Commons:Bots/Work requests

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

Shortcuts: COM:BR • COM:BWR

SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 7 days.


# Bot request Status 💬 👥 🙋 Last editor 🕒 (UTC) 🤖 Last botop editor 🕒 (UTC)
1 Hidden categories added as Category:Hidden categories 11 4 Enhancing999 2024-08-05 01:21 Fl.schmitt 2024-08-04 07:23
2 Revert additions to Category:History by Mitte27 7 5 Enhancing999 2024-06-30 11:06 Cryptic-waveform 2024-06-25 13:04
3 Convert Category:Photographs by Carol M. Highsmith to JPEG 28 7 DaxServer 2024-08-03 16:46 DaxServer 2024-08-03 16:46
4 Cities in Finland and China by month 2 2 DaxServer 2024-08-02 21:26 DaxServer 2024-08-02 21:26
5 Remove extraneous "I, " in author param of PD-self 10 5 Pelikana 2024-08-04 21:44 Jeff G. 2024-07-24 10:21
6 Images with borders (MTC) 4 2 DaxServer 2024-07-29 13:57 DaxServer 2024-07-29 13:57
7 Add OCR output to jpg 4 2 Prototyperspective 2024-09-21 11:45
8 Move "Historical images of" to "History of" 9 4 Adamant1 2024-08-10 07:27 Jeff G. 2024-08-04 10:20
9 P625 to P1259
Resolved
7 4 Schlurcher 2024-10-02 06:42 Schlurcher 2024-10-02 06:42
10 Media missing infobox template 22 4 Fl.schmitt 2024-09-04 18:31 Fl.schmitt 2024-09-04 18:31
11 Auto-addition of inferrable categories 4 2 Prototyperspective 2024-09-09 20:39
12 Generate a daily database report equivalent of Special:UncategorizedCategories 3 2 Prototyperspective 2024-10-07 16:45
13 file description cleanup: "Uploaded with Reworkhelper" 2 2 Ammarpad 2024-09-09 18:29
14 Add P1651 YouTube video ID structured data from "source" attribute of Filedesc template 17 6 Prototyperspective 2024-10-04 18:01 DaxServer 2024-09-19 07:44
15 Add missing Template:Location 14 3 Enhancing999 2024-09-21 12:37 Fl.schmitt 2024-09-20 19:57
16 Monuments database in Russia 1 1 Pigsonthewing 2024-09-24 11:14
17 Redirect Galleries per Concensus 1 1 Sadads 2024-09-30 12:26
18 Automatically removed closed requests from Deletion requests/mobile tracking
Resolved
4 3 Krd 2024-10-01 18:38 Jeff G. 2024-09-30 15:25
19 Remove double headings created due to a mass-message
Resolved
7 2 Fl.schmitt 2024-10-01 18:43 Fl.schmitt 2024-10-01 18:43
Legend
  • In the last hour
  • In the last day
  • In the last week
  • In the last month
  • More than one month
Manual settings
When exceptions occur,
please check the setting first.


Hidden categories is a system category added by __HIDDENCAT__

However, some files and even categories add it as regular categories: [[Category:Hidden categories]]

To find some: [1] (currently 468 in category namespace). Enhancing999 (talk) 13:22, 2 June 2024 (UTC)[reply]

I've reduced the numbers with Com:Cat-a-lot. The rest probably should be gone through manually. Jonteemil (talk) 23:04, 3 June 2024 (UTC)[reply]
Shouldn't they be replaced with __HIDDENCAT__? This finds those lacking that. Enhancing999 (talk) 23:18, 3 June 2024 (UTC)[reply]
I'm not sure all 128 categories really should be hidden. That's why I suggest they be gone through manually. Jonteemil (talk) 11:52, 4 June 2024 (UTC)[reply]
Currently 54 hits.  Support fixing this. [[Category:Hidden categories]] should NOT appear. — Preceding unsigned comment added by Taylor 49 (talk • contribs) 14:12, 26 June 2024 (UTC)[reply]
I think is done now. I've edited most of the remaining 43 categories using AWB. I was unsure about Category:Vector files with non-modifiable text since there, Category:Hidden categories is used as piped link.
{{Section resolved|Fl.schmitt (talk) 10:36, 14 July 2024 (UTC)}} Fl.schmitt (talk) 10:36, 14 July 2024 (UTC)[reply]

Thanks for the help. I had done a few as well. While doing the change manually helps adding more precise categories (like {{Source category}} or {{Usercat}} ) . I don't see an issue with systematically converting all uses going forward. Since July 14, a new use has been added: [2]. Maybe a bot that runs daily could include it too. Enhancing999 (talk) 11:54, 16 July 2024 (UTC)[reply]

Oops - sorry for the misunderstanding - i stopped reading too early :-) ... I've left a message on the user's talk page. Fl.schmitt (talk) 07:23, 4 August 2024 (UTC)[reply]
I asked R'n'B to include it in Russbot's tasks and cleaned up some of the noise in [3]. Enhancing999 (talk) 07:33, 4 August 2024 (UTC)[reply]
This doesn't need a bot. You can easily locate pages that have been added manually to Category:Hidden categories; see, e.g., https://quarry.wmcloud.org/query/85343. And I agree with Jonteemil that these pages should be reviewed manually; a bot has no way of knowing whether hiding the category is in fact appropriate. --R'n'B (talk) 23:52, 4 August 2024 (UTC)[reply]
The question is not if we can find them, but if we want fix them manually. We can easily empty category redirects manually too, but we don't really want to.
Did you find any in the 500 that needed manual review? Enhancing999 (talk) 01:21, 5 August 2024 (UTC)[reply]

Revert additions to Category:History by Mitte27

[edit]

Thousands of uncategorized files were added to the already-bloated Category:History. All of the edits I find were on 31 May 2024. Could some please automatically revert these edits? Thanks. Cryptic-waveform (talk) 20:55, 24 June 2024 (UTC)[reply]

I don't think it's a good idea to return it. My idea was to then move the files from "Category:History" to more specific categories. --Mitte27 (talk) 09:59, 25 June 2024 (UTC)[reply]
The current status is that thousands of files that were correctly marked as Uncategorized, and therefore easily visible to contributors doing a first round of categorization, are now erroneously categorized in a top-level category. Cryptic-waveform (talk) 13:04, 25 June 2024 (UTC)[reply]
@Mitte27: so when do you plan to move the images to more specific categories? This is clearly not an indefinite solution. —Matrix(!) {user - talk? - uselesscontributions} 18:55, 26 June 2024 (UTC)[reply]
I sorted out some photos related to the history of Russia/USSR, but I have little understanding of American history, and most of the photos in the category are related to it. In any case, this category is better than none. --Mitte27 (talk) 22:29, 26 June 2024 (UTC)[reply]
There is no reason to ever place files into extremely broad categories like Category:History. Please do not remove {{Uncategorized}} unless you are able to either accurate place a file in the most specific categories available or into a dedicated cleanup category. Pi.1415926535 (talk) 00:22, 27 June 2024 (UTC)[reply]
You could just use cat-a-lot. I don't think adding all LOC or NARA images to "History" by default is a good idea. Enhancing999 (talk) 11:06, 30 June 2024 (UTC)[reply]

Category:Photographs by Carol M. Highsmith is an excellent Library of Congress collection of very good images. Unfortunaly, all those images are in TIFF format, which means that the average file size is 100-300 MB, which is incredibly large. It causes long loading times of even the preview image (let alone the actual file), and TIFF file format is not supported by most browsers or general applications. Wikipedia discourages using TIFF files for those reasons, and this reduces the likelyhood of those excellent images being used.

