From 8842358d7ff9d6c2de86ca15abb7ea8c68e9aefc Mon Sep 17 00:00:00 2001
From: Olivier Leger <olivierleger@gmail.com>
Date: Tue, 24 Oct 2023 10:56:48 -0400
Subject: [PATCH] Skip redundant queries to DB

---
 .../commands/soft_delete_orphan_attachments.py     |  7 +++++--
 onadata/apps/logger/models/xform.py                | 14 ++++++++++----
 onadata/apps/logger/xform_instance_parser.py       |  3 +--
 3 files changed, 16 insertions(+), 8 deletions(-)

diff --git a/onadata/apps/logger/management/commands/soft_delete_orphan_attachments.py b/onadata/apps/logger/management/commands/soft_delete_orphan_attachments.py
index c2520e396..dfd5a73c0 100644
--- a/onadata/apps/logger/management/commands/soft_delete_orphan_attachments.py
+++ b/onadata/apps/logger/management/commands/soft_delete_orphan_attachments.py
@@ -11,7 +11,6 @@
 )
 from onadata.apps.logger.signals import pre_delete_attachment
 from onadata.libs.utils.logger_tools import get_soft_deleted_attachments
-from onadata.apps.viewer.models.parsed_instance import datetime_from_str
 
 
 class Command(BaseCommand):
@@ -62,7 +61,9 @@ def handle(self, *args, **kwargs):
             'to run on big databases'
         )
 
-        instance_ids = Attachment.objects.values_list('instance_id', flat=True).distinct()
+        instance_ids = Attachment.objects.values_list(
+            'instance_id', flat=True
+        ).distinct()
 
         if start_id:
             instance_ids = instance_ids.filter(instance_id__gte=start_id)
@@ -112,6 +113,8 @@ def handle(self, *args, **kwargs):
                 continue
 
             for soft_deleted_attachment in soft_deleted_attachments:
+                # Avoid fetching Instance object once again
+                soft_deleted_attachment.instance = instance
                 pre_delete_attachment(
                     soft_deleted_attachment, only_update_counters=True
                 )
diff --git a/onadata/apps/logger/models/xform.py b/onadata/apps/logger/models/xform.py
index 146a31e9a..84995d3ec 100644
--- a/onadata/apps/logger/models/xform.py
+++ b/onadata/apps/logger/models/xform.py
@@ -2,6 +2,7 @@
 import json
 import os
 import re
+from copy import deepcopy
 from io import BytesIO
 from xml.sax.saxutils import escape as xml_escape
 
@@ -129,10 +130,15 @@ def url(self):
             }
         )
 
-    def data_dictionary(self):
-        from onadata.apps.viewer.models.data_dictionary import\
-            DataDictionary
-        return DataDictionary.all_objects.get(pk=self.pk)
+    def data_dictionary(self, use_cache: bool = False):
+        from onadata.apps.viewer.models.data_dictionary import DataDictionary
+
+        if not use_cache:
+            return DataDictionary.all_objects.get(pk=self.pk)
+
+        xform_dict = deepcopy(self.__dict__)
+        xform_dict.pop('_state', None)
+        return DataDictionary(**xform_dict)
 
     @property
     def has_instances_with_geopoints(self):
diff --git a/onadata/apps/logger/xform_instance_parser.py b/onadata/apps/logger/xform_instance_parser.py
index cea69d0ba..a6c21c8de 100644
--- a/onadata/apps/logger/xform_instance_parser.py
+++ b/onadata/apps/logger/xform_instance_parser.py
@@ -366,10 +366,9 @@ def get_xform_media_question_xpaths(
     xform: 'onadata.apps.logger.models.XForm',
 ) -> list:
     logger = logging.getLogger('console_logger')
-    parser = XFormInstanceParser(xform.xml, xform.data_dictionary())
+    parser = XFormInstanceParser(xform.xml, xform.data_dictionary(use_cache=True))
     all_attributes = _get_all_attributes(parser.get_root_node())
     media_field_xpaths = []
-
     # This code expects that the attributes from Enketo Express are **always**
     # sent in the same order.
     # For example: