fix: lift over orphaned variant annotations (#1124)

maintenance/management/commands/generate_result_set.py

@@ -1,4 +1,6 @@
 """Django command for generating query result sets."""
+import json
+
 from django.core.exceptions import ValidationError
 from django.core.management.base import BaseCommand
 from django.db import transaction
@@ -63,42 +65,38 @@ def handle(self, *args, **options):
             msg = "Creating query set for all cases."
 
         self.stdout.write(self.style.NOTICE(msg))
-        msg_warning = None
+        msg_warning = ""
+        msg_orphans = ""
 
         if options["async"]:
-            count, fill_count = create_queryresultset(
+            count, fill_count, orphans = create_queryresultset(
                 options["case_uuid"], options["project_uuid"], options["all"]
             )
             msg = "Done creating result sets:"
             if (
                 count["smallvariantqueryresultset"]
                 or count["svqueryresultset"]
-                or fill_count["small_variants"]
-                or fill_count["structural_variants"]
+                or fill_count["small_variant_annotations"]
+                or fill_count["structural_variant_annotations"]
             ):
                 msg += f"""
 - SmallVariantQueryResultSets created: {count['smallvariantqueryresultset']}
 - SvQueryResultSets created: {count['svqueryresultset']}
-- User annotations added to SmallVariantQueryResultSets: {fill_count['small_variants']}
-- User annotations added to SvQueryResultSets: {fill_count['structural_variants']}"""
+- User annotations added to SmallVariantQueryResultSets: {fill_count['small_variant_annotations']}
+- User annotations added to SvQueryResultSets: {fill_count['structural_variant_annotations']}"""
             else:
                 msg += "\n- Nothing to do."
             if (
-                fill_count["small_variants_orphaned"]
-                or fill_count["structural_variants_orphaned"]
-                or fill_count["small_variants_no_query_result_set"]
-                or fill_count["structural_variants_no_query_result_set"]
-                or fill_count["small_variants_no_case_result_set"]
-                or fill_count["structural_variants_no_case_result_set"]
+                fill_count["small_variant_annotations_orphaned"]
+                or fill_count["structural_variant_annotations_orphaned"]
             ):
                 msg_warning = f"""
-WARNING! There are orphaned user annotations or cases or queries without result set:
-- Orphaned small variant user annotations: {fill_count['small_variants_orphaned']}
-- Small variant queries without result set: {fill_count['small_variants_no_query_result_set']}
-- Cases without small variant query result set: {fill_count['small_variants_no_case_result_set']}
-- Orphaned structural variant user annotations: {fill_count['structural_variants_orphaned']}
-- Structural variant queries without result set: {fill_count['structural_variants_no_query_result_set']}
-- Cases without structural variant query result set: {fill_count['structural_variants_no_case_result_set']}""".lstrip()
+WARNING! There are orphaned user annotations:
+- Orphaned small variant user annotations: {fill_count['small_variant_annotations_orphaned']}
+- Orphaned structural variant user annotations: {fill_count['structural_variant_annotations_orphaned']}""".lstrip()
+                msg_orphans = json.dumps(orphans, indent=1)
+                with open("orphans.json", "w") as f:
+                    json.dump(orphans, f, indent=1)
         else:
             task_create_queryresultset.delay(
                 options["case_uuid"], options["project_uuid"], options["all"]
@@ -108,3 +106,5 @@ def handle(self, *args, **options):
         self.stdout.write(self.style.SUCCESS(msg))
         if msg_warning:
             self.stdout.write(self.style.WARNING(msg_warning))
+        if msg_orphans:
+            self.stdout.write(self.style.NOTICE(msg_orphans))