Therefore, some bot should convert those TIFFs to JPEGs, copy the descriptions/categories and make sure the files reference each other. Further, the categories from the TIFF files should be replaced with Category:LC TIF images with categorized JPGs TheImaCow (talk) 21:59, 30 June 2024 (UTC)[reply]

@TheImaCow Thanks for finding this. I've filed for a bot Commons:Bots/Requests/ImageConverterBot -- DaxServer (talk) 15:13, 1 July 2024 (UTC)[reply]
I didn't expect someone to reply to this so quick, thank you!
I came across this series via Category:Aerial photographs of the United States and subcats, which contains many poorly categorized images from this collection. TheImaCow (talk) 16:40, 1 July 2024 (UTC)[reply]
LCCN2013631230.tif shows a jpg and several jpg-sizes are offered. Is this really needed? Enhancing999 (talk) 18:44, 1 July 2024 (UTC)[reply]
Hmm, I didn't notice that. It seems it is not necessary after all -- DaxServer (talk) 20:56, 1 July 2024 (UTC)[reply]
Maybe I'm blind, but where are those files offered? It's not the "Download/Use this file/Email a link" bar, all resolutions there only download the same low-quality preview generated by the Mediawiki software (which is shown on the file description page) TheImaCow (talk) 21:19, 1 July 2024 (UTC)[reply]
Below the image, there is a line:
"Size of this JPG preview of this TIF file: 800 × 533 pixels. Other resolutions: 320 × 213 pixels | 640 × 427 pixels | 1,024 × 683 pixels | 1,280 × 853 pixels | 2,560 × 1,707 pixels | 6,144 × 4,096 pixels."
The last one matches the tiff. Enhancing999 (talk) 21:52, 1 July 2024 (UTC)[reply]
Oh thanks I see. However this is very obscure and when embedding the file anywhere, it will always refer to the TIF version - so an seperate JPG should probably still be uploaded, like the 220,000 other TIF files in Category:LC TIF images with categorized JPGs (or the 58,000 Category:NARA TIF images with categorized JPGs)
But I don't have strong opinions on this. TheImaCow (talk) 22:11, 1 July 2024 (UTC)[reply]
Loading a file to test this -- DaxServer (talk) 05:51, 2 July 2024 (UTC)[reply]
Possibly support for tiffs was less developed when they were uploaded. I wonder how all those thousands of duplicates are curated and how much volunteer time is lost by handling two instead of just one copy of every image. WMF recently expressed their view on hosting files on Commons that aren't used on WMF sites [4]. Enhancing999 (talk) 09:22, 2 July 2024 (UTC)[reply]
Well, in theory, those TIF duplicates shouldn't need any curation, as they are supposed to be dumped into the massive categories mentioned above, and only linked from the description of the maintained JPG version.
The use of TIF is something I think is generally not needed for 99.9% of files, modern compression is more than good enough.
(I don't oppose eventually getting rid of the TIF duplicates, but there is not even consensus to delete de-facto duplicates where one version is rotated differently by single degrees, or random low quality TIF scans of generic text documents, where the same scans are also uploaded as JPG, so forget it) TheImaCow (talk) 13:28, 2 July 2024 (UTC)[reply]
Oddly, I can't figure out which one of the two maps is correct ;) Did you nominate the wrong one? For the text ones, I'd have nominated the jpg ones. The assumption that deletion doesn't save anything is incorrect: deletion reduces curation (even if theoretically none is needed, it still happens and wastes volunteer time), limits spamming of Special:search, can even save storage space as files can be purged (from non-public view) or wont be exported twice when requested.
As technology changes, I think views on this evolve. NARA's approach might have been the ideal 15 years ago, but other GLAMS that started only more recently use different approaches. Enhancing999 (talk) 12:13, 3 July 2024 (UTC)[reply]
Not sure what you mean. Both maps are exactly the same. JPG ones nominated instead? Ideally someone uploads a PDF and 307 files are replaced with one in the correct format for documents. I never said I oppose deletions, I said the exact opposite.
The NARA approach has actually changed - there have been at least two bulk uploads, one in 2011 and the other 2019.
The 2011 one uploaded nearly every image twice - one TIF+one JPG. The 2019 one uploaded only JPGs.
Looking at the NARA catalogue, files uploaded earlier have often TIF,JPG and sometimes GIF versions for download. Images uploaded 2019, presumably digitized later, have only high-resulution JPGs for download. TheImaCow (talk) 18:33, 3 July 2024 (UTC)[reply]
It's better to have the lossless files than a JPEG, as you can always make a JPEG from a lossless file, but you can't make a lossless from a JPEG. Still, while we shouldn't delete the TIFFs, we should make JPEG options. Adam Cuerden (talk) 08:52, 4 July 2024 (UTC)[reply]
If we want to offer lossless files in a reasonable sizes (2MB vs 200MB), we might want to consider offering PNGs instead of JPEGs -- DaxServer (talk) 08:57, 4 July 2024 (UTC)[reply]
@DaxServer: Please don't, PNG images look fuzzy when scaled down (due to design decisions discussed in phab:T192744) on WMF projects.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 12:59, 13 July 2024 (UTC)[reply]

What should be done now? Is there any reason to not do what has been already done successfully with hundreds of thousands of NARA/LOC files? TheImaCow (talk) 11:57, 24 July 2024 (UTC)[reply]

What leads you to describe it as "successful"? How many edits had to be made because we have the same file twice? Enhancing999 (talk) 12:02, 24 July 2024 (UTC)[reply]
"successful" because the TIF versions are dumped into the massive TIF categories linked above and linked in the "other_versions=" parameter at the information template of their respective JPG version, in case anyone needs them. JPG versions are maintained, TIF versions are just there. And there hasn't been much of an issue with as far as I can tell.
Please show some nessescary manual edits that had to be done twice, because when done right, there aren't any. TheImaCow (talk) 17:02, 24 July 2024 (UTC)[reply]
Any edit on the version you consider secondary is as waste of curation energy. Wouldn't professionally managed archives clean this up beforehand rather than waste our volunteer's time to clean it up?
Sometimes I wonder if they employ uploaders paid by the number of files uploaded. We seem to end up with books added page by page in duplicate copy from what should be a single djvu document. Enhancing999 (talk) 18:20, 24 July 2024 (UTC)[reply]
"Any edit on the version you consider secondary is as waste of curation energy." - Yes obviously and I fully agree on that. Thats why there are categories like Category:LC TIF images with categorized JPGs - these images have JPG versions which are being maintained, and the TIF files in this category are uncategorized besides being in that category/referenced from the JPG file description page. This means that there is no need to ever do any edits on TIF files in that category, as only the "categorized JPGs" are maintained.
(Topic Paid by files uploaded: I don't think so, this is simply the format those scans are stored, and with proper software to handle this, there isn't anything wrong - but Commons dosen't have the software, and i fully support efforts to convert single page uploads of books into PDF. The core problem is that Commons software is not designed to handle the same media in multiple file formats, like e.g. the Internet Archive which offers texts for download in countless different file formats from the same description page (random example.)) TheImaCow (talk) 20:03, 24 July 2024 (UTC)[reply]
These seems like an over optimization. Things work, they are a bit slower because someone thought having a very high resolution in a very badly compressible format was desirable. The side effect of very large originals is that it takes a while before the thumbnail is ready. But for 99.9% of the people that isn't a problem. Thubmnails are cached. If images are used, you therefor never have to wait for the thumbnail. You are in the .1% of people (Curators) looking at things that are NOT used. It's acceptable to wait a second in that case. —TheDJ (talkcontribs) 13:47, 24 July 2024 (UTC)[reply]
"someone thought it was desireable" - The Library of Congress used an archival format to archive the images, but we are not the LOC, and have different goals - so we should use file formats better suited for our uses.
The motto of this site is "freely usable media" - this also includes not having to have very fast internet connections or special programs to process or even fully view the file (at least edge/firefox cannot show raw .tif files in the browser, e.g. when trying to zoom in further), and this list could be expanded endlessly. We shouldn't forget our actual end users, who in general have much less knowledge about dealing with such file types, or generally anything.
Another issue not yet mentioned is that TIFF files are not indexed by Google Image Search and presumably other search engines, which is bad for obvious reasons. (search for site:commons.wikimedia.org carol highsmith, and there are only a couple hundred images which have been manually converted to JPG, but not a single of the 30,000 TIFs, appending filetype:TIFF dosen't return anything at all) TheImaCow (talk) 17:02, 24 July 2024 (UTC)[reply]
If Google is broken than we don't want users at Commons having to fix it. Enhancing999 (talk) 18:22, 24 July 2024 (UTC)[reply]
This is fully intentional, as demonstrated plenty of times, TIFF is a format simply not suitable for general web use. TheImaCow (talk) 20:05, 24 July 2024 (UTC)[reply]
Weirdly, it appears to remove jpgs, see #c-Enhancing999-20240701184400-TheImaCow-20240630215900. Enhancing999 (talk) 10:38, 25 July 2024 (UTC)[reply]
 Comment Yes, on Commons, we need JPEG. If the source may not be available in the long term, we should also upload the original TIFF versions, but that not the case here. Be sure to link both versions. That was not done for some other files uploaded from LOC or NARA, and it is now a mess. Yann (talk) 18:09, 24 July 2024 (UTC)[reply]

