Commons:Village pump/Technical

Shortcuts: COM:VP/T • COM:VPT

	SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 1 day and sections whose most recent comment is older than 30 days.

Hi!

Is there a tool that calculates the cumulative amount of file sizes within a category?

Thanks :) --PantheraLeo1359531 😺 (talk) 16:56, 13 March 2025 (UTC)[reply]

Would be best if it also allowed seeing the largest files in the category. Maybe one could use a query with deepcategory (relevant question here). Prototyperspective (talk) 00:30, 14 March 2025 (UTC)[reply]

What would be the goal that had to be achieved with that information ? —TheDJ (talk • contribs) 11:16, 15 March 2025 (UTC)[reply]

For the file sizes:

– It allows to analyse cats if they have files that may be too large to use and can get a smaller file version. It can be analyzed, how the values have changed and, for example, if datasets by (e.g.) US governments have been copied (sometimes, the number of files isn't a good indicator). --PantheraLeo1359531 😺 (talk) 10:33, 17 March 2025 (UTC)[reply]

@PantheraLeo1359531 try User:RoyZuo/FilesizeSum.js? :) RoyZuo (talk) 12:33, 4 April 2025 (UTC)[reply]

Thank you a lot :D --PantheraLeo1359531 😺 (talk) 12:41, 4 April 2025 (UTC)[reply]

@RoyZuo How do I execute this script on a category? I am a noob at coding :D --PantheraLeo1359531 😺 (talk) 12:47, 4 April 2025 (UTC)[reply]

mw.loader.load('/w/index.php?title=User:RoyZuo/FilesizeSum.js&action=raw&ctype=text/javascript');

add to line 2 of User:PantheraLeo1359531/common.js. RoyZuo (talk) 12:49, 4 April 2025 (UTC)[reply]

Thanks, I figured it out. Splended thing! --PantheraLeo1359531 😺 (talk) 13:08, 4 April 2025 (UTC)[reply]

@RoyZuo It seems some calculation errors appear, like here: Category:Black caps in Japan. It claims to contain 98 MB, but it has approx. 16 MB, or am I doing something wrong? --PantheraLeo1359531 😺 (talk) 13:19, 4 April 2025 (UTC)[reply]

"On this page: 15.34 MB (4 files)" for me. RoyZuo (talk) 14:10, 4 April 2025 (UTC)[reply]

Oh okay, I will look into it :) --PantheraLeo1359531 😺 (talk) 16:36, 4 April 2025 (UTC)[reply]

Hi everybody. I have a question about this {{Wikidata Infobox}}. I makes sense to create a category including this code <!-- {{Wikidata Infobox}} -->. or, after, comment this code {{Wikidata Infobox}} when the wikidata is not existing yet or is not yet connect? An example here Category:Companies_based_in_Heemstede,_North_Holland. I don't want to start an editwar, but this kind of edit seem to not very wise, because when the wikidata element will be created it will be more difficult to connect it or, when already present, is not easy to find it, for example Category:Klouwer (surname), it was better to look for it instead of comment it <!-- {{Wikidata Infobox}} -->. Thank you in advance for your answers. P.S. If not necessary, I would not insert at all {{Wikidata Infobox}} MrKeefeJohn (talk) 18:40, 22 March 2025 (UTC)[reply]

Create your category with {{Wikidata Infobox}} but without the HTML comment markup. The infobox is designed for such use. I have just done this on your example category. Note that the infobox there includes a "Create new Wikidata item" link, Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:01, 22 March 2025 (UTC)[reply]

@Pigsonthewing: thank you for your answer. I understand what is designed for <!-- {{Wikidata Infobox}} --> and always add it creating new categories. The question was more about if wise to remove it when the wikidata element is not created yet. MrKeefeJohn (talk) 06:58, 3 April 2025 (UTC)[reply]