utils/kickoff_orphaned_annotation_query.py

@@ -0,0 +1,41 @@
+import json
+import sys
+
+import requests
+
+BASE_URL = "http://"
+TOKEN = "XXX"
+SETTINGS_URL = f"{BASE_URL}/variants/api/query-case/query-settings-shortcut"
+QUERY_URL = f"{BASE_URL}/variants/api/query/list-create"
+
+
+headers = {"Authorization": "TOKEN " + TOKEN}
+
+
+def read_file(file):
+    with open(file, "r") as f:
+        return json.load(f)
+
+
+def main():
+    orphans = read_file(sys.argv[1])
+    for case_uuid, case_orphans in orphans.items():
+        if not case_orphans["small_variants"]:
+            continue
+        response = requests.get(f"{SETTINGS_URL}/{case_uuid}/", headers=headers)
+        if response.status_code != 200:
+            print(f"Error: {response.status_code} ({case_uuid})")
+            continue
+        response = response.json()
+        response["query_settings"]["selected_variants"] = case_orphans["small_variants"]
+        response = requests.post(f"{QUERY_URL}/{case_uuid}/", headers=headers, json=response)
+        print(
+            f"Query created for case {case_uuid}, {len(case_orphans['small_variants'])} orphans, status code: {response.status_code}"
+        )
+
+
+if __name__ == "__main__":
+    if len(sys.argv) != 2:
+        print("Usage: kickoff_orphaned_annotation_query.py <orphans.json>")
+        sys.exit(1)
+    main()

variants/queries.py

@@ -1567,6 +1567,27 @@ def extend_selectable(self, query_parts):
         )
 
 
+class ExtendQueryPartsSelectVariants(ExtendQueryPartsBase):
+    def extend_conditions(self, _query_parts):
+        condition = []
+
+        if not self.kwargs.get("selected_variants"):
+            return []
+
+        for variant in self.kwargs["selected_variants"]:
+            release, chromosome, start, reference, alternative = variant.split("-")
+            condition.append(
+                and_(
+                    SmallVariant.sa.release == release,
+                    SmallVariant.sa.chromosome == chromosome,
+                    SmallVariant.sa.start == start,
+                    SmallVariant.sa.reference == reference,
+                    SmallVariant.sa.alternative == alternative,
+                )
+            )
+        return [or_(*condition)]
+
+
 #: QueryPartsBuilderExtender classes list for cases.
 extender_classes_base = [
     ExtendQueryPartsCaseJoinAndFilter,
@@ -1640,6 +1661,7 @@ def get_qp_extender_classes(self):
             ExtendQueryPartsGeneSymbolJoin,
             ExtendQueryPartsAcmgJoin,
             ExtendQueryPartsMgiJoin,
+            ExtendQueryPartsSelectVariants,
         ]
 
 

variants/query_schemas.py

@@ -188,6 +188,8 @@ class CaseQueryV1:
     quality: typing.Dict[str, QualitySettingsV1]
     genotype: typing.Dict[str, typing.Optional[GenotypeChoiceV1]]
 
+    selected_variants: typing.Optional[typing.List[str]] = None
+
     transcripts_coding: bool = True
     transcripts_noncoding: bool = False
 
@@ -296,6 +298,7 @@ class QueryJsonToFormConverter:
 
     def convert(self, case: Case, query: CaseQueryV1) -> typing.Dict[str, typing.Any]:
         result = {
+            "selected_variants": query.selected_variants,
             "database_select": query.database,
             "var_type_snv": query.var_type_snv,
             "var_type_mnv": query.var_type_mnv,

variants/schemas/case-query-v1.json

@@ -10,6 +10,24 @@
     "genotype"
   ],
   "properties": {
+    "selected_variants": {
+      "anyOf": [
+        {
+          "type": "null"
+        },
+        {
+          "$id": "#/properties/selected_variants",
+          "type": "array",
+          "title": "The selected_variants schema",
+          "description": "Limit query to specified variants",
+          "default": [],
+          "examples": [
+            "GRCh37-1-1000000-A-G",
+            "GRCh37-1-1000001-T-C"
+          ]
+        }
+      ]
+    },
     "database": {
       "$id": "#/properties/database",
       "type": "string",