Cities in Finland and China by month

[edit]

Hello! I would like to ask you to automatically create categories for the distribution of cities in Finland and China by month. There are corresponding templates: {{MonthinFinlandbycity}} and {{MonthinChinabycity}}. MasterRus21thCentury (talk) 07:22, 17 July 2024 (UTC)[reply]

Hi @MasterRus21thCentury Could you explain it a bit more? Thanks! -- DaxServer (talk) 21:26, 2 August 2024 (UTC)[reply]

Remove extraneous "I, " in author param of PD-self

[edit]

@Pelikana noticed that there are a lot of erroneous uses of {{PD-self}} which insert "I, " before the author. Could someone please replace {{PD-self|author=I, with {{PD-self|author= in the following pages? —CalendulaAsteraceae (talkcontribs) 06:52, 24 July 2024 (UTC)[reply]

I can do that -- DaxServer (talk) 09:46, 24 July 2024 (UTC)[reply]
@CalendulaAsteraceae and @Pelikana: "I, " is there to make the assertion first person.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 10:21, 24 July 2024 (UTC)[reply]
Hi @Jeff Yes, it obviously used to be there to make the assertion first person. But I think at some point the textlines were changed and now IMHO it is a displaced element, plus very odd that it is the only not translated text element in the template, at least in these use cases. Do you mean to say the results are completely correct this way and need no change? Both lines seem grammatically faulty to me "... door de auteur, I, JohnDoe" (".. by the author, I, JohnDoe") and "I, JohnDoe allows ...". Last one should read (in Dutch) "Ik, JohnDoe sta ...." It should not read "I, JohnDoe staat ... " because this line starts in first person and ends in 3rd person. In later days (past 2007-2008) the "I, " "I, " is not in the templates anymore it seems. Peli (talk) 10:52, 24 July 2024 (UTC)[reply]
Indeed. The template uses {{int:Wm-license-pd-author-with-author-text}}, which produces the text "This work has been released into the public domain by its author, $1. This applies worldwide." The appropriate way to make this first person would to edit the page on TranslateWiki (well, the English one needs to be changed in MW code, but for other languages this is where you'd edit it), not to manually put "I, " in the author parameter. —CalendulaAsteraceae (talkcontribs) 20:50, 24 July 2024 (UTC)[reply]
I think it is a good idea to add "I, " as a suffix if the uploader is also the work's creator. Please don't replace that. For example, it may not be clear to many or people only or first check the author field where this is useful metadata, especially if the author name is different from the username in which case they would also need to check the license template. Prototyperspective (talk) 12:04, 25 July 2024 (UTC)[reply]
This is a good thing to handle in {{PD-self}} (which is a template only intended to be used by the uploader). Adding it manually means it's a huge pain to update if the wording of the template changes, and also doesn't work with internationalization. Right now
{{int:Wm-license-pd-author-with-author-text|I, Calendula}}
produces

This work has been released into the public domain by its author, I, Calendula. This applies worldwide.

in English, which is ungrammatical and frankly silly. If I switch my display language to Spanish, it instead produces

Este trabajo ha sido liberado al dominio público por su autor, I, Calendula. Esto aplica para todo el mundo.

which is even worse. If you want to change the wording of {{PD-self}}, probably the way to go is switching in the template from int:Wm-license-pd-author-with-author-text to something like int:Wm-license-pd-author-self-text that incorporates the author's name. —CalendulaAsteraceae (talkcontribs) 19:28, 25 July 2024 (UTC)[reply]
I couldn't find an existing piece of text, so I submitted a feature request at phabricator:T371057. I think that further discussion of updates to the text of {{PD-self}} should go to the template talk page, and also that this bot request should go ahead because manually adding "I, " before the author's name is a terrible way to make the template first-person. —CalendulaAsteraceae (talkcontribs) 20:28, 25 July 2024 (UTC)[reply]
You're absolutely right. Sorry, I misunderstood. It's not really clear in your initial post that this would be added to the template instead. Prototyperspective (talk) 21:02, 25 July 2024 (UTC)[reply]
I agree that this is just about cleaning up a tiny bit of lost and redundant text on a limited number of pages and would be glad if @-- DaxServer would get the green light to fix this series of typo's, on these old pages by deleting "I, ". ThanksPeli (talk) 21:44, 4 August 2024 (UTC)[reply]

Images with borders (MTC)

[edit]

Many images in Category:Independence Day 2019 in Brasília have a border. Sample: File:Comemoração da Independência do Brasil (48700486098).jpg

These should be added to Category:Images with borders. Possibly the same applies to more in from the same MTC Flickr stream. Enhancing999 (talk) 11:18, 27 July 2024 (UTC)[reply]

Assuming the border always has the mark to the website www.mctic.gov.br website at left bottom, here's what I thought of: Load the image with OpenCV and extract the left bottom part, use Tesseract to do OCR for the website text, do a sequence match with the extracted text and the website string and if the comparision is very high enough that can be categorized.
Here is a sample code: https://www.kaggle.com/code/daxserver/detecting-borders-from-brazil-mtc-flickr-images/ -- DaxServer (talk) 17:25, 27 July 2024 (UTC)[reply]
I did some screening on Category:Independence Day 2019 in Brasília by changing the background color of the page. It appears that there are a few images without a border. The ones I checked were all from other Brazilian government agencies. Sample: File:07 09 2019 - Desfile 7 de setembro. (50751888331).jpg.
The magic border locator of the crop tool does work fairly reliably on these images. Sample: https://croptool.toolforge.org/?site=undefined&title=Comemora%C3%A7%C3%A3o%20da%20Independ%C3%AAncia%20do%20Brasil%20(48700486098).jpg&page=undefined
The only problem with directly cropping them seems to be that the file description pages don't include all details from the borders. Enhancing999 (talk) 10:53, 29 July 2024 (UTC)[reply]
The magic borders module is interesting. Perhaps we can employ that to detect a border. I'll do some tests -- DaxServer (talk) 13:57, 29 July 2024 (UTC)[reply]

Add OCR output to jpg

[edit]

From the discussion at VP/T, I found a solution to a problem identified earlier: frequently we have images of streets and other with some text in it. Sometimes this is of interest, but it's not necessarily included in filename or description.

https://ocr.wmcloud.org/ would allow to extract such text and make it editable on Commons.

Ideally a bot would go through new uploads (and also some maintenance category for older files) and run https://ocr.wmcloud.org/ on it. The output (if any) could be added to the file description page, either with a template or as structured data.

