Commons:Village pump/Technical

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

Shortcuts: COM:VP/T • COM:VPT

Welcome to the Village pump technical section
Technical discussion
Village pump/Technical
 Bug reports
 Code review
Tools
 Tools/Directory
 Idea Lab



This page is used for technical questions relating to the tools, gadgets, or other technical issues about Commons; it is distinguished from the main Village pump, which handles community-wide discussion of all kinds. The page may also be used to advertise significant discussions taking place elsewhere, such as on the talk page of a Commons policy. Recent sections with no replies for 30 days and sections tagged with {{Section resolved|1=--~~~~}} may be archived; for old discussions, see the archives; recent archives: /Archive/2024/10 /Archive/2024/11.

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

How to search fields of files' Information template?

[edit]

For example, how could one search file description for a term like "Kathmandu" (as asked about by another user here) like on can search with intitle. description:"Kathmandu" does show some results but I don't know what it does and the results don't have that word in their description. I could not find info on this at mw:Help:CirrusSearch either. Info how to search specified fields of {{Information}} should be added here.

One could also use this to infer categories (such as by reading the date field and then adding it to a category by date like "Videos of {year}") as proposed here. This may also be needed for a date range filter, see phab:T329961. I'd like to search the date field but there is no information on how to do that at Help:Searching but I think it's already possible if I remember correctly. For example, I found that many files in deepcategory:"NASA videos from unidentified year" deepcategory:"Videos of 2020‎" have been miscategorized into Videos of 2020 (and thus should not be copied into "NASA videos in 2020" from there) where they have the correct date in the date field which is why I'd like to use that to correct that as well as copy them to their year category in Category:Videos from NASA by year. Prototyperspective (talk) 19:03, 6 October 2024 (UTC)[reply]

Prototyperspective (talk) 12:16, 10 October 2024 (UTC)[reply]
Here EBernhardson (WMF) said Unfortunately, the image description is simply an argument to a template. CirrusSearch doesn't do anything at that level and can't be that specific..
I think the best workaround currently would be to use the insource search operator with the field name first so for example I searched for insource:"|source=[https://soundcloud.com to identify files for Category:Audio files from Soundcloud.com. I think easily searching fields of the File pages' Information template could be enabled by
  1. Developing some regex that searches for any content after e.g. |source=
  2. Creating some alias for it so instead of writing some complex regex query every time one can simply enter e.g. info-source:"soundcloud.com"
Please comment what you think about this proposed way to make this possible and if you have any info on what would be needed for that. Prototyperspective (talk) 13:06, 23 October 2024 (UTC)[reply]
[edit]

Example: An edit to a file changed its [en] caption to "28th September 2024 Sunday", but the preview in its diff and permalink displays not that but another text: "An introduction to persistent identifiers as part of a FAIR data landscape", which is the caption from the latest revision (made in diff). Previews in diffs and permalinks should display their specified revision, not the latest. -- Wotheina (talk) 05:05, 7 October 2024 (UTC)[reply]

My guess is that it shows always the latest caption? Another note, the caption box is not visible with OS X and Safari browser when I am looking the old revisions of the file pages. --Zache (talk) 07:07, 7 October 2024 (UTC)[reply]
I think Wotheina's bug has confused me before too. Is there a Phabricator: task for it? I did a quick search but I am not the best at finding bugs there. Also the Safari bug should be filed too. Commander Keane (talk) 21:45, 23 October 2024 (UTC)[reply]

Templates in description field cause empty caption in Media Viewer

[edit]

Putting some templates inside description= field causes Media Viewer to display contents of the template (usually poorly able to describe what can be seen) instead of what is inside description=. See the problem in action: navigate to File:Bodegas de Mileștii Mici, Moldavia, 2023-11-02, DD 74.jpg, observe the nice description is there ('Wine cellars of Mileștii Mici'), click 'Open in Media Viewer' and you will be displayed only 'This file illustrates a cultural heritage monument in Moldova, ID' (additionally, ID will be cropped, but it's not that important).

Note the severity of the problem: every time a visitor clicks a photo on en.wikipedia.org, they execute the Media Viewer, and may see a confusing, broken caption like I gave as an example above. This is not the best User Experience. Another thing is that POTY voting page displays descriptions in the gallery of candidate pictures using the same logic.

What do you think? Derbeth talk 18:10, 13 October 2024 (UTC)[reply]

I also saw one more scenario, involving 'Captions' section. If you have a caption, and description= field is empty, Media Viewer correctly uses the caption. But I saw files where a caption was given but 'description' was just a template on cultural heritage. Or sometimes just {{Commonist}}. With such a form of 'description', Media Viewer replaces the useful caption with a generic text about cultural heritage, or even displays blank caption (probably in the case of commonist template, but I don't remember the exact situation). What should be the solution? Changing the page by copying caption to the 'description' field, alongside the template? Changing the page by moving the template outside of {{Information}}, for example just below it in the 'Summary' section? Changing Media Viewer to do some guessing and use 'Captions' instead of 'description' in case 'description' consists exclusively of some 'undesired' template, like cultural heritage, and nothing more (note how hard it's to implement when templates are created, removed and renamed)? --Derbeth talk 18:34, 13 October 2024 (UTC)[reply]

The same happens at Special:MediaSearch when selecting an image.
There was some discussion at Commons_talk:WMF_support_for_Commons/Upload_Wizard_Improvements about reducing clutter in "description=", but apparently, this is not a priority for WMF.
 ∞∞ Enhancing999 (talk) 19:20, 13 October 2024 (UTC)[reply]
Dear colleagues, after comparing various photos, I noticed that there are also monument templates that do not negatively affect the description in the Media Viewer and are displayed despite being placed in the summary field: e.g. 01, 02, 03 and 04. So this could be an indication that the cause lies in certain monument templates that need to be adjusted. But I could be wrong, it is just an indication that I noticed. Best regards, -- Radomianin (talk) 12:10, 14 October 2024 (UTC)[reply]
Addendum to the log: This discussion was preceded by a discussion on the colleague's user page. Best regards, -- Radomianin (talk) 09:33, 17 October 2024 (UTC)[reply]
One has "noprint noexcerpt" as classes, maybe one of them helps?
 ∞∞ Enhancing999 (talk) 14:20, 14 October 2024 (UTC)[reply]
I just checked that those classes alone don't help. I edited a template, checked File:Lodochnaya_stancia.jpg again, and both CSS classes appear in the web inspector, but the caption in MediaViewer is still the wrong one. Should be 'Boating station in Ukrainian city Kriviy Rig' (taken from 'Captions'), but is 'This is a photo of a natural heritage site in Ukraine' (template in 'description='). --Derbeth talk 08:22, 17 October 2024 (UTC)[reply]
I tried those classes too, but it doesn't seem to be sufficent. Maybe @TheDJ has further details on how it's generated.
 ∞∞ Enhancing999 (talk) 15:49, 18 October 2024 (UTC)[reply]
"Mediaviewer replaces it" Just a note, Mediaviewer does not such thing, the MetaData api does this. Of which Mediaviewer is simply the most prominent consumer. —TheDJ (talkcontribs) 12:12, 14 October 2024 (UTC)[reply]
Thanks for the explanation TheDJ. Would you like to help with reporting the bug to MediaWiki API in a way that gives the best chance it is picked for development? --Derbeth talk 08:22, 17 October 2024 (UTC)[reply]

CategoryTree arrows

[edit]

Since yesterday, I see all CategoryTree arrows in grey, even for categories with subcategories. They are still clickable, but they appear to be grey. Only blue arrows that remain are those with sub-sub-categories. How can it be fixed? Is it this issue? — Draceane talkcontrib. 07:48, 18 October 2024 (UTC)[reply]

Do you still have this problem? If so please link an example. I see blue expandable arrows at categories with subcategories. Prototyperspective (talk) 10:51, 18 October 2024 (UTC)[reply]
Ok, it's probably just me... Safemode works well. Could be connected to the recent Cat-a-lot update? — Draceane talkcontrib. 11:15, 18 October 2024 (UTC)[reply]
[edit]

I am working on the maintenance category Category:Cultural heritage monuments in Berlin without linked Wikidata and noticed that even after purging and null edit to category page and template the category members are not updated many hours after they were linked to Wikidata. Are there any ideas on the reason and suggestions how to solve this? GPSLeo (talk) 08:21, 18 October 2024 (UTC)[reply]

Normally null edits to category page connected to the Wikidata item work. In general, it's the issue discussed at User_talk:Mike_Peel#Infoboxes_missing_(2024-10).
 ∞∞ Enhancing999 (talk) 18:30, 18 October 2024 (UTC)[reply]

W3C validator has been basically down for months

[edit]

Where I and these folks are from, we cannot use the link provided by {{Valid SVG}}. The validator always says "429: Too many requests". The page loads, but the "external validation service" doesn't. Aaron Liu (talk) 20:47, 19 October 2024 (UTC)[reply]

The link may fail for many reasons. I'm often run into the problem of unable to validate because the DOCTYPE is not known; I just select SVG 1.1 and ignore the minor errors.
Chealer reports failure for File:Netscape_icon.svg. That file did not have {{Valid SVG}} or {{Igen}}, so I added them and {{Invalid SVG}}.
Igen created a working URL that said the file was valid
{{Valid SVG}} gave the message that validity was not checked. That suggests a broken template.
{{Invalid SVG}} said the file was invalid (that's its job), but the URL said it was valid.
That link is the same as above. However, a second click complained "Sorry! This document cannot be checked." The problem was the external checker was not available.
Note: I can do a control-Reload and sometimes get a good result.
The previous succeeding checks used SVG 1.1 + xHTML + MathML, but the failing clicks did not recognize a Document type. However, I could select "SVG 1.1" and revalidate resulting in the URL
The new URL reports 46 errors and 1 warning. The inkscape and rdf namespaces are not treated as warnings.
Somebody may be limiting the requests or there may be a race condition that the Nu validator does not respond quickly enough.
First, I think it is poor practice for the templates to use uri=https%3A%2F%2Fcommons.wikimedia.org%2Fwiki%2FSpecial%3AFilepath%2FNetscape_icon.svg.
The more direct method would use
But that validation fails for files without a DOCTYPE. Supply an explicit SVG 1.1:
Which, of course, does not recognize inkscape or rdf namespaces because they are not in the SVG 1.1 DTD and reports dozens of errors.
Perhaps the doctype parameter was used to encourage switching to the Nu validator. Here's the URL without the parameter.
That URL fails with the external checker not being available.
Second, the problem seems to be the validator wants to hand off the request to the Nu validator. It may do this when the SVG file does not have a DOCTYPE.
The GitHub issue used the W3C Nu validator, https://validator.w3.org/nu/ , directly. The Special URL was not an issue.
Instead of using that, go directly to Nu. Chealer offers
That old version of the file has a DOCTYPE, so the old W3C validator does a reasonable validation:
The current version of File:Flag of Cuba.svg does not have an XML processing instruction or a DOCTYPE. Force SVG 1.1
Conceivably, SVG 1.1 could be forced all the time, but that raises numerous inkscape and rdf errors.
Conclusion. I would use the Nu validator without the schema parameter.
Glrx (talk) 17:36, 22 October 2024 (UTC)[reply]

