Unit tests for ClinVar namespaces in variant data CSV export

Author: jstone-dev

tests/helpers/util/score_set.py

@@ -225,6 +225,23 @@ def link_clinical_controls_to_mapped_variants(db, score_set):
     db.commit()
 
 
+def link_clinvar_control_to_mapped_variant(db, score_set):
+    """Link the seeded ClinVar clinical control (id=1) to the first mapped variant of a score set."""
+    mapped_variants = db.scalars(
+        select(MappedVariantDbModel)
+        .join(VariantDbModel)
+        .join(ScoreSetDbModel)
+        .where(ScoreSetDbModel.urn == score_set["urn"])
+    ).all()
+
+    mapped_variants[0].clinical_controls.append(
+        db.scalar(select(ClinicalControlDbModel).where(ClinicalControlDbModel.id == 1))
+    )
+
+    db.add(mapped_variants[0])
+    db.commit()
+
+
 def link_gnomad_variants_to_mapped_variants(db, score_set):
     mapped_variants = db.scalars(
         select(MappedVariantDbModel)

tests/routers/test_score_set.py

@@ -37,6 +37,7 @@
     TEST_BIORXIV_IDENTIFIER,
     TEST_BRNICH_SCORE_CALIBRATION_CLASS_BASED,
     TEST_BRNICH_SCORE_CALIBRATION_RANGE_BASED,
+    TEST_CLINVAR_CONTROL,
     TEST_CROSSREF_IDENTIFIER,
     TEST_EXPERIMENT_WITH_KEYWORD,
     TEST_GNOMAD_DATA_VERSION,
@@ -76,6 +77,7 @@
     create_seq_score_set_with_mapped_variants,
     create_seq_score_set_with_variants,
     link_clinical_controls_to_mapped_variants,
+    link_clinvar_control_to_mapped_variant,
     link_gnomad_variants_to_mapped_variants,
     publish_score_set,
 )
@@ -3154,6 +3156,128 @@ def test_download_gnomad_file_in_variant_data_path(session, data_provider, clien
     assert "gnomad.gnomad_af" in reader.fieldnames
 
