Commons talk:Structured data/Archive 2024

Aerial photographs

How do you identify whether an aerial photo was taken from an airplane, a balloon or a drone? XRay 💬 19:20, 1 February 2024 (UTC)

By "identify" I assume you mean "indicate (if known)", not how you work it out from looking at the photo (which often you can't). - Jmabel ! talk 20:28, 1 February 2024 (UTC)

You're right, indicate. --XRay 💬 20:38, 1 February 2024 (UTC)

I'm a little surprised we don't have a property corresponding to photographic technique (Q1439691). Then you could set the value of that to drone photography (Q12033087) (for example). Can anyone see a way to do an an equivalent of that? - Jmabel ! talk 20:36, 1 February 2024 (UTC)

Please have a look to Category:Aerial photographs by camera vehicle. How to model camera vehicle? --XRay 💬 20:42, 1 February 2024 (UTC)

Slightly off topic, but shouldn't Category:Aerial photographs by Smart Drone‎ be a subcat of Category:Unmanned aerial vehicle aerial photography? - Jmabel ! talk 01:56, 2 February 2024 (UTC)

@XRay: It has been suggested on Wikidata that we could use genre (P136) with drone photography (Q12033087). - Jmabel ! talk 16:26, 4 February 2024 (UTC)

This is logical and understandable. However, another value that is often used is more general: aerial photography (Q191839). You can't tell from the value whether it was a drone, an airplane, a hot air balloon or something else. --XRay 💬 18:14, 4 February 2024 (UTC)

captured with (P4082) exists for drones and the like. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:34, 4 February 2024 (UTC)

How to depict flexed muscles with expression, gesture or body pose (P6022)?

How do you recommend modelling flexed muscles with expression, gesture or body pose (P6022) for photos like this one or other photos in Category:Biceps brachii? ? Lovelano (talk) 03:12, 5 March 2024 (UTC)

instead of p6022, try

depicts (P180)	Erin Cummings edit shown with features (P1354) supination of the forearm 1 reference inferred from (P3452) bicep curl add reference
		add value

perhaps? Arlo James Barnes 08:14, 8 March 2024 (UTC)

Some more examples might be helpful.

I've been adding SDC, but I'm really not sure if I'm doing it right. Could I have some feedback, and is there a place for more examples? (I looked at Commons:Structured data/Modeling/Depiction.)

See File:Cape Air ticket counter at MWA airport.jpg; I've described it as a ticket counter, which is part of the specific airport. Is that the best way to express this?

I've added a captured with statement for the drone model in File:Caspersen Beach aerial view.jpg, but this is showing a warning. An I doing it wrong, or should this be automatically populated from EXIF or categories like Category:Taken with DJI FC220 anyway? (Is that happening in some cases but not others?)

For construction equipment like File:Caterpillar backhoe loader at construction site in Sunnyvale, back view.jpg, I get a warning when I say it depicts the generic type of equipment, and specify the make as the manufacturing company. I'm also not sure how to specify the model number. See also File:Google self-driving Lexus RX450h rear view.gk.jpg; specifying the make and model of a car seems reasonable, but I get a warning with the make and didn't know how to specify the model.

Thanks in advance to anyone who can help. I'm intrigued by the tools, but I feel like I'm using them... not quite right. grendel|khan 11:11, 15 March 2024 (UTC)

what property to mark a file that contains use of german sign language?

How do I add info about use of german sign language to a foto or to a video? C.Suthorn (@Life_is@no-pony.farm - p7.ee/p) (talk) 17:18, 25 December 2023 (UTC)

1. Click "Add statement".

2. Add property language of work or name (P407).

3. Add item German Sign Language (Q33282).

Lovelano (talk) 21:25, 4 March 2024 (UTC)

done, https://commons.wikimedia.org/w/index.php?title=File:Diversity_2023_25_(video_capture_001).png&diff=prev&oldid=861811635 ff. C.Suthorn (@Life_is@no-pony.farm - p7.ee/p) (talk) 13:59, 19 March 2024 (UTC)

Garbage "tags" still being added in bulk

So this is what the universe looks like. And NASA.

Well-intentioned users are still being encouraged by apps to stuff the depicts statement with garbage "tag" values.