Sample file:

File:Plaque Rue Lauriers - Gournay-sur-Marne (FR93) - 2021-10-04 - 1.jpg

Input:

https://ocr.wmcloud.org/?image=https%3A%2F%2Fupload.wikimedia.org%2Fwikipedia%2Fcommons%2F3%2F36%2FPlaque_Rue_Lauriers_-_Gournay-sur-Marne_%2528FR93%2529_-_2021-10-04_-_1.jpg&engine=google&psm=3&line_id=null

Output:

"PER PONTEM AD FORTUNAM GOURNAY-SUR-MARNE RUE DES LAURIERS"

Enhancing999 (talk) 15:16, 2 August 2024 (UTC)[reply]

Also please see the discussion at VP/T linked above. Just briefly adding support to this wish and two notes:
it would likely be a problem to scan all files on WMC and/or all new uploads, instead one could let the bot run only categories where this may be useful. Secondly, rather than writing a new bot it would probably be better to add this functionality to some bot that already writes e.g. structured data to lots of files (however SD can't be searched on WMC can it?) like SchlurcherBot. Prototyperspective (talk) 10:50, 4 August 2024 (UTC)[reply]
I think there are already some bots who scan all uploads .. it could obviously be added to those. If SD is used, we should make sure it's searchable. Enhancing999 (talk) 13:47, 4 August 2024 (UTC)[reply]
I think the place to put OCR results would be a new field for the file summary box that is collapsed by default. This way all users can easily find and see this info and it can be searched. The mentioned VP/T thread is now archived to here. I think adding a way to categorize based on OCR results would be quite useful. However, not extremely useful so I don't know if it's necessarily worth the effort to develop a categorization-based-on-OCR tool or extension for the ocr-tool. This is why after briefly asking about it here I only listed the task I meant to use this for at the new page Commons:Categorization requests. Prototyperspective (talk) 11:45, 21 September 2024 (UTC)[reply]

Move "Historical images of" to "History of"

[edit]

Per note at Category:Historical images by country (as conclusion from Commons:Categories for discussion/2019/09/Category:Historical images), the content of the categories at Special:PrefixIndex/Category:Historical images of should be moved to "History of". This seems to involve more than 10'000 categories, see PetScan:29034509. I think the resulting redirect could afterwards be tagged for speedy deletion. Enhancing999 (talk) 18:59, 2 August 2024 (UTC)[reply]

i dont think it's a good idea to handle this problem without human supervision.
i would rather do these instead:
  1. prohibit new categories with the word from being created.
  2. let users slowly move the files to the appropriate categories (by time).
RZuo (talk) 20:42, 2 August 2024 (UTC)[reply]
"history of ..." is not any better. everything is history. RZuo (talk) 20:43, 2 August 2024 (UTC)[reply]
Right, any cutoff for "history" will change every second/minute/hour/week/month/year/century/millennium. See also Commons:Categories for discussion/2024/08/Category:History by country.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 10:20, 4 August 2024 (UTC)[reply]
there's specific interest related to "history" of something.
for example, historians of asian history should go under "history of asia".
but to dump files into "history of xx" is no more better than dumping them in "xx" or "historical images of xx". all files of xx can perfectly fit into all those three variations.
most of these "historical images of xx" basically contain all photographs before the advent of digital photography, especially black and white photographs.
so i'd rather users move these cats to or create for example "xx in the 19th/20th century". RZuo (talk) 12:49, 4 August 2024 (UTC)[reply]
i have an idea of a bot moving files according to the time/date, but i need probably 1 or 2 years to code something like that up. RZuo (talk) 12:53, 4 August 2024 (UTC)[reply]
I don't think this is the place to re-discuss the CfD. If you think the closure is problematic, ask an admin to re-open it. Enhancing999 (talk) 12:57, 4 August 2024 (UTC)[reply]
There is just no way this can be done manually. If there are cases you think would be problematic, please state them here. Enhancing999 (talk) 20:56, 2 August 2024 (UTC)[reply]
  •  Support Per Enhancing999. There's currently 33732 categories for "historical images", which is way to many for anyone to deal with manually. This also isn't the place to relitigate the CfD. Nor do I think doing so would go anywhere anyways since it was open for 4 years and has been closed since last year. So there has been plenty of time for people to raise concerns about it. Most of these categories only contain a couple of images to begin with and they aren't "historical" either. The idea that we should let users slowly move the files to the appropriate categories when it's only a couple of images per category to begin with is totally ridiculous and would just waste everyone's time. There's no reason people can't better categorize the images once they are moved to "history of xx" categories. That's where most of the images were in the first place. Regardless, this should totally be done by a bot instead of forcing users to waste time doing it manually. --Adamant1 (talk) 07:27, 10 August 2024 (UTC)[reply]

P625 to P1259

[edit]

i just made mistakes of adding coordinate location (P625) to files. is there a bot already running that automatically migrates p625 statements to locality of creation (P1071) coordinates of the point of view (P1259)? RZuo (talk) 10:37, 6 August 2024 (UTC)[reply]

made a mistake again. it seems p1071 is meant for wd items of actual places, not coords.
Commons:Structured data/Modeling/Location, important info.--RZuo (talk) 11:29, 6 August 2024 (UTC)[reply]
OpenRefine can do that. I can help if you'd like. -- DaxServer (talk) 10:58, 6 August 2024 (UTC)[reply]
manual is easy, but this should be a continuous bot job.--RZuo (talk) 11:29, 6 August 2024 (UTC)[reply]
See Module_talk:Coordinates#Usage_of_P625?. Multichill (talk) 17:21, 13 August 2024 (UTC)[reply]
I'm working on this on an ongoing basis. --Schlurcher (talk) 06:42, 2 October 2024 (UTC)[reply]
Checkmark This section is resolved and can be archived. If you disagree, replace this template with your comment. Schlurcher (talk) 06:42, 2 October 2024 (UTC)

There are about 339,000 files in the category Media missing infobox template. Even using add_information.php (or the gadget), the task is too huge to be done manually. I assume that would be a nice job for a bot. A simple search/replace wouldn't be sufficient, since some file pges contain {{Filedesc}} and {{License-header}} which should be preserved. Additionally, some files have information on sources, e.g. 1884 South Penn RR.jpg. Those should be used for the source parameter of the information template. Fl.schmitt (talk) 19:22, 10 August 2024 (UTC)[reply]

Maybe a list could be generate from the category about the most used files and these be done manually? Also, please keep in mind COM:GOF. Enhancing999 (talk) 07:09, 12 August 2024 (UTC)[reply]
Good idea - restricting on the most used files is reasonable. Additionally, i thought about grouping by uploader / author which would facilitate automatic editing. Fl.schmitt (talk) 07:25, 12 August 2024 (UTC)[reply]
I tried Special:Search/switzerland incategory:"Media_missing_infobox_template" and then used Petscan:29082230 to find the uploaders.
This found images like File:Runs_Kapelle.jpeg by "Ikiwaner" who uploaded plenty of own pictures which is clearly indicated, but even add-information can't complete it.
One would think that we'd have more pictures of these places almost 20 years later, but sometimes we don't. Enhancing999 (talk) 08:30, 12 August 2024 (UTC)[reply]
Looks very interesting! The problem with add-information.php is that it has to transform arbitrary input, which is IMO almost impossible. With pre-structured data (known author/uploader, known structure of file description), maybe the task can be automated to a certain extend. Limiting the input by location is a good idea! Fl.schmitt (talk) 08:55, 12 August 2024 (UTC)[reply]
add-information.php seems relatively good based on the input, but a review seems necessary.
Even when filtering by uploader can give large range of complicated cases (especially old imports from other wikis). Adding a search for "own photograph" (or similar) can simplify things. Enhancing999 (talk) 12:02, 12 August 2024 (UTC)[reply]

Maybe we could list groups of a similar cases somewhere, so someone else can determine if they want to assess them further (or they are all actually similar). Samples:

Enhancing999 (talk) 15:01, 12 August 2024 (UTC) updated[reply]