Tech News: 2024-43

[edit]

MediaWiki message delivery 20:48, 21 October 2024 (UTC)[reply]

This just materialized on a file I was editing, so I quickly turned the red link blue and found that it's popped up on 800+ files. What is this? Why is it here? —Justin (koavf)TCM 16:05, 24 October 2024 (UTC)[reply]

I noticed this too. I got as far as figuring out that {{#invoke:ISOdate|ISOdate|1=2006-07-24}} adds the category. I'm not too versed in Lua modules, so can't really continue the hunt. Cryptic-waveform (talk) 19:02, 24 October 2024 (UTC)[reply]
And so does {{#invoke:DateI18n|Date|year=2000|month=1|day=1}}. Cryptic-waveform (talk) 19:06, 24 October 2024 (UTC)[reply]
May be related to https://github.com/wikimedia/mediawiki-extensions-JsonConfig/commit/db05e3251393d734605d20c98a63dbd7f93d9426. Cryptic-waveform (talk) 19:13, 24 October 2024 (UTC)[reply]
@Cscott: Is this related to your change? And is it a desired outcome? Cryptic-waveform (talk) 19:15, 24 October 2024 (UTC)[reply]
@Koavf, @Cryptic-waveform & @Cscott - FYI, I used the CropTool today and the extracted photo file has the this same hidden category as the the original photo file here: File:Zheng Jianbang in 2023 (cropped).jpg.
  • 1. Does this category still apply to cropped images extracted from files within this new category?
  • 2. Was a there any "notice" or other advance communication about this small, but highly noticable "hidden" category? -- Ooligan (talk) 22:11, 24 October 2024 (UTC)[reply]
  1. I guess? ¯\_(ツ)_/¯
  2. No.
Justin (koavf)TCM 22:34, 24 October 2024 (UTC)[reply]

Related: https://en.wikipedia.org/wiki/Wikipedia:Categories_for_discussion/Log/2024_October_24#Category:Pages_using_the_JsonConfig_extension. Cryptic-waveform (talk) 19:21, 24 October 2024 (UTC)[reply]

See also mw:Extension:JsonConfig. No idea why this is something that needs to be tracked. —Justin (koavf)TCM 19:43, 24 October 2024 (UTC)[reply]

Requested fix

[edit]

Commons:Administrators'_noticeboard#Interface_admin_request. —Justin (koavf)TCM 22:38, 24 October 2024 (UTC)[reply]

The category is empty now. Good to delete? Regards, Aafi (talk) 13:46, 5 November 2024 (UTC)[reply]
Perhaps also good to keep a look at phab:T378352. Regards, Aafi (talk) 13:47, 5 November 2024 (UTC)[reply]
Emptying was predicted to complete at the end of december. File File:Chrysalis5504.jpg had Category:Pages using the JsonConfig extension and it was red, whereas the cat itself was empty. No change after "null edit". Fixed after real edit. Now nazi file File:Deutsches_Reichsgesetzblatt_33T1_126_0846.jpg has that cat (see above). Strange. Taylor 49 (talk) 19:29, 6 November 2024 (UTC)[reply]
File:Deutsches_Reichsgesetzblatt_33T1_126_0908.jpg pretends to be in "Category:Pages using the JsonConfig extension" (red and empty, should NOT be there) and in "Category:Nazi symbols" (OK), OTOH it does NOT show "Category:Nazi symbols status", but when looking into that cat, it is listed there. Commons is severely broken today. Taylor 49 (talk) 19:34, 6 November 2024 (UTC)[reply]
Taylor 49, last time I checked Nazi symbols status had approximately 27,000 files and presumably all of them had Category:Pages using the JsonConfig extension. Currently, there are 14,829 files and it looks like all of them have Category:Pages using the JsonConfig extension. I don't know exactly what happened to the rest of the files.
Deutsches Reichsgesetzblatt 33T1 126 0846.jpg and Deutsches Reichsgesetzblatt 33T1 126 0908.jpg are not longer in Category:Pages using the JsonConfig extension and are normally categorised. Is it safe to assume for the rest of the files? Also Category:Nazi symbols and Category:Nazi symbols status are also correctly behaving now. Ratekreel (talk) 04:17, 7 November 2024 (UTC)[reply]
A few files have "fixed themselves" since yesterday, but there is still the discrepancy in both directions:
It's ultimately a BUG, and a new one. The legacy flaw of all WMF wikis is that sometimes a file claims to be in an existing cat, but when looking on that cat the file is not there. What we can observe here is new and worse than that. Taylor 49 (talk) 14:40, 7 November 2024 (UTC)[reply]

Should Template:Compressed version be used with Lossy WebP?

[edit]

It's unclear to me what the role of Template:Archival version is now in 2024. Although it makes a lot of sense in a world where GIF, PNG, and JPEG were the "Big Three" commonly supported image formats on the web (ignoring BMP, which I think did enjoy an era of support most associated with the early web), it seems a bit technical whether or not the Template:Archival version & Template:Compressed version templates should still be used on new uploads today with the advent of WebP.

I'm not an expert, but I think modernly, Mediawiki is happy to take PNG and shunt it to WebP automatically, as a display format that saves bandwidth without conscious change of workflow on part of the user. The transparent nature of these bandwidth savings is less applicable to JPEG (which can also be automatically converted to WebP for bandwidth savings), but when that's done, the converted WebP still must deal with the limitations of the legacy JPEG format. Namely: lack of support for alpha transparency, and converting lossy-to-lossy (necessarily degrading the image, versus if the user had just embraced lossy WebP to begin with).

