1239756 - Purge taggit_tag table of dead tags

Status: RESOLVED FIXED

(developer.mozilla.org Graveyard :: Code Cleanup, defect)

Description

[Rob Hudson [:robhudson]]

As content types have been removed from MDN, we have tags in the taggit_tag table that no longer have content associated with them. It might be wise to clean this up.

Comment 1

[Safwan Rahman (:safwan)]

How to determine which tags are not associated with any content?

I have run this queryset and found following result


>>> tags_content = TaggedItem.objects.all().values_list("tag__id", flat=True)
>>> Tag.objects.exclude(id__in=tags_content).count()
9204

Seems like, the 9204 is the unused Tag

Comment 2

[John Whitlock [:jwhitlock]]

Tags that have no related content should be removed, but tags to models that are no longer present also need cleaning up.

>>> count_data = TaggedItem.objects.values('content_type_id').annotate(Count('content_type_id'))
>>> for data in count_data:
>>>     ct_id = data['content_type_id']
>>>     tag_count = data['content_type_id__count']
>>>     ct = ContentType.objects.get(id=ct_id)
>>>     is_model = ct.model_class() is not None
>>>     print("%s (%s): %d" % (ct.model, "is model" if is_model else "NOT model", tag_count))

Result:

userprofile (NOT model): 33129
filter (is model): 35
user (is model): 8005

The 33,129 TaggedItems can be cleared, and then the ContentTypes for removed for models that have been dropped from the code.

This is production database cleanup, not a code change.

Comment 3

[MDN Team (:mdn-dev)]

Commits pushed to master at https://github.com/mozilla/kuma

https://github.com/mozilla/kuma/commit/123131edae378d69d4bcc0c4f276d46bb6ad89bf
[Bug 1239756] Purge taggit_tag table of dead tags

https://github.com/mozilla/kuma/commit/2f315de27ff521d966d30798349c946727241106
Merge pull request #4047 from safwanrahman/taggit

[Bug 1239756] Purge taggit_tag table of dead tags

Comment 4

[John Whitlock [:jwhitlock]]

Manually ran cleanup script in stage and prod for stale content due to removed tables:

https://gist.github.com/jwhitlock/20eb434466bab9b192a23476cf03a8ce

Manually ran cleanup of unused tags in stage and prod.

Migrations should be no-ops when applied.

Comment 5

[John Whitlock [:jwhitlock]]

Migration run in production and staging.