We could do subcategories of Media missing infobox template for maps, logos, coats of arms, insignia, currency, flags and personal templates. There is already one for artwork.
Interesting to compare the early digital photos with others we have: sometimes it still looks the same, others lack any comparable one, sometimes it's clearly aged, sometimes it gives a historic comparison, sometimes in a larger set we lack clearly better ones.
BTW, image notes seem to be handled badly by add-information (they get mixed into the description). Headers handling could be improved too. I don't think I ever had one that didn't need editing (that seems to be the idea anyways).. besides, I try to complete them. Enhancing999 (talk) 13:38, 13 August 2024 (UTC)[reply]
@Enhancing999: Great work, this is very helpful. I've started with the maps provided by Tschubby, because it seems that most of the file description shares the same structure. Please check Revision #909185535 of Karte Gemeinde Troinex.png for a regex-based replacement by pywikibot. IMO, this looks ok. Fl.schmitt (talk) 16:49, 13 August 2024 (UTC)[reply]
The problem with File:Karte Gemeinde Troinex.png is that it wasn't uploaded by Tschubby, so {{Own}} isn't applicable.
Supposedly that file and File:Carte Commune Troinex.png are based on a file that was initially uploaded at de:File:Karte Gemeinde Troinex.png, see https://de.wikipedia.org/w/index.php?title=Spezial:Logbuch&logid=283755 . Normally the file description page would include copy of the upload log from dewiki, but it doesn't. File:Glacier.zermatt.arp.750pix.jpg had some details I added after "own".
BTW, Tschubby is still very active, so he might have a view how he prefers them to be done or do them directly himself. Enhancing999 (talk) 17:03, 13 August 2024 (UTC)[reply]
If it's the same file, initial upload was: [5]. Enhancing999 (talk) 17:09, 13 August 2024 (UTC)[reply]
hmm - ok - yes, seems I was too optimistic... it's clear that getting this done by a bot will never reach the quality of manually checking / editing all the parameters. So we will have to decide which grade of completeness is achievable / required. Searching for other / derived / source versions can only be done manually, I think. So if this is a requirement, there's no way to get this task done by a bot, not even a small part of this task.
What's possible IMO is to group the files by the structure of their description, maybe additionally by uploader and year/month of upload, and do a regex-based replacement. This may lead to incomplete Information/Map/Artwork templates, e.g. if there's no information regarding the source.
Regarding the parameters:
  • Setting the source parameter may be possible (1) if the source is stated in the description or (2) if uploader is identical with author. In other cases, the source can't be set automatically.
  • Setting the exact upload date will be very difficult if we use pywikibot's replace script. If using the upload's year and month is sufficient, one could group the files accordingly, based on a PetScan search. This depends on the required/acceptable grade of precision.
Fl.schmitt (talk) 18:13, 13 August 2024 (UTC)[reply]
Still trying to get the {{Upload date}} template working... Fl.schmitt (talk) 16:51, 13 August 2024 (UTC)[reply]
I try to avoid upload date. Weirdly, add-information tends to get even the exif date wrong. For Tschubby's municipality maps, it may be sufficient to add the year they are meant to be current (borders don't change that frequently). Enhancing999 (talk) 15:00, 15 August 2024 (UTC)[reply]
when you can identify some common pattern in some file sets, Commons:AWB or jwb might be a good tool. RZuo (talk) 22:04, 14 August 2024 (UTC)[reply]
What would be cool for add-information is if one could use it with some defaults (description language, author, date, {{Taken on}}-location, source, other fields, license, etc) for a given subset.
Also, a few bugs might be worth fixing (licence header formatting, keeping image annotations together, placement of coordinates template, exif dates) if others plan to use it (I'm mostly done with the subset I'm looking into). Enhancing999 (talk) 15:05, 15 August 2024 (UTC)[reply]
Enhancing999, Thank you for tackling this long neglected problem. I like your divide-and-concur approach, and I agree with RZuo that Commons:AWB might be a good tool to use. That is what I used when some years ago I was adding infoboxes. Another possible approach might be to start adding com:SDC data like author, description and date with QuickStatements tool. If you do that than you can just add {{Information}} template with no parameters and it will display SDC data. See File:Indoor_Climbing_Kid.jpg for example. If you have any questions about this approach I can explain with more details. --Jarekt (talk) 04:21, 16 August 2024 (UTC)[reply]
Good idea indeed. This could simplify adding only one aspect at a time (not everything can be determined with the same ease). Once sufficient data for {{Information}} is available, the template could be added. We just need to be careful that basic information available as statements is also otherwise visible.
BTW, one would think that it's an old issue, but sometimes even recent uploads don't have a template (or someone deleted it).
If it's thought helpful for others, I can create subcategories for some or most of the above groups (obviously they should be deleted easily once empty or if a better one can be found).
If it's easy to add by bot, a subcategory for frequently used files could be helpful. (it's doable with PetScan for a relatively small set, but not for all 337000 files in the category). In the subset I checked few had more than 30 main namespace uses (sample, now with template). Enhancing999 (talk) 11:21, 16 August 2024 (UTC)[reply]
Flickr might a good start to add {{Information}} through statements only. We currently have ca. 4500 files mentioning Flickr. Some 2100 have both creator and source. An issue with some of these seems to be that they are blank. I brought this up at Schlurcherbot. Wouldn't the various Flickr templates also include source and creator? Enhancing999 (talk) 15:01, 16 August 2024 (UTC)[reply]
@Enhancing999 thank you for creating the hiddencat - this makes it easier to get a clearly defined set of files as input for bulk modifications! I'm currently working on a bot that should be able to work through the grouped files, preferably writing SDC data wherever possible. But there are some points where I'm not sure about:
Fl.schmitt (talk) 16:11, 22 August 2024 (UTC)[reply]
Commons_talk:Structured_data might find you help on the question specifically for structured data.
If {{Information}} has no date, the line doesn't even appear as missing. Sample: Special:Diff/914493166.
I noticed some uploaders use {{Own}} and link directly their username at Wikipedia. Not sure how bots handle this.
Reimports from Wikipedia are tricky in general. See also: Commons:Village_pump#c-Jarekt-20240817151300-Asclepias-20240817140600 Enhancing999 (talk) 19:29, 22 August 2024 (UTC)[reply]
@Enhancing999 - it took some time, but my bot solution is almost ready for action. Since handling weak-structured data is tricky, the bot first prepares (and actually prepared) just a "simulation" result, without any "live" modifications of Commons pages. This "simulation" result shows the proposed modifications for a certain set of file pages lacking {{Information}}. The bot tries to add as much information as possible by SDC (esp. Date and Author) and doesn't repeat those values in the generated {{Information}} template, since the template uses those SDC values by default. So, the template may look "incomplete" (for reference, see e.g. File:Karte Bodensee Birnau.png where I added as much as SDC as possible manually, leaving the respecting files in the Information template empty). The simulation result is available on gitlab in two formats: plain txt and SQL (sqlite). Before filing a bot request, I would be glad about any critical feedback regarding the proposed modifications. Fl.schmitt (talk) 18:31, 4 September 2024 (UTC)[reply]

Auto-addition of inferrable categories

[edit]

Could somebody create a bot to add categories that are inferrable given the structured data (SD) that other bots have added or the combination of the file's existing categories?

I previously asked if this functionality could be added to User:SchlurcherBot here but it seems like it won't be done. Schlurcher (talk · contribs) said there already are some bots adding categories – maybe instead of creating new bot(s) for this it would be best to add this functionality to these (please name these because I don't know what the relevant bots would be).


SchlurcherBot for example already reads and parses the date field, it would be nice if a bot did the same and then added "Category:Videos of 2023" if that's in the field which the bot already writes into the "inception" field. It could also be put in a hidden subcat like "Uncategorized videos of 2023" so people can check these and/or it doesn't clutter the category which could also be specific to videos that show something specific to the year like an event (this doesn't seem to be the case currently).