Usually a Wikidata infobox is not added when there is no Wikidata item to connect it to / that it's connected to. Adding an infobox nevertheless is I think an implicit suggestion/request for others to create a Wikidata item for the subject of the category. For these, I usually just create the Wikidata item but there are exceptions when it's not clear how the item would look like. In regards to your case (example?) I don't think there are Wikidata items for 'Companies based in {region}' but maybe I'm wrong or these could be created – however usually items are created for companies with some location information so there's no Wikidata item for that. They could maybe be created as dynamic list pages somehow though. Prototyperspective (talk) 09:51, 3 April 2025 (UTC)[reply]

That is the context in which I gave my answer. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:29, 3 April 2025 (UTC)[reply]

@Prototyperspective: I do understand your opinion, but in my experience quite often it's just a problem of missing connection. Just todady a couple of example Category:Companies based in Udine wikidata and Category:Companies based in Udine wikidata MrKeefeJohn (talk) 19:30, 11 April 2025 (UTC)[reply]

Don't know what you mean. This isn't what you wrote about earlier. Yes, there are categories that are not connected to their Wikidata items. You could for example always add the infobox and then check if the Wikidata item exists via search and if not create it or if not just leave the empty infobox as is. Prototyperspective (talk) 21:35, 11 April 2025 (UTC)[reply]

I mean that if you don't add a {{Wikidata Infobox}}, or remove it or comment like this <!-- {{Wikidata Infobox}} -->, it will be more difficult to connect with wikidata elements and/or with already existing articles or categories in some wiki. And it seems to me that it makes more difficult the bots' operation. MrKeefeJohn (talk) 06:56, 12 April 2025 (UTC)[reply]

Why would that be the case, I don't think so. That infobox template is added by a bot. Prototyperspective (talk) 10:16, 12 April 2025 (UTC)[reply]

Hello, I just observed something that's puzzling me. I thought about moving File:Claire Leger speaks to media about the Tarnak Farms incident.jpg on the rationale COM:FR#FR2, as the name "Léger" contains a spelling mistake (it's currently lacking the acute). I'm using a Chrome browser on my Android tablet to write this. If I switch to the Commons desktop mode to use "move and replace", the checkbox "leave a redirect behind" is active and greyed out, which is the usual behaviour on files in use that I'm accustomed to. But if I'm switching to the mobile view and access https://commons.m.wikimedia.org/wiki/Special:MovePage/File:Claire_Leger_speaks_to_media_about_the_Tarnak_Farms_incident.jpg through a 3-dot-menu on the file page, then I would be able to suppress the redirect anyway. Why is there such a difference between the desktop gadget and the special page function in the mobile view? Regards, Grand-Duc (talk) 02:55, 24 March 2025 (UTC)[reply]

Probably because no one ever looked at making the Commons-specific utility that is used in the desktop mode work for mobile as well. —TheDJ (talk • contribs) 16:45, 25 March 2025 (UTC)[reply]

if you start the move and replace gadget in the file history tab of the file on pc, you will also have the option of suppressing redirects. :)

i guess the gadget's logic is checking the "global usage" section on a file page. so when it cannot see that section, it assumes there's no global usage = suppress is possible. RoyZuo (talk) 08:54, 3 April 2025 (UTC)[reply]

I added a category with HotCat here and one with Cat-a-lot here and for some reason the categories were added inside the brackets of the crop-template that was already there as if the categories were part of the template syntax. As a result, when another user cropped the image with CropTool the categories were (automatically) removed together with the crop-template, see here. That looks like a chain of things going wrong. Nakonana (talk) 18:20, 1 April 2025 (UTC)[reply]

Several images are affected: see File talk:Bundesarchiv Bild 101I-133-0703-38, Polen, Ghetto Litzmannstadt, Kinder.jpg#Categories (and that list might not be comprehensive). Nakonana (talk) 18:27, 1 April 2025 (UTC)[reply]

I suppose the tools' logic is to find the last [[Category:...]] and then open a new line after it and add the new cats.

The logic makes sense, since it seeks to put categories together.