For instance this recent edit tagged the above image of part of the surface of the Earth from the ISS as depicting the ISS, "Expedition 22", "outer space", NASA, and the universe. The edit summary includes #suggestededit-add-tag 1.0" and the edit is tagged "Mobile edit Mobile app edit Android app edit Suggested Edits edit".

How do we get this stopped? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:21, 10 March 2024 (UTC)

Seconding this request. I find myself clearing crap like this off of the file pages of my pictures every day or two. How much of this must be accumulating on file pages that no one is watching? - Jmabel ! talk 18:33, 10 March 2024 (UTC)

Strange, I never had crap tags added to the files I uploaded (not to say it is impossible, but it never happened to my pictures). Ymblanter (talk) 18:35, 10 March 2024 (UTC)

Better instructions in the apps or a visual example? I'm speculating here but if people see a blank page they have no idea how to use correct properties so something to steer people towards right approach would be useful? Let's say people want to add "creation date" but the property they need is called "inception" - something to steer people towards right property might be useful. Another thing entirely is how well-made the translations for different languages are.. Ipr1 (talk) 21:29, 21 March 2024 (UTC)

Imagine being a new user and you see an empty box on the page, so maybe you would assume that is where you put things and magic happens. Adding a statement? A claim? A qualifier? It takes a more experienced users to know what those are and how they relate to values they see on the page. Ipr1 (talk) 21:40, 21 March 2024 (UTC)

I frequently have crap tags added to my pictures, which I remove.

It isn't a matter of better teaching clueless newbies what is a good value to add as a "depicts" on someone else's picture. It's a matter of not encouraging clueless newbies to add these. - Jmabel ! talk 07:01, 22 March 2024 (UTC)

Subtitles hardcoded into video track？

Secretary Janet Yellen on the Child Tax Credit

maybe there should be some properties about subtitles embedded in the video track? in contrast to subs from timed text files? RZuo (talk) 19:31, 23 March 2024 (UTC)

What properties to use to state conservation status and types of damage of old photographs

Hello! I'm trying to add structured data to several old photographs of my city we've uploaded to Commons, like this one: https://commons.wikimedia.org/wiki/File:Plaza_Constituci%C3%B3n_-_0001FMHC.tif I have a big dataset several datapoints available, including "conservation status" (good, bad or regular) and "types of damage" (like silver mirroring Q112581162, mutilation, scratchs, inscriptions...). Apparently all these data follows the General International Standard Archival Description conventions, but I still can figure out which Wikidata properties should I use. Any ideas? Thanks! Piracalamina (talk) 15:58, 4 April 2024 (UTC)

Tool to copy sdc from file to file

is there one for problems like https://commons.wikimedia.org/w/index.php?title=Commons_talk:CropTool&oldid=867301107#Structured_Data ? we need a tool like this. RZuo (talk) 07:31, 12 April 2024 (UTC)

Maybe d:MediaWiki:Gadget-moveClaim.js could be adjusted to do this, @Matěj Suchánek, Melderick? --Marsupium (talk) 16:22, 15 April 2024 (UTC)

Aerial point of view

What would be the appropriate way to depict with structured data that a photograph was taken from an aerial point of view? Sdkb ^talk 15:54, 16 April 2024 (UTC)

Commons_talk:Structured_data/Modeling/Location#Aerial_photographs have some dicussion related to this. -- Zache (talk) 17:10, 16 April 2024 (UTC)

Thanks for the link! Let's centralize discussion there. Sdkb ^talk 18:38, 16 April 2024 (UTC)

Maintenance category broken?

Quality images missing SDC Commons quality assessment currently has over 14,000 files in it. But from a random sample of 15 files in this category, I found 9 to actually already include the statement. Clearing the category's cache didn't help. This makes that category not really helpful any more. Any ideas, what could cause this problem?