Likewise, the bot already writes the display resolution to the SD but does not add the respective Category:Videos by display resolution subcat. If not nearly all videos are in there I don't see why this category (its subcats) could be useful. If it was added to videos, then one could use this for statistics, petscans and maybe other things. The same goes for the WebM videos category which is currently up for deletion. Most webm videos are missing there so the category is largely useless. (Note that these two are exceptional cases: most WMC categories are useful.) If files were in there one could for example use this as a workaround to find videos in petscan which currently can't filter for videos except when combining categories with the Category:Videos by file format cat. A bot could also populate the Category:4K videos.

Some further examples of inferrable cats:

  • "Videos of 2024 from the United States" (depending on the license tag or other categories of the file or the coordinates)
  • Category:Audio files of 1906 (see here)
  • when a video file is in a category like "Muscidae" it should be added to Category:Videos of Muscidae
  • when a video file is in category "Azeliinae", which is a subfamily of Muscidae, it should also be added to "Videos of Muscidae" because that is the closest category to Muscidae which has a subcategory for videos (it would be best if a bot moved it to the more specific subcat once it is created)
  • files in Category:Animals in water should be moved to Category:Elephants in water if it's also in a subcat of elephants
  • it could make sure artworks like paintings (e.g. somewhere in a subcat of Visual arts) are in an "in art" subcat so paintings of elephants eating are not directly in the Elephantidae eating category and can e.g. be easily filtered out using the deepcategory search operator (or viewed alongside other such artworks)
  • a video in a subcategory of Category:Black and white films should go into Category:Black and white videos
  • After some delay to allow removal of metadata, categories like Category:Taken with Canon PowerShot A480 (example) or Category:Photographs by exposure time‎ subcats based on the file exif metadata

More difficult:

More things could be added and refined to such an automatic categorization system(s) over time. There can be rare exceptions but having things auto-categorized with exceptional errors would be better than things missing and requiring lots of manual maintenance/subcategorization and there would be ways to deal with that (for example for video files in Category:Short films it would create a 'suggestion' to add Category:Short films videos) and move things out of ill-inferred categories where usually another cat of the file is false.

Here I suggested that video2commons adds as many of inferrable cats right away when uploading, such as "Videos with English subtitles" if it imports en subtitles.

It would be a big endeavor but addition of categories that are inferrable from other categories of a file would be very useful, very much improve the reliability / completeness / usefulness of categories, and free up categorizers' time for other tasks.

Maybe focusing this discussion on inferrable technical criteria cats for videos would be best for now. Prototyperspective (talk) 23:05, 17 August 2024 (UTC)[reply]

I'll rarely add some more inferrable categories to the list. Added 2 and edited 1 now. --Prototyperspective (talk) 12:50, 30 August 2024 (UTC)[reply]

Videos without audio should actually be fairly straightforward, since it's possible to detect if a video has no audio channel at all.
I think implicit in what you wrote above is that a bot should add subcategories of Category:Vidoes by display resolution, which is what I came here to suggest. —Justin (koavf)TCM 14:03, 9 September 2024 (UTC)[reply]
How is that possible? (Is that possible without having to download the full video from WMC?) Yes, there are lots of inferrable categories like the ones about resolution and rather than handling each individually I think it needs a broader approach where many inferrable cats are added at once. Display resolution is just one and probably not even among the most useful ones (eg not really suitable much for showing results sorted by resolution or specifying min resolution). Since the deepcategory search operator very often doesn't work one can use Category:Videos by file format in PetScan as a workaround to show videos of whatever subject one is interested in but it's missing quite a lot of videos. Prototyperspective (talk) 20:39, 9 September 2024 (UTC)[reply]

Generate a daily database report equivalent of Special:UncategorizedCategories

[edit]
initial request and related discussion
Generate a daily database report equivalent of Special:UncategorizedCategories

For each page, output:

  • Page name
  • Date time last edited
  • Creator name
  • User who last edited
  • Wikidata item (if available)

Ideally formatted in a template. Enhancing999 (talk) 14:27, 24 August 2024 (UTC)[reply]

Is there a way to find categories modified / created in the last 24 hours? Fl.schmitt (talk) 17:53, 24 August 2024 (UTC)[reply]
https://commons.wikimedia.org/w/index.php?sort=last_edit_desc&search=contentmodel%3Awikitext&ns14=1
https://commons.wikimedia.org/w/index.php?sort=create_timestamp_desc&search=contentmodel%3Awikitext&ns14=1 . RZuo (talk) 19:34, 24 August 2024 (UTC)[reply]
Thanks! But it's difficult to build a stable PagePile from a query... - here's a Quarry #85732 searching for Categories created today and a Petscan (psid: 29146248) based on that quarry (same for yesterday: Quarry #85736 / Petscan (psid: 29146272)). Would be nice if pywikibot could "update" the quarry automatically, since the pagepile is only a "snapshot" of the quarry result at the time the pagepile was created. Or the bot would have to query a replica db itself... Fl.schmitt (talk) 21:27, 24 August 2024 (UTC)[reply]
Something like Quarry:history/85731/922919/895636 should work (1187 rows today, ran in 27 minutes). I'm trying to add the Wikidata item. User names seem more difficult, but I guess we can do without. Enhancing999 (talk) 21:34, 24 August 2024 (UTC)[reply]
Username did work too Quarry:history/85733/922940/895656 (1167 rows, ran in 60 minutes). Username seems interesting too. Enhancing999 (talk) 21:47, 24 August 2024 (UTC)[reply]
with Wikidata item Quarry:history/85731/923002/895717 (1162 rows, ran in 27 minutes; 109 have items). Maybe I can combine the two. Enhancing999 (talk) 21:50, 24 August 2024 (UTC)[reply]
An initial output for now at Commons:Report_Special:UncategorizedCategories. Enhancing999 (talk) 22:48, 24 August 2024 (UTC)[reply]
@Enhancing999: Excellent work! Is there a way to create such a report automatically based on a Quarry? Fl.schmitt (talk) 10:01, 25 August 2024 (UTC)[reply]
That's the point of my bot request: the idea is that the bot runs the report daily and add its result to the page. I'm still working on the optimal format.
(BTW, as another bot adds infoboxes to categories with items, there shouldn't be hardly any on the list. I reported this to User talk:Mike Peel.) Enhancing999 (talk) 10:08, 25 August 2024 (UTC)[reply]
Here is the combined query: Quarry:history/85733/923138/895851. It seems it runs quicker now that we worked to limit the number of categories on it. Enhancing999 (talk) 13:14, 25 August 2024 (UTC)[reply]
@Enhancing999: I've found thatt with Pywikibot, it's quite easy to get the metadata (last time edited, last user editing, wikidata item and so on) for any page or category. Here's a small python script that collects the metadata and creates a CSV file as result. CSV output is only for testing purposes - with some more effort, the script could output a wikitable (or even a static HTML page). The difficult part is getting a PagePile containing the uncategorized categories (doesn't even need to be a PagePile, but working on PagePiles is quite straightforward with Pywikibot). Fl.schmitt (talk) 06:01, 26 August 2024 (UTC)[reply]
So it might be quicker to use a basic query and then enriched that afterwards. Enhancing999 (talk) 07:06, 26 August 2024 (UTC)[reply]
Yes - while it's important to include the namespace as page_namespace and the category name as page_title (column aliases in SQL) - otherwise PagePile isn't able to create a valid result. Anyway, a bot would still require some time to work, since it needs to request all categories and their metadata. Depending on the server load, it seems that a bot is able to process about 6-7 Categories per Minute. So, processing ca 1,000 categories will still take between 2 and 3 hours (even when preloading the categories in batches). I'll have to check if there's a way to speed up things. Fl.schmitt (talk) 12:36, 26 August 2024 (UTC)[reply]
Then it might be easier to do it all in SQL: Quarry:history/85733/923437/896130 took about 15 minutes, but speed on quarry varies much. Enhancing999 (talk) 12:49, 26 August 2024 (UTC)[reply]
What a great query! Thanks for building this (also @Tjmj: ). Could you or somebody else fork it once more to also create a report for redcats (cats with only nonexisting categories)? Doesn't this query mean this is nearly solved? One would only need to get its result output and write it to some page on WMC and have a script do that regularly. Quite possibly there already is a tool for each of both. Prototyperspective (talk) 14:16, 26 August 2024 (UTC)[reply]
Update request (the reports were created a while ago and manually updated)

The following reports should be updated by bot:

  1. Commons:Report_Special:UncategorizedCategories (based on Quarry:query/86077, takes >10 minutes to run)
  2. Commons:Report_UncategorizedCategories_with_infobox (Quarry:query/85877, takes ∼1 minute to run)

Notes:

  • When updating, after running the query, the resulting categories need to be null-edited and then the queries run again. Otherwise we get false positives due to template based categorizations (notably {{Wikidata Infobox}}).
  • The count by user is added when it's formatted.
  • The lines should be in a template for easier formatting.
  • If it's easier to update, I could merge the two reports.
  • Ideally, the reports are updated 6AM and 6PM UTC, so Europeans and Americans don't get too many entries that have already been dealt with.