 
+def test_download_clinvar_namespace_in_variant_data_path(session, data_provider, client, setup_router_db, data_files):
+    """ClinVar namespace returns clinical_significance and clinical_review_status columns with correct values."""
+    # The ClinVar control seeded in setup_router_db has db_version="11_2024", mapping to namespace clinvar.2024_11.
+    clinvar_namespace = "clinvar.2024_11"
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set_with_mapped_variants(
+        client, session, data_provider, experiment["urn"], data_files / "scores.csv"
+    )
+    link_clinvar_control_to_mapped_variant(session, score_set)
+
+    with patch.object(arq.ArqRedis, "enqueue_job", return_value=None) as worker_queue:
+        published_score_set = publish_score_set(client, score_set["urn"])
+        worker_queue.assert_called_once()
+
+    response = client.get(
+        f"/api/v1/score-sets/{published_score_set['urn']}/variants/data"
+        f"?namespaces={clinvar_namespace}&drop_na_columns=false"
+    )
+    assert response.status_code == 200
+    reader = csv.DictReader(StringIO(response.text))
+    assert f"{clinvar_namespace}.clinical_significance" in reader.fieldnames
+    assert f"{clinvar_namespace}.clinical_review_status" in reader.fieldnames
+
+    rows = list(reader)
+    # The first variant is linked to the ClinVar control; check its values.
+    assert rows[0][f"{clinvar_namespace}.clinical_significance"] == TEST_CLINVAR_CONTROL["clinical_significance"]
+    assert rows[0][f"{clinvar_namespace}.clinical_review_status"] == TEST_CLINVAR_CONTROL["clinical_review_status"]
+    # Other variants have no linked control for this version; they should be NA.
+    assert all(row[f"{clinvar_namespace}.clinical_significance"] == "NA" for row in rows[1:])
+    assert all(row[f"{clinvar_namespace}.clinical_review_status"] == "NA" for row in rows[1:])
+
+
+def test_download_clinvar_namespace_with_no_matching_version(
+    session, data_provider, client, setup_router_db, data_files
+):
+    """When no controls match the requested ClinVar version, all rows return NA."""
+    # clinvar.2023_01 does not match the seeded control (11_2024), so all rows should be NA.
+    clinvar_namespace = "clinvar.2023_01"
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set_with_mapped_variants(
+        client, session, data_provider, experiment["urn"], data_files / "scores.csv"
+    )
+    link_clinvar_control_to_mapped_variant(session, score_set)
+
+    with patch.object(arq.ArqRedis, "enqueue_job", return_value=None) as worker_queue:
+        published_score_set = publish_score_set(client, score_set["urn"])
+        worker_queue.assert_called_once()
+
+    response = client.get(
+        f"/api/v1/score-sets/{published_score_set['urn']}/variants/data"
+        f"?namespaces={clinvar_namespace}&drop_na_columns=false"
+    )
+    assert response.status_code == 200
+    reader = csv.DictReader(StringIO(response.text))
+    assert f"{clinvar_namespace}.clinical_significance" in reader.fieldnames
+    assert f"{clinvar_namespace}.clinical_review_status" in reader.fieldnames
+
+    rows = list(reader)
+    assert all(row[f"{clinvar_namespace}.clinical_significance"] == "NA" for row in rows)
+    assert all(row[f"{clinvar_namespace}.clinical_review_status"] == "NA" for row in rows)
+
+
+def test_download_multiple_clinvar_namespaces_in_variant_data_path(
+    session, data_provider, client, setup_router_db, data_files
+):
+    """Multiple ClinVar namespaces produce distinct column sets; only the matching version has real data."""
+    matching_ns = "clinvar.2024_11"  # matches db_version="11_2024" seeded in setup_router_db
+    non_matching_ns = "clinvar.2023_01"  # no controls with this version
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set_with_mapped_variants(
+        client, session, data_provider, experiment["urn"], data_files / "scores.csv"
+    )
+    link_clinvar_control_to_mapped_variant(session, score_set)
+
+    with patch.object(arq.ArqRedis, "enqueue_job", return_value=None) as worker_queue:
+        published_score_set = publish_score_set(client, score_set["urn"])
+        worker_queue.assert_called_once()
+
+    response = client.get(
+        f"/api/v1/score-sets/{published_score_set['urn']}/variants/data"
+        f"?namespaces={matching_ns}&namespaces={non_matching_ns}&drop_na_columns=false"
+    )
+    assert response.status_code == 200
+    reader = csv.DictReader(StringIO(response.text))
+    fieldnames = reader.fieldnames
+    # Both namespaces produce columns.
+    assert f"{matching_ns}.clinical_significance" in fieldnames
+    assert f"{matching_ns}.clinical_review_status" in fieldnames
+    assert f"{non_matching_ns}.clinical_significance" in fieldnames
+    assert f"{non_matching_ns}.clinical_review_status" in fieldnames
+
+    rows = list(reader)
+    # Matching version: first variant has data.
+    assert rows[0][f"{matching_ns}.clinical_significance"] == TEST_CLINVAR_CONTROL["clinical_significance"]
+    assert rows[0][f"{matching_ns}.clinical_review_status"] == TEST_CLINVAR_CONTROL["clinical_review_status"]
+    # Non-matching version: all rows are NA.
+    assert all(row[f"{non_matching_ns}.clinical_significance"] == "NA" for row in rows)
+    assert all(row[f"{non_matching_ns}.clinical_review_status"] == "NA" for row in rows)
+
+
+def test_invalid_clinvar_namespace_returns_422(client, setup_router_db, data_files):
+    """A clinvar namespace with an out-of-range month (13) is rejected with 422."""
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set(client, experiment["urn"])
+
+    response = client.get(
+        f"/api/v1/score-sets/{score_set['urn']}/variants/data?namespaces=clinvar.2024_13"
+    )
+    assert response.status_code == 422
+
+
+def test_unrecognized_namespace_returns_422(client, setup_router_db, data_files):
+    """An entirely unrecognized namespace string is rejected with 422."""
+    experiment = create_experiment(client)
+    score_set = create_seq_score_set(client, experiment["urn"])
+
+    response = client.get(
+        f"/api/v1/score-sets/{score_set['urn']}/variants/data?namespaces=unknown_namespace"
+    )
+    assert response.status_code == 422
+
+
 ########################################################################################################################
 # Fetching clinical controls and control options for a score set
 ########################################################################################################################