Tag cloud should not show empty tags

[Stefano Maffulli]

It turns out that spammers create tags for their spammy posts and when the posts are (rightfully) deleted, the tags remain. I would suggest that we either don't show the tags that have 0 questions or automatically remove empty tags from the database https://ask.openstack.org/en/tags/

[Tom Fifield (fifieldt)]

I found this as well.

Raising the priority, because some of the content they are trying to spam, eg phone numbers, is still in there.

[Tom Fifield (fifieldt)]

This is still a problem. Raising the priority again, because if we don't fix this, the spammers win.

[Evgeny Fadeev (evgeny-fadeev)]

todo: comment on why tags are allowed to stay empty in some cases

[Evgeny Fadeev (evgeny-fadeev)]

We could auto-delete unused tags from questions of blocked users.

This means actually that tags will be deleted from the questions that are marked as deleted via user blocking.
After that tag usage will be recounted and if it's 0, we delete the tags for good.

For cleanup this will also require going through all blocked users, their deleted questions and
clean up the tags.

[Stefano Maffulli (smaffulli)]

That sounds like a decent plan. Let's do it this way.

[Tom Fifield (fifieldt)]

This problem is still on the site.

There are empty tags including spam phone numbers in

https://ask.openstack.org/en/tags/

[Tom Fifield (fifieldt)]

Hi,

Any progress on this problem? It's been affecting the site for 6 months so far...

[Tom Fifield (fifieldt)]

This continues to be a problem, with the recent spam there are now thousands of spam tags

[Stefano Maffulli (smaffulli)]

Got this suggestions from Evgeny as a workaround:

```
you could first delete questions that are marked as deleted,
then run "python manage.py delete_unused_tags"

To delete threads:

1) back up the database

2) do the following:

# python manage.py shell
>>> from askbot.models import Thread
>>> Thread.objects.filter(deleted=True).delete()
>>> exit()

# python manage.py delete_unused_tags

```

[Tom Fifield (fifieldt)]

Tags continue to be listed that have sexual meanings. Raising this back to Critical level.

[Ian Y. Choi (ianychoi)]

Thank you, Tom Fifield. I did not know that my bug report was related to this report.

Comparing to my bug report on https://bugs.launchpad.net/openstack-community/+bug/1505068 ,
it seems that some guys executed the script mentioned by Stefano Maffulli on July 15 2014.
But the problem is that irrelevant sexual Korean tags are coming again...

So I want to suggest in another way: is it possible not to allow tags with non-English characters (e.g., disallowing Korean characters)?
In my opinion, the ask.openstack.org content could contain non-English characters, but
showing only English and numerical character tags in ask.openstack.org/en would be nice and
tags with other-language characters could be shown on ask.openstack.org for other languages
such as https://ask.openstack.org/zh/tags/ .

[Tom Fifield (fifieldt)]

Marking the other report as a duplicate of this one, but copying the info from Ian Y. Choi in here so we can discuss in one place:

One member in OpenStack Korea user group has recently found that there are irrelevant Korean tags on https://ask.openstack.org/en/tags/ like
: 강동퀸 강남건마 강남안마 강남오피 강남핸플 대전오피 마곡건마 망원건마 명일건마 부산유흥 부천핸플 분당오피 상수건마 서면오피 선릉오피 수원오피 일산건마 일산오피 천안오피 강남야구장 강남키스방 강남풀싸롱 강동티파니 부천키스방 선릉풀싸롱 역삼야구장 역삼풀싸롱 바카라사이트 연산동풀싸롱.

I am not sure how ask.openstack.org askbot collected and extracted those irrelevant tags,
but I am sure that those tag words are never related to OpenStack. Those Korean words have sexual meanings.

At first, I tried to identify which ask.openstack.org articles contain those words,
but I could not identify now because when I click one of those tags, there are no search results on the tag like
: https://ask.openstack.org/en/questions/scope:all/sort:activity-desc/tags:%EA%B0%95%EB%8F%99%ED%80%B8/page:1/ .

In sum, there are two kinds of bugs or to be fixed on ask.openstack.org:
1) Collects irrelavant tag words on ask.openstack.org
2) Cannot search articles using Korean tags or search keywords

[Tom Fifield (fifieldt)]

OK, so since the script was run, we have not had another spam flood - since the spam was caught in the moderation queue.

So, it seems even posts that end up in the moderation queue have their tags saved and posted on the site?

I have manually cleaned a few of the tags, but this problem really needs to be solved.

[Tom Fifield (fifieldt)]

This was

[Tom Fifield (fifieldt)]

This was not yet fixed.

[Tom Fifield (fifieldt)]

To check, scroll to the bottom of the tag cloud and try a tag or two.

[Stefano Maffulli (smaffulli)]

I did, the all seemed legit and had 1 :) I must have lucked out.

On 07/25/2016 06:10 PM, Tom Fifield wrote:
> To check, scroll to the bottom of the tag cloud and try a tag or two.
>
> ** Changed in: openstack-community
> Importance: Critical => High
>

[Tom Fifield (fifieldt)]

yeah, sometimes hard to find!

example:

https://ask.openstack.org/en/questions/scope:all/sort:activity-desc/tags:unlimited/page:1/