I could imagine a little more could be added to the logic. It can try to find the last \n\s*\[\[Category\:.+\]\]\s* instead. RoyZuo (talk) 18:43, 1 April 2025 (UTC)[reply]

I think that it would work that function find_insertionpoint() in Gadget-Cat-a-lot.js line 916 would contain also following replace when it is cleaning the wikicode:

.replace( /\{\{Crop\|category=(\s|\S)*?\}\}/g, replaceByBlanks );

I think that HotCat is using similar cleanup.

-- Zache (talk) 18:58, 1 April 2025 (UTC)[reply]

Thanks for your ideas. I have one suggestion: there could be one or more blanks between | and category= that needs to be considered. Kind regards, --Sebastian Wallroth (talk) Sebastian Wallroth (talk) 06:32, 2 April 2025 (UTC)[reply]

The much simpler solution would be to move the category link inside the template and only write the category name when using the template. GPSLeo (talk) 08:41, 2 April 2025 (UTC)[reply]

Whom/Where to ask to make that happen? Nakonana (talk) 08:38, 6 April 2025 (UTC)[reply]

I changed the template[1]. GPSLeo (talk) 09:05, 6 April 2025 (UTC)[reply]

Thank you! Nakonana (talk) 09:06, 6 April 2025 (UTC)[reply]

Is this solved then? Prototyperspective (talk) 11:35, 6 April 2025 (UTC)[reply]

I think? I haven't tested how CropTool is handling the change, though. Nakonana (talk) 18:00, 7 April 2025 (UTC)[reply]

Hello, maybe it is just me but I keep getting an error every time I tried to open any page.

• “MediaWiki internal error”…”Fatal exception of type "Wikimedia\Rdbms\DBUnexpectedError"

The error will go away if I refresh couple times, but it keeps happening when I tried to click on another page. It has been going on for a while now. Anyone know what this error is about? Tvpuppy (talk) 05:08, 2 April 2025 (UTC)[reply]

i'm getting the same for 2-3 days now. bugs have been filed https://phabricator.wikimedia.org/T389734 https://phabricator.wikimedia.org/T390811 . RoyZuo (talk) 05:43, 2 April 2025 (UTC)[reply]

Ah so it wasn’t just me. Thanks for letting me know. Tvpuppy (talk) 05:55, 2 April 2025 (UTC)[reply]

Hi!

I would like to create a table where numbers are displayed in the second row. These numbers should represent the number of files in a category including their subcats. Which template can I use to do this?

Thank you --PantheraLeo1359531 😺 (talk) 14:09, 3 April 2025 (UTC)[reply]

Is there a tutorial on how to create GeoJSON maps like this. I was able to download QGIS and add the QuickOSM plugin. I ran a query for all roads in Reykjavík and filtered the layer to only show the name of the street. Then I selected 'make permanent'. But the file I got was not usable for making GeoJSON map on Commons. Steinninn ♨ 11:52, 4 April 2025 (UTC)[reply]

If there is, please add it to (somewhat-new) Category:Wikimedia projects and maps. Prototyperspective (talk) 13:16, 4 April 2025 (UTC)[reply]

When I use the mobile uploader, usually it succeeds in picking names for my files that don't overwrite each other (for example Category:Views from the Internationale chairlift). With File:Blewett Pass - Tronsen Ridge ski tour.jpg, they were just uploaded on top of each other. Weirdly I had just done a previous multi-upload for the name File:Alpental.jpg that worked perfectly. What's the easiest way to fix this? Buidhe (talk) 03:37, 5 April 2025 (UTC)[reply]

@Buidhe I think this is a serious bug that you should write to https://github.com/commons-app/apps-android-commons , or follow Commons:Mobile app#Issues. RoyZuo (talk) 09:40, 6 April 2025 (UTC)[reply]

Hello! I have submitted a request to rotate File:ATTRITOR GRINDING MACHINE - NARA - 17496809.jpg by 180° via @SteinsplitterBot. Now the thumbnail has the desired orientation, but the actual image still remains upside down. What is the correct procedure to handle such cases? --Aciarium (talk) 10:49, 7 April 2025 (UTC)[reply]

@Aciarium: it looks ok for me. See COM:PURGE. Commander Keane (talk) 12:07, 7 April 2025 (UTC)[reply]