The reports may appear short now, but not too long ago they were at 4000 categories total. I think this was partially due to Special:UncategorizedCategories having ran only once a month.

The reports would be similar to w:Wikipedia:Database_reports/Uncategorized_categories.
 ∞∞ Enhancing999 (talk) 12:08, 29 September 2024 (UTC)[reply]

Please also update these new reports with a bot:
  1. Commons:Report UncategorizedCategories with redcats (Quarry:query/86864)
  2. Commons:Report UncategorizedCategories with only infobox categories (Quarry:query/86867)
I suggest that these are updated twice a month at first. Frequency could be increased as needed.
Here's how I update the reports manually (info how this is done for the two reports above doesn't seem to be included): I go to the query page click Download data and select csv. Then I open the csv in VSCodium (Visual Studio Code) and use this to add [[:Category: to the start and ]], to the end of every line as well as replacing all linebreaks. There also is a page 2 with only the first 500 items. I requested the queries here so thanks to Matěj Suchánek. Changing the output to be ordered alphabetically would improve it. "redcats" refers to nonexisting categories – further explanations are at the top of these reports.
By the way, I think the resulting categories need to be null-edited is too unclear. Prototyperspective (talk) 16:45, 7 October 2024 (UTC)[reply]

file description cleanup: "Uploaded with Reworkhelper"

[edit]

At File:Bahnhofshalle_Zuerich-2.jpg, when adding {{Information}}, I noticed:

Uploaded with ''[http://tools.wikimedia.de/~luxo/reworkhelper.html Reworkhelper] {{Wayback|url=http://tools.wikimedia.de/~luxo/reworkhelper.html |date=20080419030339 }}''

in the source. There are other similar: Special:Search/insource:"tools.wikimedia.de/~luxo/reworkhelper.html" (1970)

I'd remove this from the file description pages and replace it with a category, e.g. Category:Uploaded with Reworkhelper, similar to others at Category:Files by upload tool.

@Luxo FYI
 ∞∞ Enhancing999 (talk) 13:46, 27 August 2024 (UTC)[reply]

I have opened a bot approval request to do this. – Ammarpad (talk) 18:29, 9 September 2024 (UTC)[reply]

Add P1651 YouTube video ID structured data from "source" attribute of Filedesc template

[edit]

To assist with de-duplication of Common Criteria YouTube video uploads, a bot could search uses of Template:Filedesc for a "source" attribute that contains YouTube links in the format of "^https:\/\/www\.youtube\.com\/watch\?v=([-_0-9A-Za-z]{11})$". The captured 11 character YouTube video ID can then be added as structured data for the video file using Wikidata's P1651 property. Dhx1 (talk) 01:54, 29 August 2024 (UTC)[reply]

Also see related request (approved) for adding P12120 Flickr photo IDs as structured data at Commons:Bots/Requests/FlickypediaBackfillrBot. Dhx1 (talk) 02:00, 29 August 2024 (UTC)[reply]
I have exactly the same idea and wanted to propose something in a few days.
@Multichill Is your bot capable of adding YouTube data in SDC, and perhaps also additional data related to it? -- DaxServer (talk) 07:22, 29 August 2024 (UTC)[reply]
@DaxServer: Focus is on own work so not really doing much identifier properties at the moment. Did propose one at d:Wikidata:Property proposal/VIRIN. Maybe make it a bit more generic bot to add identifiers based on templates?
For YouTube a bit of cleanup is needed. Probably switch to something like {{From YouTube}}, which can add a tracker category so a robot has a category to work on. Same goes for {{ID-USMil}} and other templates that have an identifier property. Multichill (talk) 10:59, 1 September 2024 (UTC)[reply]
We currently have 166,884 files resulting from this search. That template doesn't mention categorization.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 15:28, 1 September 2024 (UTC)[reply]
Jeff, the template mentions that it categorizes into Category:Media from YouTube. Are we talking about the same {{From YouTube}}? -- DaxServer (talk) 18:42, 1 September 2024 (UTC)[reply]
@DaxServer: Thanks, I missed it at first.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 21:58, 4 September 2024 (UTC)[reply]
@Multichill Yes, please make the bot generic so it can be expanded with ease 🙏 -- DaxServer (talk) 20:26, 1 September 2024 (UTC)[reply]
Once files have the ID could this please also be added to Commons:Video2commons so people don't accidentally reupload videos that are already here (under a different title)? I think it could and probably should already check for the YouTube ID via an insource search for that ID doesn't it? Prototyperspective (talk) 22:29, 9 September 2024 (UTC)[reply]
Pinging @Don-vip who maintains the V2C -- DaxServer (talk) 07:48, 10 September 2024 (UTC)[reply]
Created https://github.com/toolforge/video2commons/issues/214. I'm happy to accept pull requests for this. vip (talk) 11:24, 10 September 2024 (UTC)[reply]
Once this is implemented could some tool scan for duplicate uploads using the tool please? I think this may work best if the bot adding it would simply store all IDs that it adds along with the URL (or file title or some file ID) so that after added all the IDs one can simply have it show all duplicate IDs. Prototyperspective (talk) 09:46, 28 September 2024 (UTC)[reply]
Actually the structured data is not needed and as far as I can see adds nothing directly useful. I still think it would be good to add this data. However, much more important would be to have video2commons check whether the same video has already been uploaded which can be done using the insource search operator – issue here: A scan for whether the video has been already been uploaded. Prototyperspective (talk) 18:01, 4 October 2024 (UTC)[reply]
@Dhx1 @Multichill @Jeff G. @Prototyperspective @Don-vip - How does this schema look like? -- DaxServer (talk) 18:00, 12 September 2024 (UTC)[reply]
If this looks okay, we can move forward? -- DaxServer (talk) 13:44, 18 September 2024 (UTC)[reply]
@DaxServer: YouTube handle (P11245) & YouTube channel ID (P2397) give constraint violations. The first looks like a good qualifier so I'll update the constraints, just not sure about YouTube channel ID (P2397). Multichill (talk) 21:26, 18 September 2024 (UTC)[reply]
@Multichill Reading the https://support.google.com/youtube/answer/11585688, it seems the channel IDs are the ones that are persistent while the handles can be changed and can be claimed by others. If we were to add the handle and not channel ID and if the handle is relinquished and claimed by another, we would be directing to the wrong creator. I think channel ID should be added, not sure about the handle -- DaxServer (talk) 07:44, 19 September 2024 (UTC)[reply]

Add missing Template:Location

[edit]

Some images seem to have coordinates in SDC, but not display them on the file description page.