@Mike Peel, @Dschwen: could QICbot possibly set Commons quality assessment (P6731) Wikimedia Commons quality image (Q63348069) upon promotion and make this maintenance category obsolete? MB-one (talk) 10:34, 11 March 2024 (UTC)

This was being done by @Multichill's bot, I'm not sure why that stopped? I haven't figured out how to set SDC values using pywikibot/the API yet, it's on my to-do list, but I'm not sure when I'll get to it. Thanks. Mike Peel (talk) 11:27, 11 March 2024 (UTC)

@Mike Peel Maybe that helps you to get started: User:SchlurcherBot/commonsapiaddclaimsoauth Let me know anything else you need. --Schlurcher (talk) 06:55, 23 May 2024 (UTC)

Indeed https://commons.wikimedia.org/w/index.php?sort=last_edit_desc&search=-haswbstatement%3AP6731+hastemplate%3AQualityImage+filetype%3Abitmap&title=Special%3ASearch&profile=advanced&fulltext=1&advancedSearch-current=%7B%22fields%22%3A%7B%22hastemplate%22%3A%5B%22QualityImage%22%5D%2C%22filetype%22%3A%22bitmap%22%7D%7D&ns6=1 shows far fewer results and no false positives so far. MB-one (talk) 18:46, 11 March 2024 (UTC)

@MB-one and Mike Peel: Jobs User:BotMultichillT used to run have been disabled by the Wikimedia Foundation, see phab:T319912. Didn't really have the motivation to fix it yet. Multichill (talk) 17:40, 15 March 2024 (UTC)

@Multichill: so do I understand correctly that you decided that with this change of technical environment, you are abandoning all your bot tools and have no interest in migrating them? Are you willing to work at all with some other user who might be interested in migrating them? I suspect there are a fair number of these someone would like not to lose. - Jmabel ! talk 18:40, 15 March 2024 (UTC)

I've had to convert several of my bot scripts over to the new system (e.g., setup script for the venv, then new job submission script, with the run script loading the venv), would be happy to help here if you need. Although, this is why I run most of my scripts on an environment I have full control over... Thanks. Mike Peel (talk) 18:59, 15 March 2024 (UTC)

Update: Multichill's bot BotMultichillT has started thankfully started to tag images in Quality images. But Quality images missing SDC Commons quality assessment still lists about 5,000 files at any given time, although the actual number of files missing this tag hasn't exceeded around 150 in a month now. I wonder why.--MB-one (talk) 13:37, 22 May 2024 (UTC)

@MB-one: There seems to be something amiss due to Template:Quality image and its categorisation and the data it pulls from the WD ether. The metadata may need to be refreshed, however, it is so not visible and is beyond the scope of normal people to push purge, view, manipulate.  — billinghurst sDrewth 02:04, 23 May 2024 (UTC)

Based on my experience you have to perform a null edit as a purge is not sufficient. I can do this ealy next week for the category. --Schlurcher (talk) 06:51, 23 May 2024 (UTC)

Yeah, I'm slowly recovering some of the bots. Category:Quality images missing SDC Commons quality assessment is empty now and looks like all quality images have the statement. Multichill (talk) 11:42, 24 May 2024 (UTC)

Proposal to add perceptual hashes to SDC

Just FYI, I made a Village pump proposal for adding pHash checksum (P9310) and Imagehash difference hash (P12563) values to commons images. Currently, FinnaUploadBot have been adding these only to images uploaded from Finna.fi --Zache (talk) 17:00, 17 May 2024 (UTC)

So this is essentially a front-end to Wikidata?

Do I understand this correctly: every file on Commons is also a node (item?) on Wikidata, in the sense that it is assigned a QID, and thus you can start making statements about it — the structured data tab is essentially a front-end to the Wikidata database? The main difference to editing statements on Wikidata appears to be that this interface does not show the subject QID anywhere! Or are you using other URIs as resource identifiers? (Being explicit with such technical details would have helped me understanding what this was all about.)