@Commander Keane: My bad, I wasn't aware that I had to purge the site after this action. Now it looks ok also on my side. Thanks for clarifying!! --Aciarium (talk) 13:49, 7 April 2025 (UTC)[reply]

Does someone have the means to extract the image from the British Museum's website? My usual methods are failing. JayCubby (talk) 14:49, 7 April 2025 (UTC)[reply]

@JayCubby https://www.britishmuseum.org/collection/image/1613085647 maybe? RoyZuo (talk) 08:36, 12 April 2025 (UTC)[reply]

MediaWiki message delivery 18:48, 7 April 2025 (UTC)[reply]

When I use the upload tool, PDF is not one of the options, and it does not recognize a PDF as a file type when I try to upload. When I fill in the information, the tool fails with nonsensical errors. This is especially problematic as Foundation policy directs that PDFs (which are heavily used on English Wikisource and other language Wikisource projects) which are free by Commons rules should be uploaded here and not locally. TE(æ)A,ea. (talk) 01:30, 9 April 2025 (UTC)[reply]

@TE(æ)A,ea. which tool do you mean? can you plz write down the link? RoyZuo (talk) 02:37, 9 April 2025 (UTC)[reply]

Hello @TE(æ)A,ea., I assumed you have found a solution to your problem since you uploaded File:Japanese.pdf after you posted here. Tvpuppy (talk) 02:47, 9 April 2025 (UTC)[reply]

• RoyZuo: It’s the default upload tool, Special:Upload. Tvpuppy: I used the upload wizard to upload the file from my computer; I generally avoid this, as the wizard is an inferior upload tool. As I would like to avoid using the wizard in the future, this is still an issue. TE(æ)A,ea. (talk) 03:41, 9 April 2025 (UTC)[reply]

  @TE(æ)A,ea. Special:Upload only up to 100 MB. RoyZuo (talk) 05:07, 9 April 2025 (UTC)[reply]

  it's clearly written "Maximum file size: 100 MB" right under where you pick you file. ¯\_(ツ)_/¯ RoyZuo (talk) 05:09, 9 April 2025 (UTC)[reply]

• Do you guys just not assume good faith here? RoyZuo: Even if the file size is too large, it still recognizes the file type; and, there is a separate option (on the default Upload tool) for uploading from a whitelisted (or allowlisted, whatever that means) Web-site, and that option doesn’t have a 100 MB limit. I had used that option, but it didn’t recognize PDF as a file type and never actually uploaded. TE(æ)A,ea. (talk) 12:31, 9 April 2025 (UTC)[reply]

  @TE(æ)A,ea. i just uploaded File:Walksincitycant00graygoog text.pdf by upload from url.

  when you dont give any useful info about the bugs, and make other users second guess the problem, the problem wont be solved. RoyZuo (talk) 12:50, 9 April 2025 (UTC)[reply]

on Special:Upload, there's a line "Permitted file types: tiff, tif, png, gif..."

after or around the page finishes loading, something deletes ", pdf" from the list of extensions. sorry my skill is too low to find out what does that. RoyZuo (talk) 08:16, 9 April 2025 (UTC)[reply]

The relevant code, and an explanation, is at MediaWiki:Upload.js#L-141--L-148. Omphalographer (talk) 00:14, 10 April 2025 (UTC)[reply]

I uploaded a video a while ago File:TTC Line 2 T-1 Subway arrives at Pape station (2023-08-08) 001.webm, it had some frame rate issues and never got around to fixing it, until now. However I am unable to re-upload it as the file exceeds 100MB. How can I overwrite the video file with the new one? PascalHD (talk) 02:02, 10 April 2025 (UTC)[reply]

@PascalHD User talk:Rillke/bigChunkedUpload.js. RoyZuo (talk) 06:36, 10 April 2025 (UTC)[reply]

@RoyZuo Thanks! It worked perfectly! PascalHD (talk) 21:38, 10 April 2025 (UTC)[reply]

This is the reason why this issue exists: phab:T382814 Missing info on how to upload a file larger than 100 MB in the 'Upload a new version of this file' page – You're not the only one who doesn't know that. Prototyperspective (talk) 21:52, 10 April 2025 (UTC)[reply]

I have here files I ripped from the SLSA's website (Category:Files_from_the_State_Library_of_South_Australia), which are higher-resolution than the existing files.