I'm new to Wikimedia, and trying to determine the best way to upload my images so that:

  1. Pages that include my images continue to load fast
  2. My images are stored at the highest resolution possible and in the best quality possible

To satisfy both of these goals, I thought I should hijack Template:Archival version and Template:Compressed version as an "existing structure" and seemingly "best practice" put into place during an era when it was expected that the Template:Compressed version would be a JPEG. Lending legitimacy to this, the Template has a (deprecated) field named "Filetype", implying that at least in the past, it was appropriate to use the template with more than just JPEG images.

So I went ahead and tried it:

However, there are currently a few problems with this.

  1. Although Template:Compressed version claims filetype can be omitted and then defaults to an "intelligent guess", it seems that in reality, it has not been able to detect usage of the template with WebP, and instead defaults to the text to "This image is a JPEG version [...]" when used with WebP. If this is a valid usecase of the template, the text can be changed to support detection of WebP easily enough.
  2. Although I hoped that by using WebP for the compressed thumbnail version instead of JPEG I would gain access to alpha transparency (improving dark mode support on Wikipedia), it seems that the thumbnail generated at w:Connection Machine filled the transparency with the colour white, which is especially unfortunate for this primarily black object where it is hard to discern detail against a white background. Although in this specific case, I may want to consider a permanently dark background to set this object against instead of transparency, it can be imagined that there are cases where it would be nice for thumbnails to adapt to Light Mode and Dark Mode versions of Wikipedia.
    I am told that the lack of support for transparent thumbnails of WebP images is caused by "Phab:T283646 which leads to this ImageMagick issue" but I wonder if it might also have been done on purpose, to avoid issues where people have uploaded transparent images in the era of Wikipedia not supporting dark mode, not expecting them to be set against anything other than a white background...
  3. I'm not completely sure that any of this makes sense for me to worry about, and perhaps it is actually fine to upload only a 60 MB PNG and let Mediawiki worry about bandwidth efficiency. I need to check in with people more familiar with the technology involved, which is what I'm doing right now here at the Village Pump.
I would note that Mediawiki doesn't seem to be doing a great job with my 60 MB PNG, still serving it as a 4 MB thumbnail image at least in File: namespace, while the WebP I uploaded is 2 MB at full resolution (barely discernible loss of quality), is converted to a 1.13 MB PNG for display in File: namespace, a 75% savings.
I'd note it is a little frustrating that it converted my WebP to PNG, but I suppose this was done to ensure compatibility. https://caniuse.com/webp claims WebP support is up to 95.69% among browsers that are currently used. The remaining 4.3% is still a sizable portion to support, especially in the context of Wikipedia, but it seems like it would at least be possible to read my useragent and determine that like almost all browsers released in the past 5 years, my browser supports WebP and can be served a smaller file.

In summary, please leave your thoughts on if using Lossy WebP in Template:Compressed version is valid and useful, or if there is some better technical way going forward, such as uploading only a Lossless WebP (which unlike PNG, does properly support photo metadata). —Hubcapp(talk) 18:37, 25 October 2024 (UTC)[reply]

1. {{Archival version}} is to indicate "don't mess with this, so that others can mess with the same reference at a later time". Now we could say that this is somewhat dated, as we already don't allow most people to change the version of an image, asking people to upload them separately.
2. Storing VERY high resolution images, means that the thumbnail has to take a huge leap from large to small. And it needs to do this with settings that are 'OK' for most of the images. This at times can result in suboptimal results. For that reason, it may be better to make your own slightly smaller version, applying self selected settings, and then uploading that smaller version of the file, for 'general use' in the Wikipedia, where often images get downscaled to something like 220px or smaller. However, in MOST cases, the images aren't used that widely and that critical that this is warranted, and it can always be done later.
3. Compressed version: This template indeed was not handling webp. I have updated its dependency to provide support for this.
4. In general, you should not worry about data and bandwidth usage. Where required and useful, as well as permitting time and money, the engineers will use newer technologies as they become ubiquitously available. In many cases having a simpler more 'wasteful' flow, makes up by being way faster and easier to manage and understand in the short term. (At the same time, this is not a license for uploading multi gigabyte files with little value and/or use). —TheDJ (talkcontribs) 15:44, 29 October 2024 (UTC)[reply]
Re 2.: There exists Template:No scaled down dupes, which implies it is improper to upload a lower resolution version of an image. Do you mean it's actually OK to upload a "smaller" version (in terms of resolution), or by "smaller" did you mean "more optimally compressed but still at full resolution"? And to clarify, you're saying that the Archival/Compressed versions dichotomy is in fact still useful for new uploads (but generally only for very high traffic images where it matters more)?
Re 3.: Thank you for clarifying it's OK to use WebP in Template:Compressed version and providing support for it!
Re 4.: Is it perhaps a good idea to upload *only* Lossless WebP? In my case, the resolution of my camera is roughly 6000 by 4000, so I'm not in multi-gig territory. However "Very Large Image" seems to be both a matter of opinion, and a definition that has changed as technology has advanced.
Thanks, —Hubcapp(talk) 02:34, 30 October 2024 (UTC)[reply]

Wikimaps Warper edits not working

[edit]

Somehow Special:OAuthListConsumers/view/d3308b00209d49e5a1d764aee019f83c stopped working for me. The tool sets the bounding box for maps depending on definitions of control points at https://warper.wmflabs.org . It's unclear who currently maintains the tool. @Chippyy: seems inactive.

For several, including Here, I had to complete the infobox fields manually. It doesn't seem to be a problem of the other content on file description pages, as this didn't get updated either.

