn
+    published_date = __table__.c.published_date
+    current_mapped_variant = __table__.c.current_mapped_variant