Do you have any examples of statements relating one file to another? (I didn't encounter anything like that, browsing a few of the examples in the list of properties.) Practical application: relating different BSicons to each other. 88.129.117.158 14:59, 7 July 2024 (UTC)

No, each image does not have a QID. Each media item instead has an identifier consisting of "M" followed by its page ID, but as you say the interface mostly hides this. I don't think there's any way to relate files to one another directly; I've certainly never seen one. Notably, when Wikidata wants to refer to Commons files, it does it by name and not by using the "M" IDs. --bjh21 (talk) 15:31, 7 July 2024 (UTC)

Bot to remove duplicate statements

Hi folks, sometimes we end up with duplicate statements on files. This often happens when bots edit the same file twice. I set up bot to clean that up about once a month (code). It goes through the latest dump at https://dumps.wikimedia.org/commonswiki/entities/ (the copy on Toolforge) and removes any duplicate statements. It hashes the statement for quick processing so it doesn't catch cases (yet) like this one where the qualifiers are in different order. Multichill (talk) 20:29, 7 July 2024 (UTC)

List of bots involved

i think it'd be good for information, reference and documentation purposes to have a page that lists all bots and how they are involved in editing sdc. then we will know, whether a specific task in mind is already taken charge of, or if any bot malfunctions which tasks are not carried out, etc. RZuo (talk) 10:41, 6 August 2024 (UTC)

for example, i want to know now whether some bot is already in charge of migrating accidental addition of coordinate location (P625) to location of creation (P1071). RZuo (talk) 10:43, 6 August 2024 (UTC)

See Commons:Structured_data/Modeling#Bots --XRay 💬 10:55, 6 August 2024 (UTC)

thx a lot.

also i made a mistake above. should be moving p625 to coordinates of the point of view (P1259) instead of p1071. RZuo (talk) 11:57, 6 August 2024 (UTC)

Cannot find haswbstatement:P1259

https://commons.wikimedia.org/w/index.php?search=haswbstatement:P1259 shows nothing despite File:Martin-Luther-Kirche Behringen 2024-04-09 03.jpg for example. why? this shouldnt be expected, right? RZuo (talk) 11:58, 6 August 2024 (UTC)

@RZuo: Ah yes, I remember looking at this before. I think CirrusSearch can only see properties whose values are text or Wikidata items. If you use the API to get cirrusbuilddoc for that page (API sandbox link) you'll find that statement_keywords contains source of file (P7482), copyright license (P275), copyright status (P6216), captured with (P4082), media type (P1163), and checksum (P4092), but not creator (P170), inception (P571), data size (P3575), or coordinates of the point of view (P1259). bjh21 (talk) 17:37, 6 August 2024 (UTC)

Tracked in Phabricator
Task T371929

thx a lot for the tips! RZuo (talk) 19:36, 6 August 2024 (UTC)

You can have a look to the cirrusdump: [1] --XRay 💬 20:04, 6 August 2024 (UTC)

Inception for files where only Upload date is available

Is there a preferred way to add SDC inception (P571) claims for files with unknown creation date, but known {{Upload date}}? As far as I see, there's no "upload date" qualifier on Wikidata. Is it ok to simply use the upload date for inception (P571)? Fl.schmitt (talk) 20:30, 22 August 2024 (UTC)

@Fl.schmitt: how about "unknown value" with qualifier latest date (P1326) and the upload date. - Jmabel ! talk 04:30, 23 August 2024 (UTC)

@Jmabel - that's a good solution! I've used it already at Karte Bodensee Birnau.png - and {{Information}} is able to use latest date (P1326) as default if there's no date parameter - great! I wasn't able to set "unknown date" manually, but pywikibot was able to create such a claim with snaktype "somevalue". Fl.schmitt (talk) 07:16, 23 August 2024 (UTC)

VIRIN property proposal

See d:Wikidata:Property proposal/VIRIN. Multichill (talk) 13:27, 24 August 2024 (UTC)

Test environment for SDC?

Is there a installation for testing SDC? I found https://test.wikipedia.org but that doesn't have SDC. https://test-commons.wikimedia.org seems to be closed.
 ∞∞ Enhancing999 (talk) 19:06, 2 September 2024 (UTC)

On https://commons.wikimedia.beta.wmflabs.org many testing should be possible but you might need to request rights on Beta Wikidata to create properties. GPSLeo (talk) 19:31, 2 September 2024 (UTC)

YouTube SDC

Please see Commons:Bots/Work requests#Add P1651 YouTube video ID structured data from "source" attribute of Filedesc template for the sample schema, which is more or less identical to Commons:Structured data/Modeling#Upload from a platform like Panoramio, Geograph or Flickr -- DaxServer (talk) 10:32, 13 September 2024 (UTC)