perf(users): further optimize /api/v2/users list endpoint for large datasets DEV-1236 (#6450)

### 📣 Summary
Fix remaining performance issues on the `/api/v2/users` endpoint when
handling very large datasets not covered by #6447

### 📖 Description
Although PR #6447 significantly improved the /api/v2/users list
performance, slow responses were still observed with very large
datasets. This update introduces additional optimizations to handle
those edge cases efficiently.

Author: noliveleger

kpi/views/v2/user.py

@@ -88,7 +88,6 @@ def get_queryset(self, *args, **kwargs):
             self.queryset = (
                 self.queryset.exclude(pk=settings.ANONYMOUS_USER_ID)
                 .select_related('extra_details')
-                .annotate(assets_count=Count('assets', distinct=True))
                 .order_by('id')
             )
 
@@ -106,6 +105,28 @@ def list(self, request, *args, **kwargs):
 
         return super().list(request, *args, **kwargs)
 
+    def paginate_queryset(self, queryset):
+        if not (page := super().paginate_queryset(queryset)):
+            return None
+
+        user_ids = (user.pk for user in page)
+        counts_map = {
+            asset['owner_id']: asset['assets_count']
+            for asset in (
+                Asset.objects.filter(owner_id__in=user_ids)
+                .values('owner_id')
+                .annotate(assets_count=Count('id'))
+            )
+        }
+
+        def _page_generator():
+            # Inject and yield on-the-fly assets count
+            for user in page:
+                user.assets_count = counts_map.get(user.pk, 0)
+                yield user
+
+        return _page_generator()
+
     @action(detail=True, methods=['GET'],
             url_path=r'migrate(?:/(?P<task_id>[\d\w\-]+))?')
     def migrate(self, request, task_id: str = None, **kwargs):