Forum:Image maintenance bot

From the RuneScape Wiki, the wiki for all things RuneScape
Jump to: navigation, search
Forums: Yew Grove > Image maintenance bot
Archive
This page or section is an archive.
Please do not edit the contents of this page.
This thread was archived on 26 March 2010 by Quarenon.

Hi, I had a couple of ideas for a bot that would perform some menial tasks with regards to image maintenance. Here's a list of what I had in mind:

  1. Tag images that are only for personal use with {{dper}} and post {{NoPP}} on the uploader's talk page. An image would trigger this if a) The image is used only in the User namespace and b) The image is used only in a small number of pages (maybe 5 or less?)
  2. Tag JPEG images that are missing a {{jpeg}} template.
  3. Categorize animated GIFs into Category:Animations if not already.
  4. Tag images exceeding 2 MB for deletion per RS:IMAGES.
  5. Remove "==Category==" from image descriptions. As far as I know, the section is always empty because the categories are shown elsewhere.

The bot would find these by occassionally checking the list of recently uploaded images; doing a run on existing images would be possible, too.

Not all of these ideas have to be approved at once, so feel free to oppose/support specific bits and the bot can run just the ideas that everyone agrees on. --Quarenon  Talk 16:08, 8 August 2009 (UTC)

Discussion

Support all - as nom. --Quarenon  Talk 16:09, 8 August 2009 (UTC)

Comments/Questions

  1. It would be better to "tag" non-mainspace images. Some images are used only in "User talk", "RuneScape", etc.
    • Could you estimate the number of images? We have roughly 21,000 images in total.
    • And, are you going to check the "WhatLinksHere" for each image? If yes, then I oppose. This would burden the servers too much.
  2. We have about 500 JPEGs in total, and 50 are listed in the category. So I support this, since there is only 100 images to process.
  3. We have roughly 3,000 GIFs. How would you differentiate normal GIFs from animated GIFs?
  4. Hopefully not speedy delete?
    • We should evaluate each image to check whether it can be compressed. Usually, animations are the ones bigger than 2MB.
    • If they cannot be compressed, then remove the link(s) to the image before tagging the images for deletion.
  5. We have about 20,000 images with "==Category==" in the description. Although I agree that they have to be removed, I remove them only when if I see them. I don't think this is important enough to be removed by a bot. I oppose this.

Handling 21,000 images can be huge task even for a bot. OMG! Ideally, the bot should target only specific images for processing.   az talk   18:27, 8 August 2009 (UTC)

Varied -

  1. Tag images that are only for personal use with {{dper}} and post {{NoPP}} on the uploader's talk page. An image would trigger this if a) The image is used only in the User namespace and b) The image is used only in a small number of pages (maybe 5 or less?) Oppose - Personal images are quicky found when one looks at someone's userpage, a bot is not needed. Plus, most images not in signatures only appear in pages maybe 2-3 times at most.
  2. Tag JPEG images that are missing a {{jpeg}} template. Support - If it does this correctly.
  3. Categorize animated GIFs into Category:Animations if not already. Support - This would also help in finding SD animations that arn't catgorized.
  4. Tag images exceeding 2 MB for deletion per RS:IMAGES. Neutral - For speedy deletion? I think it should just be tagged for compression... if we have a tag for that, anyway.
  5. Remove "==Category==" from image descriptions. As far as I know, the section is always empty because the categories are shown elsewhere. Support - Those sections are usless.

But like az said, it should only target the specific images that need to be handled, not the entire database of images. — Enigma 19:00, 8 August 2009 (UTC)

Other considerations

  1. I support tagging images only used in certain namespaces. To me, we should probably not tag images used in the Main, RuneScape, MediaWiki, and perhaps Template namespaces. I would guess that most personal images end up in the User, User talk, and Talk namespaces. I would also suggest something besides {{dper}}, because someone, might try to get an image for an article only to find it deleted before they can link to it.
  2. There are a few JPEG's that we keep for historical reasons (they can't be recreated) so if a JPEG has Template:Historical image on it, we don't want to tag it with {{jpeg}}
  3. Also if an animation is tagged with any subcategory (Agility animations, Animation lacks anti-aliasing, Emote animations, Special attack animations, & Spell animations) we don't want to redundantly add it to the general animation category as well.
    • Also, can we determine if a GIF is animated or not? If so, we could tag non-animated GIF's for replacement. If not, this would put a lot of still GIF's in the animation category incorrectly.
  4. Maybe we could tag it with a new template saying it's too big, and then work on the list by hand
  5. I'm pretty sure the "Category" heading is added by the rich text editor. Is there anyway we can change the settings of that? I agree that the heading is useless and can be deleted. Can we keep this from being added in the future?

