perf: optimize find_all_tags_for with caching and indexing

[magajh]

Description

This is a backport of the changes made in #126 to solve this issue #119

This PR introduces optimizations to the find_all_tags_for method and related functionality to improve performance when working with large datasets in the Tag model. Specifically:

1. Caching for ContentType Lookups:

   • Added cache to store ContentType objects for frequently used model names. This reduces redundant database queries for ContentType.objects.get(), especially in high-traffic scenarios or repeated calls.

2. Database Indexes:

   • Added an index on the target_type and target_object_id fields in the Tag model.
   • This index improves query performance for operations that filter tags by target_type and target_object_id, particularly in the find_all_tags_for method.

3. Optimized Query Logic:

   • Removed the list() call on target.values_list("id", flat=True) in the find_all_tags_for method.
   • This ensures the query is fully managed by the database, avoiding unnecessary memory overhead and improving scalability for large result sets.

Why These Changes Were Made

• The find_all_tags_for method was identified as a performance bottleneck, particularly for large datasets in the Tag table (~6 million records).
• Redundant ContentType lookups were causing unnecessary database queries.
• Queries filtering by target_type and target_object_id lacked an index, leading to slower performance on large datasets.

[magajh]

@eduNEXT/heimdall just so you know these changes were already deployed some time ago to the nimbus saas repository to fix #119, and they are working fine

[DeimerM]

Hi @magajh thanks a lot for your help! LGTM!