The files are named as follows: stitched_John_L_Bonython_12yrs_1887-B7723_43.jpg.jpg corresponds to file:John_L_Bonython_12yrs_1887-B7723_43.jpg

The files are here: https://files.catbox.moe/nmqlbs.zip

Thanks a bunch JayCubby (talk) 18:39, 12 April 2025 (UTC)[reply]

I have uploaded them  ${\displaystyle 999}$REAL 💬 ⬆   19:31, 12 April 2025 (UTC)[reply]

Thank you very much, 999real! JayCubby (talk) 21:26, 12 April 2025 (UTC)[reply]

Can somebody please take a look at my edit request for {{Created with}}? — Mikhail Ryazanov (talk) 21:09, 13 April 2025 (UTC)[reply]

MediaWiki message delivery 00:20, 15 April 2025 (UTC)[reply]

Does somebody know how Category:Media missing infobox template gets added to files?

Because it seems to be missing for quite a few files, such as this. For various structured access to files such as via the Commons app, the MediaViewer, an API, etc, having a standardized populated information template is very useful. It could (and to some extent) can also be searched / queried and metadata be gathered via the standardized format.

If whatever added that category and its subcats is not running anymore it would be good if it was restarted or if it's running improved so that more files are added to it.

Additionally, is there any effort to add the {{Information}} to files at scale? The cat currently contains as many as 334,529 files. For most of these, some script could add the template and add the source, author and description data from whatever has been added instead if there is anything. Prototyperspective (talk) 19:50, 17 April 2025 (UTC)[reply]

Looking at AbuseFilter 153 - Cross-wiki upload filter the way I understand it is that it prevents uploads of files matching that specific edit summary that are not images, or they are pngs, or jpegs less than 2Mpix or under 60kB.

But the bit (user_age < 3600*24*180 | user_editcount < 50) indicates that uploaders must be at least 180 days old (or have at least 50 edits)? (mw doc page says user_age is in seconds). Unless they are confirmed or autopatrolled. I am guessing that the Commons abuse filter can't see the user_age and user_edits in outside projects and that is why there was a propsal to block new users from uploading (that meant Wikipedia non-autoconfirmed right?). But if that user_age is applied to an uploader's Commons age and edit count, how do ~500 edits per day get through? For example File:Guillotina de 1789.jpg's uploader wasn't 180 days old anywhere and didn't have 50 edits (global account info). I must be misunderstanding something.

I will ping @Billinghurst as they last edited the filter. Commander Keane (talk) 08:47, 19 April 2025 (UTC)[reply]

@Commander Keane So I looked into this and realized that the statement !("confirmed" in user_groups | "autopatrol" in user_rights) also matches autoconfirmed users, see https://phabricator.wikimedia.org/T181024. This means that the filter has de-facto been exempted for user accounts more than 4 days old, which is unfortunate because it means the filter has probably been missing a very large number of files. This is a good catch and we should change this immediately, however given that the filter has been like this since 2016 it might result in a large increase in filter hits (maybe complaints from other wikis). Dylsss (talk) 23:54, 19 April 2025 (UTC)[reply]

How to mass rename/move (files)/delete (redirects, where files are already've been renamed) for IP ~400 objects at least changing obvious email (that is COM:ADVERT) "Media.Krysiak@gmail.com" to "Klub Lewicy" (according to flickr account name it uploaded from) as it's too hard to template it one by one, you know.

Also - how is possible something like deleting a revision comment meant here and here? 203.57.51.188 16:39, 21 April 2025 (UTC)[reply]

MediaWiki message delivery 20:56, 21 April 2025 (UTC)[reply]

User talk:Basque mapping#How is the pixel size your Flickr photo upload to Commons at (4,438 × 6,101 pixels), larger than the maximum size on this Flickr account?

