Topic: There should be a bot that automatically removes invalid tags.

Posted under Tag Alias and Implication Suggestions

I have been removing invalid tags from post and there should be a bot that does it automatically. Who agress with me?

Invalid tags are usually supposed to be replaced with something else, not removed. Otherwise somebody would've already made a bot - it'd only take me 10 minutes.

EDIT: looking at your edit history I see you've actually mostly removed the invalid_tag tag rather than any other tags with the invalid category. I guess this one doesn't really need any manual intervention and could be automatically removed? I'll set up a bot if other people think it's fine to do.

For tags like gold instead of being removed by a bot it'd need manual intervention - replacing it with numerous tags such as gold_coin or gold_jewelry

EDIT2: apparently there's tag bots that indiscriminately just remove the gold tag so maybe nobody actually cares.

Updated

faucet said:
EDIT: looking at your edit history I see you've actually mostly removed the invalid_tag tag rather than any other tags with the invalid category. I guess this one doesn't really need any manual intervention and could be automatically removed? I'll set up a bot if other people think it's fine to do.

I think there might already be a bot that does that, though it runs periodically instead of constantly. Either that, or someone who takes great effort to deal with it. Within a day or two, the number of posts with invalid_tag drops significantly, regardless of how many were tagged.

faucet said:
EDIT2: apparently there's tag bots that indiscriminately just remove the gold tag so maybe nobody actually cares.

I might say that falls under tagging abuse. Like any other ambiguous/disambiguation tag, it should be looked at to find what the tagger meant by it. An attempt to tag gold dragon (and so could get gold_scales or yellow_fur)? A gold_coin? gold_(metal)? It's fine to autoremove invalid_tag because it leaves no hint as to what was attempted to be tagged and was probably unimportant, but other tags that are simply in the Invalid category should not be autoremoved until the image is looked over for applicable missing tags it could've meant.

Pup

Privileged

watsit said:
I think there might already be a bot that does that, though it runs periodically instead of constantly.

As well as the other things it checks my bot removes invalid_tag and tags in the the invalid category from posts when it first checks them. I could always stop it removing the category tags though.

Edit:
Also, given I log my bot's changes, I could always re-add all the invalid category tags that it's removed.

Updated

-1

posts with the invalid_tag were probably tagged by an inexperienced user. Even if there isn't any hint about what the user tried to tag, there is most likely other mistakes that need to be fixed.

Pro tip: removing invalid tags isn't that helpful if you don't identify the source of the problem. Sometimes you might need to tell someone that they are making a mistake.

pup said:
As well as the other things it checks my bot removes invalid_tag and tags in the the invalid category from posts when it first checks them. I could always stop it removing the category tags though.

Yes, it should stop indiscriminately removing tags from the Invalid category, because disambiguation tags should go there (there's a big open BUR for it, as well as various other BURs that move disambiguations to Invalid). The Invalid category is otherwise fairly useless compared to aliasing it all to invalid_tag if all its tags are automatically cleaned away. invalid_tag itself can be automatically cleaned away while any fundamentally useless tag can be left aliased to it (e.g. 0, 20_percent_cuter, etc), but other Invalid tags that indicate a mistake or which should be replaced with something less ambiguous should be left for someone to manually check and fix.

Updated

No, don't bot invalid tags. They need to be reviewed and the tags fixed. If there are invalid tags and the person who tagged them didn't notice, it's an indicator that the tag quality is already in question. Indiscriminately removing invalid tags is tag vandalism.

The tags on posts with invalid tags should be reviewed, updated, fixed, and then invalid tags can be removed at the same time. It shouldn't be the sole action of the edit.

Pup

Privileged

watsit said:
Yes, it should stop indiscriminately removing tags from the Invalid category, because disambiguation tags should go there [..]

kiranoot said:
No, don't bot invalid tags. They need to be reviewed and the tags fixed.

I didn't realise about the disambiguated tags, I'll update my bot and re-add the invalid tags it removed. I think I was only removing invalid_tag because it was being done by a previous bot that'd stopped, then also included the invalid category.

Pup

Privileged

Just a quick bump to say I've re-added the 15k invalid_tags and the tags in the invalid category my bot removed.