These all sounds like good ideas though. I support them with tweaking and thanks for bringing more bot ideas to the table. Air rune.png Tollerach hates SoF Fire rune.png 19:55, 8 August 2009 (UTC)

Comments

  1. I think I am hearing from this thread that we want to have some sort of new template/category to encompass all non-RS/mainspace images?
  2. I can definitely exclude the JPEGs marked as historical.
  3. Animated GIFs can be determined by reading the GIF file header; it is known whether a GIF is animated by downloading and reading two specific bytes in the file, so downloading the entire file is not necessary. I can have it skip adding Category:Animations if it already contains a category that contains the word "animation" in it (or any other category for that matter).
  4. Yeah, I will change the proposal so it gets tagged with something else other than with deletion. If we use a category for large images then the bot can put images in there and a person can evaluate whether to keep it (and compress it down if needed) or tag it for deletion. The MediaWiki API provides for a way to search images on the basis of a minimum or maximum filesize so it should not be difficult to actually find these images.
  5. I don't know where the category header comes from, but if it's possible to stop it from being added in the first place, that would be great. As far as processing the large number of images for which this applies, I agree that it's not a big deal; the bot doesn't have to try to fix all of it at once. Just doing a few hundred or so edits each day would mean it's making some progress on a task that few people want to do by hand.

Don't forget that this proposal isn't just about checking existing images, which can indeed be a slow process due to the number of images. This sort of thing can be done on recently uploaded/changed images that occasionally have missing categories/templates, which translates to less work that a human needs to do. --Quarenon  Talk 04:00, 9 August 2009 (UTC)

So is this a modification of the proposal? If so, then I Support all the way. — Enigma 04:39, 9 August 2009 (UTC)

The ==Category== heading is added automatically whenever a user selects a category from the dropdown list in Special:Upload. (I think the "None selected" template is added if the user forgets to add the category.) Usually, in other wikis, the dropdown list is used for adding "Licensing" information (i.e. non-free image, copyleft image, etc.)

MediaWiki:License adds the header; and MediaWiki:Licenses contains the list of categories. As it involves MediaWiki software, I don't know if we could disable this automatic "header addition" without disabling the dropdown list in Special:Upload. If we blanked MediaWiki:License, the "Category" text from the Special:Upload page may disappear, and instead of having ==Category== added to the image description, we might end up with ==== being added instead.   az talk   04:05, 10 August 2009 (UTC)

I researched this a bit and it seems that blanking MediaWiki:License results in it defaulting back to the "Licensing" text Frown. We certainly want to keep the category drop-down list on the upload form, but it may not be possible to have that and omit the header at the same time. --Quarenon  Talk 11:32, 10 August 2009 (UTC)
Turns out there is a way to hide the header using CSS (thanks to mendel-de at #wikia):
.ns-6 a#Category + h2 {
	display: none;
}
--Quarenon  Talk 14:14, 10 August 2009 (UTC)
OK, so this has been inactive for a looooong time. I'll get Quarenon to post here on the status of the bot. It seems Quarenon's second proposal had no opposition so bot should go ahead. And are we all for adding the above code to MediaWiki:Common.css or wherever we're supposed to add it; that ==Category== really pisses me off. Cheers, Chicken7 >talk 12:12, December 9, 2009 (UTC)
I'll go ahead and implement that change now that I have access to the site CSS file. There hasn't been much participation in this thread about the bot but since I have more time in the next few weeks I'll start working on the coding part Lol --Quarenon  Talk 22:14, December 9, 2009 (UTC)

If bytes 310-31B (Hexadecimal) = encoded NETSCAPE2.0 then it is an animated file, per wikipedia:Gif#Animated_.gif Hello71 21:14, December 12, 2009 (UTC)

Cleanup section

Support - Per all @ above. --Fruit.Smoothie 05:06, December 10, 2009 (UTC)

Oppose - All these can be done by users. We shouldn't create bots for everything. Bots should only be used for stuff which normal users can't or find it too much of a hassle to do. --Nup(T) 01:19, December 15, 2009 (UTC)

Comment - Indeed, except for the category header, which is prevalent on a very large number of images. If we do ever decide to clean those up, I think we would certainly want a bot to do it. --Quarenon  Talk 08:57, December 15, 2009 (UTC)

Closed - QBot will begin tagging newly uploaded JPEGs as well as old images which may be missing the tag. Historical JPEGs and JPEGs that come from the RuneScape site (e.g. updates) will be excluded. No consensus achieved on the other points. --Quarenon  Talk 00:02, March 26, 2010 (UTC)