Can someone explain how photo uploads grow in size from a source that has a smaller size. Please, see the link above and it's associated links to help me understand how these changes occur.

• My earlier upload from the exact same Flickr page is equal too the maximum size file available from the United States Department of State Flickr account here: [9] https://www.flickr.com/photos/statephotos/54295399868/

Thank you, -- Ooligan (talk) 03:02, 22 April 2025 (UTC)[reply]

@Ooligan: The history of File:Official portrait of Secretary Marco Rubio.jpg says that it came from https://www.state.gov/wp-content/uploads/2025/01/Official-Photo-Secrertary-Marco-Rubio.jpg which now appears to have been moved to https://www.state.gov/wp-content/uploads/2025/01/Marco-Rubio.jpg and which is 4438x6101 and matches the file that was uploaded. So I guess linking it to Flickr was just a convenience, as it's perhaps considered more stable than a WordPress attachment, and is the same file. Sam Wilson 07:44, 22 April 2025 (UTC)[reply]

given a random pdf here, how to check if it contains a text layer (that has a lot of text)?

i found https://www.reddit.com/r/pdf/comments/npz1ak/how_to_identify_which_pdfs_contain_a_text_layer/ . not sure if something already does that for commons files. RoyZuo (talk) 17:20, 22 April 2025 (UTC)[reply]

the practical use of this for me is, for a certain book category, there're a dozen different pdfs inside.

now i wanna know which one of them might already have the text. if some of them do, i can extract that and feed into wikisource directly, instead of wasting my time to ocr page by page. RoyZuo (talk) 17:23, 22 April 2025 (UTC)[reply]

You can use the API, it will return the text layer contents of a pdf. For example: https://commons.wikimedia.org/w/api.php?titles=File:Serial_Programming.pdf&action=query&prop=imageinfo&iiprop=metadata Bawolff (talk) 20:08, 22 April 2025 (UTC)[reply]

Should we not run a bot, to categorise as, say, "PDFs with text layer" and "PDFs with no text layer"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:17, 27 April 2025 (UTC)[reply]

I am using AutoWikiBrowser version 6.4.0.0 at Commons. I've found the software is almost useless because I cannot input the "replace with" field in the advanced settings.
Options --> Find and replace --> Advanced settings --> Find --> Replace with:
Find field can be inputted, but Replace with field cannot be. Please fix it as soon as possible. Thank you. --トトト (talk) 05:01, 24 April 2025 (UTC)[reply]

• It has been already reported at https://phabricator.wikimedia.org/T391609. I hope it will be fixed soon. --トトト (talk) 11:44, 27 April 2025 (UTC)[reply]

MediaWiki message delivery 19:28, 28 April 2025 (UTC)[reply]

I'm gonna try making a tool that goes from file revisions or previews back to file description page (in case you land on those direct links). does anyone know of any documentation on this? RoyZuo (talk) 12:02, 30 April 2025 (UTC)[reply]

Two digits from a hexadecimal MD5 hash and prefixed dates. See MW:Manual:MediaWiki file usage.

• File:Tartaric acid 2.svg
• https://upload.wikimedia.org/wikipedia/commons/0/06/Tartaric_acid_2.svg
• https://upload.wikimedia.org/wikipedia/commons/archive/0/06/20250419191555%21Tartaric_acid_2.svg

previews have similar information

• https://upload.wikimedia.org/wikipedia/commons/thumb/0/06/Tartaric_acid_2.svg/525px-Tartaric_acid_2.svg.png?20250419191557

The 525px-... states the image size. SVG files may have lang prefixes (this one is made up for German as spoken in Switzerland):

• https://upload.wikimedia.org/wikipedia/commons/thumb/0/06/Tartaric_acid_2.svg/langde-ch-525px-Tartaric_acid_2.svg.png?20250419191557

Of course the lang and size would be easy to strip.

Glrx (talk) 18:57, 30 April 2025 (UTC)[reply]