API/tasks to remove duplicate versions of concepts

core/common/tasks.py

@@ -11,6 +11,7 @@
 from django.core.exceptions import ValidationError
 from django.core.mail import EmailMessage
 from django.core.management import call_command
+from django.db.models import F
 from django.template.loader import render_to_string
 from django.utils import timezone
 from django_elasticsearch_dsl.registries import registry
@@ -709,3 +710,24 @@ def calculate_checksums(resource_type, resource_id):
                     instance.get_latest_version().set_checksums()
                 if not instance.is_versioned_object:
                     instance.versioned_object.set_checksums()
+
+
+@app.task(ignore_result=True, retry_kwargs={'max_retries': 0})
+def delete_duplicate_concept_versions(source_mnemonic, source_filters=None, concept_filters=None):  # pragma: no cover
+    source_filters = source_filters or {}
+    concept_filters = concept_filters or {}
+    from core.sources.models import Source
+    source = Source.objects.filter(version='HEAD', mnemonic=source_mnemonic).filter(**source_filters).first()
+    deleted = 0
+    for concept in source.concepts_set.exclude(id=F('versioned_object_id')).exclude(
+            is_latest_version=True).filter(**concept_filters).iterator(chunk_size=100):
+        source_versions = concept.sources
+        count = source_versions.count()
+        if not concept.comment and count == 0 or (count == 1 and source_versions.first().version == 'HEAD'):
+            if not concept.references.exists() and not concept.expansion_set.exists():
+                if not concept.mappings_from.exists() and not concept.mappings_to.exists():
+                    print(f"***Deleting: {concept.uri}***")
+                    concept.delete()
+                    deleted += 1
+
+    print("***DELETED***", deleted)

core/common/views.py

@@ -15,7 +15,7 @@
 from pydash import get, compact, flatten
 from rest_framework import response, generics, status
 from rest_framework.generics import ListAPIView, RetrieveUpdateDestroyAPIView
-from rest_framework.permissions import AllowAny, IsAuthenticated
+from rest_framework.permissions import AllowAny, IsAuthenticated, IsAdminUser
 from rest_framework.response import Response
 from rest_framework.views import APIView
 
@@ -1089,8 +1089,8 @@ def task_response(self, task, queue='default'):
             status=status.HTTP_202_ACCEPTED
         )
 
-    def perform_task(self, task_func, task_args, queue='default'):
-        is_async = self.is_async_requested()
+    def perform_task(self, task_func, task_args, queue='default', is_default_async=False):
+        is_async = is_default_async or self.is_async_requested()
         if self.is_inline_requested() or (get(settings, 'TEST_MODE', False) and not is_async):
             result = task_func(*task_args)
         else:
@@ -1108,3 +1108,22 @@ def perform_task(self, task_func, task_args, queue='default'):
                 return self.task_response(task, queue)
 
         return result
+
+
+class ConceptDuplicateDeleteView(BaseAPIView, TaskMixin):  # pragma: no-cover
+    permission_classes = (IsAdminUser, )
+
+    def post(self, _):
+        source_mnemonic = self.request.data.get('source_mnemonic', None)
+        source_filters = self.request.data.get('source_filters', None) or {}
+        concept_filters = self.request.data.get('concept_filters', None) or {}
+        if not source_mnemonic:
+            raise Http400(detail='source_mnemonic is required.')
+
+        from core.common.tasks import delete_duplicate_concept_versions
+        result = self.perform_task(
+            task_func=delete_duplicate_concept_versions,
+            task_args=(source_mnemonic, source_filters, concept_filters),
+            is_default_async=True
+        )
+        return result

core/urls.py

@@ -27,7 +27,7 @@
 from core.common.constants import NAMESPACE_PATTERN
 from core.common.utils import get_api_base_url
 from core.common.views import RootView, FeedbackView, APIVersionView, ChangeLogView, StandardChecksumView, \
-    SmartChecksumView
+    SmartChecksumView, ConceptDuplicateDeleteView
 from core.concepts.views import ConceptsHierarchyAmendAdminView
 from core.importers.views import BulkImportView
 
@@ -91,6 +91,9 @@
     ),
     path('manage/bulkimport/', BulkImportView.as_view(), name='bulk_import_urls'),
     path('toggles/', include('core.toggles.urls'), name='toggles'),
+    path(
+        'concepts-duplicate-versions-delete/',
+        ConceptDuplicateDeleteView.as_view(), name='concepts-duplicate-versions-delete'),
 ]
 
 handler500 = 'rest_framework.exceptions.server_error'