Special:Search/haswbstatement:P1259 -hastemplate:"Module:Coordinates" doesn't find them. Is there way to complete them?
 ∞∞ Enhancing999 (talk) 15:14, 29 August 2024 (UTC)[reply]

@Enhancing999: completing the file description isn't too difficult once the files are found. Fl.schmitt (talk) 17:53, 1 September 2024 (UTC)[reply]
I'm trying to figure out where I saw it. I think it was Flickrbackfillerbot who had added the coordinates. As they weren't visible on the description, I had added some myself and then the template complained about a differences.
 ∞∞ Enhancing999 (talk) 17:55, 1 September 2024 (UTC)[reply]
@Enhancing999: BTW: If there are SDC coordinates, it seems to be sufficient to add empty {{Location}}/{{Object location}} templates - see test at Revision #918241588. Fl.schmitt (talk) 18:35, 1 September 2024 (UTC)[reply]
Yes, the question is to which files.
 ∞∞ Enhancing999 (talk) 18:41, 1 September 2024 (UTC)[reply]
One brute-force solution is to parse the database dumps - both the wikibase entities and wikitext content - and determine the files -- DaxServer (talk) 18:50, 1 September 2024 (UTC)[reply]
Maybe we need to have MediaWiki re-add Category:Pages with coordinates to files.
 ∞∞ Enhancing999 (talk) 18:55, 1 September 2024 (UTC)[reply]
@Enhancing999 / @DaxServer: Missing {{Location}} seems to be quite often the case for imports from Flickr. Analyzing one randomly chosen multistream dump archive, containing 918,128 file pages, there are 5,228 file pages with coordinates of the point of view (P1259) claim missing a {{Location}} template: PetScan #29249152 / Pagepile #60582. Fl.schmitt (talk) 18:50, 9 September 2024 (UTC)[reply]
Thanks. I struggled finding my sample again.
Sample from that list: Boeing 757-200, Diocletians_Palace
It also includes some with really "special" templates: File:Wym-bykaont (wersja Józefa Gary).ogg, File:Davy Mourier - Cartoonist 2013 - P1600722.jpg that do display them.
 ∞∞ Enhancing999 (talk) 19:05, 9 September 2024 (UTC)[reply]
OK :-) then please subtract 3,176 files pages from the "Wymysorys pronunciation" category from my 5,228. Obviously, I've hit a special dump... but why is the location visible in Davy Mourier - Cartoonist 2013 - P1600722.jpg, while it isn't visible in Diocletians Palace? Both are using the {{Information}} template? Fl.schmitt (talk) 19:19, 9 September 2024 (UTC)[reply]
There is: Template:Festival Cartoonist 2013. Go figure. Maybe location should be removed from that template.
 ∞∞ Enhancing999 (talk) 19:22, 9 September 2024 (UTC)[reply]
Ok. As far as I see, the dump doesn't have the templates transcluded. Cases like Template:Festival Cartoonist 2013 can be detected only if those templates containing coords are known beforehand. Thus, scanning the dumps won't yield a complete result set. So I would propose to restrict the task on files with {{Information}} template, SDC coordinates of the point of view (P1259) and missing {{Location}}.
Is there consent to solve the issue in those cases by adding {{Location}}? Does this need additional discussion? Alternatively, since Module:Information already uses locality of creation (P1071) and location (P276), would it be realistic to get coordinates of the point of view (P1259) added to Module:Information, too? This would solve the issue without the need of a bot and also for the future. Fl.schmitt (talk) 16:10, 11 September 2024 (UTC)[reply]
@Enhancing999: see Module_talk:Information#Proposal:_Evaluate_SDC_P1259_(coordinates_of_the_point_of_view) - let's wait and see... Fl.schmitt (talk) 19:57, 20 September 2024 (UTC)[reply]
The 1% missing mentioned there seems a lot that we are missing. I still favor adding {{Location}} directly, {{Information}} might take a long time to update or change.
Template:Location from Template:Festival Cartoonist 2013 might be in the template links table if it's correctly refreshed. In rare cases having two location templates shouldn't be the end of the world.
Also, shall we have MediaWiki add Category:Pages with coordinates again?
 ∞∞ Enhancing999 (talk) 12:37, 21 September 2024 (UTC)[reply]

Monuments database in Russia

[edit]

Per discussion at Commons:Village pump#Monuments database in Russia.

There are >25K sub-categories of Category:Galleries of cultural heritage monuments in Russia (and about 275 in its subcategory, Category:Galleries of cultural heritage monuments in Crimea) named in the format (for example) Category:WLM/1010021052. That example duplicates Category:Threshing barn from Berezovaya Selga. The corresponding Wikidata item, Threshing barn from Berezovaya Selga (Q106488771), has a Wiki Loves Monuments ID (P2186) value of RU-1010021052 (note the "RU-" prefix). That Wikidata item is linked to the alphanumerically named, not numbered, category.

For each of those 25K categories, we need a bot to do the following:

  1. Find the Wikidata item with the Wiki Loves Monuments ID (P2186) value (e.g. RU-1010021052)
    1. If no Wikidata item is found, write a log entry and skip to the next category
  2. Find the Commons category that the Wikidata item is linked to
    1. If no Commons category is found; or if the linked category is of the numeric type, write a log entry and skip to the next category
  3. Redirect the numeric category (e.g. Category:WLM/1010021052) to the latter category (e.g. Category:Threshing barn from Berezovaya Selga)
  4. Ensure that the latter category transcludes {{Wikidata infobox}}

An alternative at 1.1 would be to create a Wikidata item; populating with data from e.g. https://ru-monuments.toolforge.org/wikivoyage.php?id=1010021052 - but this could be done later. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:14, 24 September 2024 (UTC)[reply]

Redirect Galleries per Concensus

[edit]

Per the proposal here: Commons:Village_pump/Proposals#Summary_Proposal_by_Alachuckthebuck, we are approaching the point where the task could be automated. It will require a bot to see if the gallery title matches a category name, and can be redirected without effecting the countervandelism and patrolling tools, Sadads (talk) 12:26, 30 September 2024 (UTC)[reply]

Automatically removed closed requests from Deletion requests/mobile tracking

[edit]

When I started a deletion request it got automatically added to the bottom of the page, which had gotten so long that the latest requests were no longer loading in properly (you can scroll to the bottom to see what that looked like). I went ahead and removed the closed requests manually, but this seems like the sort of thing bots could take care of as well. ReneeWrites (talk) 15:12, 30 September 2024 (UTC)[reply]

@ReneeWrites: You could ask Krd if some of his bot's functionality could help with this. Archival of the page has been a problem for 11 years.   — 🇺🇦Jeff G. please ping or talk to me🇺🇦 15:25, 30 September 2024 (UTC)[reply]
✓ Done --Krd 18:38, 1 October 2024 (UTC)[reply]
Checkmark This section is resolved and can be archived. If you disagree, replace this template with your comment. --Krd 18:38, 1 October 2024 (UTC)

Remove double headings created due to a mass-message

[edit]

I sent a POTY message to targetes listed here, and unfortunately the heading was duplicated. Most likely because it existed on the "page to be sent as message" and the message tool asked to supply a subject line. Could someone run a bot and have these duplicate instance of the heading removed? Regards, Aafi (talk) 17:33, 1 October 2024 (UTC)[reply]

@Aafi: Revision #931176428 Is this ok? Fl.schmitt (talk) 18:01, 1 October 2024 (UTC)[reply]
If yes, I can run the same edit on the other affected User talk pages. Fl.schmitt (talk) 18:01, 1 October 2024 (UTC)[reply]
@Fl.schmitt, yes. This perfectly fixes the issue. Regards, Aafi (talk) 18:09, 1 October 2024 (UTC)[reply]
Great - I've started running pwb replace, may take some while to complete.
Checkmark This section is resolved and can be archived. If you disagree, replace this template with your comment. Fl.schmitt (talk) 18:18, 1 October 2024 (UTC)
Fl.schmitt (talk) 18:18, 1 October 2024 (UTC)[reply]
Since it's more than 3k pages, i'll file a bot request - this will take some additional time since i'll have to wait for the new bot account to be autoconfirmed. Fl.schmitt (talk) 18:43, 1 October 2024 (UTC)[reply]