It does work for @Kognos at [10] and a few other contributors [11].
 ∞∞ Enhancing999 (talk) 10:19, 27 October 2024 (UTC)[reply]

Also, it doesn't appear to depend on the absence of SDC statements: [12] failed too.
It doesn't seem to work for other users any more either: [13] by @Groupsixty.
 ∞∞ Enhancing999 (talk) 11:25, 27 October 2024 (UTC)[reply]
I'm afraid I have no idea why it's not working for others. I have had problems with the warper in the past, but it has been working OK, if sometimes a bit slow, recently. As you have noticed, I use the warper quite a lot. I set the template to Map, then click on the "Georeference..." button. If I am not currently logged on to the warper I get a logon page, select log on through Commons, then allow Oauth. I then georeference the map, and the bounds are automatically updated. I guess this is no different from what others are trying to do. I agree that the lack of any clear path to support is frustrating. I have been quite unclear in the past as to where any requests for help should be posted, and have had little success in getting support. The Wikimaps Warper category is useful, but there is still no indication as to who actually owns the project. Any enlightenment appreciated! Kognos (talk) 18:09, 28 October 2024 (UTC)[reply]
Maybe it's a browser issue, It used to work for me last year. The category is at Category:Wikimaps Warper.
 ∞∞ Enhancing999 (talk) 18:39, 28 October 2024 (UTC)[reply]
@Susannaanas do you know who is maintaining wikimaps warper? --Zache (talk) 18:27, 28 October 2024 (UTC)[reply]

Can category content be viewed by date?

[edit]

Hi. Category content is always sorted by alphabet as the default. This can be unconvenient when parts of series become scattered all over the list. Is it possible or could it be possible to overview the contents by date at the push of a button? This would be pretty neat. I will give two examples : 1 sorted by alphabet and 2 sorted by date, (more coherent). As you can see this link is very long and can be hard to put into a short working link (such as a text button). I would like (the option) to keep the layout ot example 1 in favor of listed as in 2, because it shows images next to eachother and gives overall a better overview per viewed segment of pages, like as if looking trough a category. This project idea would make the category contents available, to more people in a more appealing format, to have series more or less grouped by upload-date, and not scattered by alphabet. This would be more ergonomic and visually intuitive. Can this be done? Thanks. Peli (talk) 14:26, 28 October 2024 (UTC)[reply]

Currently if you enable Help:Gadget-DeepcatSearch it lets you 1 click to a deepcat search of the cat you are viewing. with a few more clicks you can get to a sorted-by-upload-date list.
i have considered expanding its functionality to include more direct links to oldest-first, newest-first... lists.
it's also not difficult to make some user js to make those links. RoyZuo (talk) 14:48, 28 October 2024 (UTC)[reply]
Honestly, having a one-click sort by date would be super appreciated. I could only wish to have the ability to make custom js. Huntster (t @ c) 17:46, 28 October 2024 (UTC)[reply]

Ah cool. But it would not be available to all visitors unless they install it? I wanted to do a post script of previous request. :: Earlier trials to install it on that page failed, but now the link is on the page. This fixes my issue, and keeps this as wanted extention to all cat pages if possible in any way. To me this sorting method means a incredible release of the pressure to rush to categorize everything (over 100k items) just to be able to overlook some parts of this material as more coherent series. Especially now the tool Cat-a-lot has become a much too slow tool for this scale projects. Thanks. I had that gadget enabled and just now explored the sort options, seems to work well to find other points of departure to explore this huge 'unsorted' category. Peli (talk) 14:26, 28 October 2024 (UTC)[reply]

  • I think it would be great if that gadget would be enabled by default. However, for that to happen the deepcategory problems (at least phab:T376440) would need to be solved.
  • The link currently shows items directly in the category as well as in subcategories – if you only want to see items directly in the category, change "deepcategory" to "incategory".
  • Another issue is that one can only sort by upload date, but not by date of what's in the media file (the date field in the file's {{Information}}) – see e.g. phab:T329961 and the things linked there. Cat-a-lot has recently been improved so should be faster again.
Prototyperspective (talk) 21:43, 28 October 2024 (UTC)[reply]
If you want to sort by date, ppl need to get on the CommonsData bandwagon, because that is the only way that will ever become a reality. —TheDJ (talkcontribs) 15:18, 29 October 2024 (UTC)[reply]
Disagree. What would be needed is to make use of the Information template – see #How to search fields of files' Information template?. I don't care whether the data from that template is 1. read and written to structured data and 2. kept in sync with structured data or whether it can be searched another way, I just disagree that this is required and also see an issue where this data would need to be written to over 100 million files and thus may not be feasible while one can already search data in the file information template via insource which could readily be improved upon. (If it doesn't become a reality otherwise, that is because of ideology and/or having invested much into a somewhat redundant techcomponent that makes it hard for some to see or consider, possibly–likely better, alternatives.) Prototyperspective (talk) 16:44, 29 October 2024 (UTC)[reply]
Is that (whatever it means or implies) really needed, just to find out the upload-date? This can be done already in special search. Pelikana (talk) 15:29, 29 October 2024 (UTC)[reply]
No, the comment was about the date the media was recorded. I think I overread "upload-date" and the title was unclear on whether you mean upload-date or date the file was taken but it's also relevant in general so I thought it was relevant enough to mention. Prototyperspective (talk) 16:50, 29 October 2024 (UTC)[reply]
@TheDJ do you mean sdc or...? RoyZuo (talk) 19:06, 1 November 2024 (UTC)[reply]
Ok. tx Any way to get around alphabetical order can be helpful in some huge unsorted categories. On smaller cats and sorted cats by crafted sortkeys the option would be needless and to be avoided. Maybe it could be just an optional piece of code to put in the header of the most relevant huge unsorted categories. I have another topic to ask attention for, it is on the [talk page of visual file change]. Open question and request about scrolling by keyboard. Peli (talk) 10:21, 31 October 2024 (UTC)[reply]
Requested more than a decade ago (phab:T71417) and triaged as "Low" priority. 😒 — Draceane talkcontrib. 20:16, 1 November 2024 (UTC)[reply]
The API also supports by date added to the category (Not neccessarily the same thing as date created) - https://commons.wikimedia.org/w/api.php?action=query&list=categorymembers&cmtitle=Category:Uncategorized_images_of_the_Rijksmuseum&cmsort=timestamp&cmlimit=max . In theory it would be easier to add support for sorting categories in that way than other forms of sort by "date". Someone would still have to do it though. Bawolff (talk) 06:31, 4 November 2024 (UTC)[reply]
That would also be very useful. Maybe that should be added to that linked phab issue? I don't think it would be as useful as sorting by date taken however. Prototyperspective (talk) 10:54, 4 November 2024 (UTC)[reply]

