optimise the database query by using RAW sql

contentcuration/contentcuration/viewsets/sync/endpoint.py

@@ -4,16 +4,15 @@
 bulk creates, updates, and deletes.
 """
 from celery import states
+from django.db import connection
 from django.db.models import Q
-from django_celery_results.models import TaskResult
 from rest_framework.authentication import SessionAuthentication
 from rest_framework.permissions import IsAuthenticated
 from rest_framework.response import Response
 from rest_framework.views import APIView
 
 from contentcuration.models import Change
 from contentcuration.models import Channel
-from contentcuration.models import CustomTaskMetadata
 from contentcuration.tasks import apply_channel_changes_task
 from contentcuration.tasks import apply_user_changes_task
 from contentcuration.viewsets.sync.constants import CHANNEL
@@ -124,29 +123,44 @@ def return_changes(self, request, channel_revs):
         return {"changes": changes, "errors": errors, "successes": successes}
 
     def return_tasks(self, request, channel_revs):
-        custom_task_objects = CustomTaskMetadata.objects.filter(
-            channel_id__in=channel_revs.keys(),
-            task_id__in=TaskResult.objects.filter(
-                status__in=[states.STARTED, states.FAILURE]
-            ).exclude(task_name__in=[apply_channel_changes_task.name, apply_user_changes_task.name]).values_list("task_id", flat=True)
-        )
+        sql_query = """
+        WITH CustomTaskCTE AS (
+                SELECT c.*, t.task_name, t.traceback, t.status
+                FROM contentcuration_customtaskmetadata c
+                INNER JOIN django_celery_results_taskresult t ON c.task_id = t.task_id
+                WHERE c.channel_id = ANY(%(channel_ids)s::uuid[])
+            )
+            SELECT t.task_id, t.task_name, t.traceback, c.progress, c.channel_id, t.status
+            FROM CustomTaskCTE c
+            JOIN django_celery_results_taskresult t ON c.task_id = t.task_id
+            WHERE t.status = ANY(%(statuses)s)
+            AND t.task_name NOT IN %(exclude_task_names)s;
+        """
+
+        params = {
+            'channel_ids': list(channel_revs.keys()),
+            'statuses': [states.STARTED, states.FAILURE],
+            'exclude_task_names': (apply_channel_changes_task.name, apply_user_changes_task.name)
+        }
+
+        with connection.cursor() as cursor:
+            cursor.execute(sql_query, params)
+            result = cursor.fetchall()
 
         response_payload = {
             "tasks": [],
         }
 
-        for custom_task in custom_task_objects:
-            task = TaskResult.objects.get(task_id=custom_task.task_id)
+        for row in result:
             task_data = {
-                "task_id": task.task_id,
-                "task_name": task.task_name,
-                "traceback": task.traceback,
-                "progress": custom_task.progress,
-                "channel_id": custom_task.channel_id,
-                "status": task.status,
+                "task_id": row[0],
+                "task_name": row[1],
+                "traceback": row[2],
+                "progress": row[3],
+                "channel_id": row[4],
+                "status": row[5],
             }
 
-            # Add the task data to the response_payload
             response_payload["tasks"].append(task_data)
 
         return response_payload