Now there's over 13K invalid_tags, which is causing the annoyances with alias suggestions. I thought it was fine for invalid_tag itself to be autoremoved because it's not helpful to keep around? It's just the others in the Invalid category that should be left to be checked over. There's currently now 20 pages (more than 1,425 posts) of invalid_tag status:deleted, and surely deleted posts have no reason to keep invalid_tag?

Pup

Privileged

watsit said:
Now there's over 13K invalid_tags, which is causing the annoyances with alias suggestions. I thought it was fine for invalid_tag itself to be autoremoved because it's not helpful to keep around? It's just the others in the Invalid category that should be left to be checked over. There's currently now 20 pages (more than 1,425 posts) of invalid_tag status:deleted, and surely deleted posts have no reason to keep invalid_tag?

Kira said not to bot invalid tags so I undid all the changes my bot made to them, without my bot removing them the deleted posts would still have those tags regardless. I'd have been fine just re-adding the category ones but given Kira's an admin and has said not to bot invalid tags I'd rather wait for an admin to say invalid_tag is ok to be removed, on deleted posts or otherwise, before removing them again.

pup said:
Kira said not to bot invalid tags so I undid all the changes my bot made to them, without my bot removing them the deleted posts would still have those tags regardless.

Yeah, I know. I'm just bringing the issue up for discussion/clarification.

I just recall a previous post on the topic where invalid_tag would still be used for "tags that I feel shouldn't appear on posts" and "it's a matter of weighing if anything can be gained from the tag appearing on the post, and if it is actually correctable or not", which to me suggests invalid_tag is reserved for tags that should just be removed outright because it has too much work to too little gain to try and fix up manually (separate from other tags in the Invalid category, which should be fixed up manually). Especially with how often invalid_tag gets tagged on posts, and when tags are newly aliased to invalid_tag for the aforementioned reasons after they've been used hundreds or thousands of times, it's impractical to keep up with by manually reviewing them one by one.

Pup

Privileged

watsit said:
Yeah, I know. I'm just bringing the issue up for discussion/clarification.

I just recall a previous post on the topic where invalid_tag would still be used for "tags that I feel shouldn't appear on posts" and "it's a matter of weighing if anything can be gained from the tag appearing on the post, and if it is actually correctable or not", which to me suggests invalid_tag is reserved for tags that should just be removed outright because it has too much work to too little gain to try and fix up manually (separate from other tags in the Invalid category, which should be fixed up manually). Especially with how often invalid_tag gets tagged on posts, and when tags are newly aliased to invalid_tag for the aforementioned reasons after they've been used hundreds or thousands of times, it's impractical to keep up with by manually reviewing them one by one.

I definitely agree with that, the only argument against removing invalid_tag is to say if somebody didn't check their tags after uploading and didn't remove the tag then there could be other tagging errors or general bad tagging on the post, though you could catch a lot of badly tagged posts with tagcount:<x, so it's not much of an argument against it. Not to mention that the tags that need extra checking should be in the category, not aliased, as you said.

pup said:
I definitely agree with that, the only argument against removing invalid_tag is to say if somebody didn't check their tags after uploading and didn't remove the tag then there could be other tagging errors or general bad tagging on the post, though you could catch a lot of badly tagged posts with tagcount:<x, so it's not much of an argument against it. Not to mention that the tags that need extra checking should be in the category, not aliased, as you said.

Hi

I think it is good to see the invalid_tag right after posting a new picture for the uploader or for someone that made an edit but for right now there are about 13k Invalid tags, so it should be ok that those tags get only a certain lifetime, like maybe 1-3 days after the last change.
After that time, usually no one cares enough to do the work of manually change it or even just delete it.
So I think it should be ok to just erase them after a certain time.

I don't know I'm here since 2012 at some point the invalid tag got established and over the time refined, so we get less bad tags overall here.
I thought they usually got just deleted in just some intervals, as most people don't know or care about them.
Personally I have them on my profile in the Tagging projects list as I noticed some admins got them on their profile too, as things you do like a side quest.

Sure it is good if you can replace them with a meaningful tag but if you look closely there will almost always a tag you could add to 90% of the posts here .

Updated

  • 1