Upload Wizard very slow

[edit]

For the past few days, Upload Wizard has been very slow. Uploads that used to take secs now take mins. The behaviour I see: when I start uploads, it runs at 1-2 MB/sec for a few seconds. Then it slows down to < 1KB/sec for long periods. Intermittently, there is a burst of 1-2 MB/sec for a few seconds.
When I use chunked upload for a new version of a file, with chuck size 4 MB, the chunks upload fast, but then finishing and posting takes several minutes.
Others have reported similar issues @廣九直通車 and MPF: . Would appreciate any suggestions and help. Tagooty (talk) 14:51, 28 October 2024 (UTC)[reply]

I just uploaded 3 images with Upload Wizard, each with around 10MB size. It took me almost 9 minutes to upload, compose and queue the files, and after populating the files with descriptions and categories, it takes me another 3 minutes to finish the whole upload procedure. Moreover, during the first 9 minutes, I remember Upload Wizard reported errors with server connection. What's the problem? Should I send the matter to Phabricator?廣九直通車 (talk) 08:04, 29 October 2024 (UTC)[reply]
Dito, also have errors when uploading larger files --PantheraLeo1359531 😺 (talk) 08:08, 29 October 2024 (UTC)[reply]
Interestingly, I also tried direct upload from Flickr. Everything works perfectly fine in this case. Interesting...廣九直通車 (talk) 10:08, 29 October 2024 (UTC)[reply]
Today, I'm unable to upload any picture with Upload Wizard. I have opened a new thread on a help page about it. -- Jakubhal 17:33, 29 October 2024 (UTC)[reply]
Hi all, I would suggest you to open a Phabricator ticket about the problems you're having, and then kindly link it back to me, so that I could solicitate some action from the devs. Is it possible? Sannita (WMF) (talk) 18:43, 29 October 2024 (UTC)[reply]
I guess this is phab:T378276 Bawolff (talk) 02:50, 31 October 2024 (UTC)[reply]
Well, as of 31 October, it appears that the problem is largely solved. Just uploaded a total of 55MB of files, and the process is smooth. This thread may be kept in case things break again, however.廣九直通車 (talk) 08:29, 31 October 2024 (UTC)[reply]
Seems like the underlying cause is phab:T378385. Basically, the upload wizard (chunked upload) jobs share infrastructure with other jobs that only get triggered rarely. All of a sudden flagged revisions extension started making a lot of a type of job that normally does not get triggered very often. This meant that the UploadWizard jobs had to wait for those to complete first, slowing things down. It seems like that situation resolved, so upload wizard should be back to normal now. Bawolff (talk) 21:27, 31 October 2024 (UTC)[reply]
Sounds like weird planning when batch jobs meant to do stuff in off-hours to reduce load for day-to-day operations in one wiki sink regular operations in another wiki.
 ∞∞ Enhancing999 (talk) 07:13, 1 November 2024 (UTC)[reply]
Similiarly: phab:T378385#10282871
 ∞∞ Enhancing999 (talk) 22:33, 1 November 2024 (UTC)[reply]
@Bawolff Not sure if phab:T378276 is resolved and the root cause properly addressed as task phab:T379035 is titled "consider".
@Sannita (WMF): is there any progress on your side?
 ∞∞ Enhancing999 (talk) 12:53, 7 November 2024 (UTC)[reply]
The immediate issue is resolved - its no longer slow. The other task is a follow up to try and reduce the risk of a similar thing happening in the future. Bawolff (talk) 13:59, 7 November 2024 (UTC)[reply]
Well, it's no longer slow each time a user aborts an upload, but that doesn't mean the problem was resolved.
Given the current configuration that has Commons uploaders wait while MediaWiki performance maintenance at Wikipedia, it's likely to happen again. We don't want to open and close tickets merely because a user closed their browser.
 ∞∞ Enhancing999 (talk) 14:06, 7 November 2024 (UTC)[reply]
As a general principle, tasks are supposed to represent something to do. The phab:T379035 is supposed to represent the long term fix, where the other one is supposed to represent making things work right now & figuring out what happened. Since the long term fixes stuff is encapsulated in the other task, it makes no sense to keep the T378385 open since there is nothing to do on that task that isn't covered by T379035 [That said, the issue is possibly reoccurring right now]. Bawolff (talk) 10:24, 9 November 2024 (UTC)[reply]
phab:T378276 is about an existing, severe problem that needs to be solved.
phab:T379035 about an investigation that may or may not fix the problem. For casual readers, it might not even clear that is concerns Commons and that it could fix a severe issue at Commons.
 ∞∞ Enhancing999 (talk) 11:34, 9 November 2024 (UTC)[reply]
Thanks to those working on these issues. For the record, I recently posted to the Help Desk here: Help_desk#Upload Wizard broken for me. RobbieIanMorrison (talk) 20:11, 9 November 2024 (UTC)[reply]

Good news: phab:T379035#10318390 notes the deployment, so category updates at Wikipedia will no longer have Commons uploaders wait while uploading. Also, if I understood the fix correctly, uploads by url wont impact upload wizard performance any more. Thanks for everyone involved in finally investigating and fixing this.

It would be good to have links to the reports allowing to check those upload queues for saturation.
 ∞∞ Enhancing999 (talk) 14:04, 14 November 2024 (UTC)[reply]

Tech News: 2024-44

[edit]

MediaWiki message delivery 20:52, 28 October 2024 (UTC)[reply]

Russian txt in English field

[edit]

Is it possible to automatically switch all the descriptions created by Aleksandr Patil from English to Russian? He apparently is not aware that he is using the English field. Hjart (talk) 21:09, 1 November 2024 (UTC)[reply]

I guess we need a friendly bot owner. Ymblanter (talk) 21:55, 1 November 2024 (UTC)[reply]

How to copy texts from Wikipedia articles via browser automatically?

[edit]
Illustration of process I'm looking to automate more

Does somebody know how to automatically fetch Wikipedia article content into local files? The caveat here is that the API probably can't be used since the content is modified via CSS based on the CSS classes of the text content. Scraping the full HTML files and then converting these to text locally after the CSS modifications would probably be too difficult – I think automating the browser would be easier and maybe there are further methods.
The content I'm looking to copy is the article contents up to the See also/Notes/References section. It would be best if all tables removed via hiding their CSS classes mark the file somehow so depending on the table one can insert the table in list format manually (this is under development, it would be best if it automatically converted tables or added a note that 'Table XYZ is not included here'). Prototyperspective (talk) 12:27, 3 November 2024 (UTC)[reply]

Template:LicenseReview, where to find the dropdown menu messages?

[edit]

Hello,

as a license reviewer, I'll have the option to click either "[license +]" or "[license -]". On "license -", I get a dropdown choice of different options, like "all rights reserved" or "non-commercial". Where are these text strings located? Where is a good place to ask for additions to them (maybe here?)? I'd like to have something like License review failed -> Reason: "Source link dead & archived versions unusable." to document the reality. Regards, Grand-Duc (talk) 22:01, 3 November 2024 (UTC)[reply]

Hi Grand-Duc! If the file was uploaded long ago you might want to read/comment Commons:Village_pump#Almost_400k_files_need_license_review. Asking for new options could be placed at MediaWiki talk:Gadget-LicenseReview.js. --MGA73 (talk) 13:20, 7 November 2024 (UTC)[reply]

Tech News: 2024-45

[edit]

MediaWiki message delivery 20:46, 4 November 2024 (UTC)[reply]

User:FlickreviewR 2 appears to be broken

[edit]

As per User:MGA73's report on the bot creator's talk page, there has been a surge of failed Flickr license reviews by User:FlickreviewR 2, with a large number of them being returned as "size_not_found" without crop or other image processing. An example can be found at here, where I asked the bot to review for 5 times, and the bot returned that result for all instances.

So far we have ~300 pending files at Category:Flickr images needing human review. Is the bot broken? I would like to hear more comments before sending the matter to Phabricator, thanks.廣九直通車 (talk) 07:29, 7 November 2024 (UTC)[reply]

@MGA73: as original user who reported the problem. Also it appears that the problem happens intermittently per the bot's contribution log: sometimes the bot can correctly review an influx of files, but it also fails to review an equivalent amount of files too.廣九直通車 (talk) 07:38, 7 November 2024 (UTC)[reply]
@廣九直通車: To me it seems like the best chance of review is if there are only a few files for review at the same time. So if someone upload 100 files the bot will crash. I have asked my bot to request a new review for 3-4 files per minute and it reduced the number of unreviewed filed from more than 2000 files to 20. But new files show up non-stop. --MGA73 (talk) 09:58, 7 November 2024 (UTC)[reply]
@MGA73: Thanks, should I file a bug report on Phabricator? I'm afraid that the bot owner won't be responding to our concerns for some while.廣九直通車 (talk) 05:33, 9 November 2024 (UTC)[reply]
@廣九直通車: Yes please, that would be great! --MGA73 (talk) 08:38, 9 November 2024 (UTC)[reply]
+1 --PantheraLeo1359531 😺 (talk) 10:55, 9 November 2024 (UTC)[reply]
Task created at Phab:T379459, FYI.廣九直通車 (talk) 13:36, 9 November 2024 (UTC)[reply]
According to Phabricator, only the bot owner can fix the problem. Maybe I'll send an email to notify the owner?廣九直通車 (talk) 05:52, 11 November 2024 (UTC)[reply]
I got the email. I'll take a look later today or tomorrow. --Zhuyifei1999 (talk) 11:09, 16 November 2024 (UTC)[reply]

template:information undisplayable name

[edit]

I found the error. even if ・・・・ inputs template to them name but {{Author missing}}. thanks. --eien20 (talk) 22:28, 9 November 2024 (UTC)[reply]

Multilingual SVG file

[edit]

Can someone this file be multilingual. I mean that it can be translated to other languages. I want to translate it to Hebrew using this tool. -- Geagea (talk) 13:43, 10 November 2024 (UTC)[reply]

@Geagea:
You will not be able to translate some of the text because it has been converted to curves (Commons does not support textPath).
I restored your initial translations.
I added a switch near the top of the file so MediaWiki will know it is a multilingual file. IIRC, MediaWiki only looks at the first few megabytes. If it does not find a translation by then, it assumes the file is monolingual. Your file is large and confuses MediaWiki.
To see what text you can translate, load the SVG file into a browser and select all the text (control+A on Windows).
Glrx (talk) 16:02, 10 November 2024 (UTC)[reply]
Thanks @Glrx. I see that my translations are ok now. But if you are saying that it won't work for all the translations I'll use this tool in a new file. Thanks a lot. -- Geagea (talk) 23:53, 10 November 2024 (UTC)[reply]
Hi. You can make the file multilingual, but do not replace ka text with he text. That makes a mixed-up monolingual file.
SVG Translate will make a multilingual file. It will not translate the 2 dozen labels that were converted to curves (they have aria-label attributes on the g container).
Glrx (talk) 02:44, 11 November 2024 (UTC)[reply]
ok, thanks. -- Geagea (talk) 08:43, 11 November 2024 (UTC)[reply]

Garbled characters in SVG image

[edit]
Fiji ....

See File:Fiji Province.svg. How can I fix it?--BigBullfrog (talk) 14:54, 10 November 2024 (UTC)[reply]

@BigBullfrog:
The file was using the Fantasy font. On Commons, that apparently turns Latin characters into special characters (which is not UTF-8). I overrode that choice with a style block that specifies text elements should use sans-serif.
The file also has Arabic in it (with display="none"), but it should use translation units rather than translation planes.
Glrx (talk) 15:30, 10 November 2024 (UTC)[reply]
@Glrx: It doesn't work in some Wikipedias (e.g. en wiki, de wiki). --BigBullfrog (talk) 18:36, 11 November 2024 (UTC)[reply]
Those links work for me. You may need to refresh the page without using the browser cache. Glrx (talk) 20:25, 11 November 2024 (UTC)[reply]

Problem with thumbnails

[edit]

File:Yiannis Boutaris Drasi.jpg is doing a problem with thumbnails, it can't be used in Wikipedia, Regards!! --Ezarateesteban 22:03, 10 November 2024 (UTC)[reply]

Creation of new pages logged as performed unknown action "create/create"

[edit]

It seems that the creation of new pages currently shows up in the logs as [User] performed unknown action "create/create". Just reporting this for the technically minded... Gestumblindi (talk) 00:12, 11 November 2024 (UTC)[reply]

Yikes. filed —TheDJ (talkcontribs) 09:58, 11 November 2024 (UTC)[reply]

Please help me clean up this template

[edit]

I created {{AlphaFold}} to better address a frequent source of AI images that I was seeing, but I'm not good enough with templates to set up proper langswitching, etcetera. Could someone take a look and make sure I didn't break anything? Thanks! The Squirrel Conspiracy (talk) 04:48, 11 November 2024 (UTC)[reply]

Tech News: 2024-46

[edit]

MediaWiki message delivery 00:03, 12 November 2024 (UTC)[reply]

NamespaceInfo and subpages

[edit]

Special:NamespaceInfo lists namespaces that can have subpages. Sample: "User talk" can have subpages, "Files" can't.

Currently the following namespaces have subpages enabled, but I'm not sure if they should:

  • Category
  • Creator
  • Institution


 ∞∞ Enhancing999 (talk) 10:32, 12 November 2024 (UTC)[reply]

Problem with Commons categories in Multilingual sites in Wikidata, consequences for Commons

[edit]

In Wikidata there are Wikidata items that have a label that starts with "Category", have a statement "instance of" and have a link to "category's main topic" (and vice versa in the linked Wikidata item about the subject with a "topic's main category"); here referred to as Wikimedia category and subject Wikidata item.

I always thought that such Wikidata items in Commons are automatically put through to for example the Wikidata Infobox and corresponding Wikipedia's. Example: d:Q9097136 is such a Wikimedia category; its 'Multilingual sites' contains a link to the Commons category Category:Spijkenisse and it has no image, that is in the corresponding subject Wikidata item d:Q488545. The Wikidata Infobox in Category:Spijkenisse shows the image and other information from the subject Wikidata item and it shows even the link to the subject Wikidata item, but this subject Wikidata item does not have any 'Multilingual sites'. And the EN-WP article w:en:Spijkenisse has a link to the correct Commons category (under "Tools"), though there is no link in the Wikimedia category to the EN-WP article. So my conclusion is that there is an automatically procedure that links both Wikidata items to each other and their content to all the places where they are needed.
 Question Is this correct?

Problem: But now there is a discussion on Commons:Categories for discussion/2024/09/Category:Redirects connected to a Wikidata item where someone pointed out that there are situations in which this automatism does not work and so people invent work-arounds that causes other problems (in this case often an empty gallery page is created with a redirect to the Commons category, and this empty gallery page is used in the 'Multilingual sites' of the subject Wikidata item, see all pages in c:Category:Redirects connected to a Wikidata item for examples). One problem is a link in a Commons category to a Wikidata "Family tree" template:

  • Open the c:Category:Carlo_Buonaparte and the "Family tree" template. On the side of the sons you can see "Joseph Bonaparte" does not have any link, just a "[d]" you can click to enter his d:Q7726 Wikidata entry: There you can see the 'Multilingual sites' field has no entries, as the Category:Joseph Bonaparte is instead linked to d:Q31993664. If you are looking for images you have to do a long scroll to the Property:P373 in order to jump back to Commons. So even if the Commons category is present in the property "Commons category" of d:Q7726 it makes no direct link.

 Question Can this problem be solved? Is this the right place to ask for such a solution? I already have asked the same question on d:Wikidata:Project_chat/Archive/2024/11#Problem_with_Commons_categories_in_Multilingual_sites but I did not get an answer. JopkeB (talk) 05:39, 14 November 2024 (UTC)[reply]

I think it's fairly easy to fix: One should just add Commons sitelinks to the item with the articles they describe, whenever possible.
 ∞∞ Enhancing999 (talk) 12:42, 16 November 2024 (UTC)[reply]