News post idea: Working with Structured Data on Commons

I’ve been thinking about writing a post describing some ways to work with Structured Data on Commons, from a community perspective. Contents could include:

  • What’s already possible
    • Manual editing (duh)
    • search
    • Direct API usage, for experts (compare Multichill’s T223746)
    • QuickStatements (very rudimentary support)
    • a user script I hope to publish within the next few days published a few days ago, AC/DC
  • What’s coming soon
    • SPARQL querying (T141602)
    • proper QuickStatements support, hopefully (T181062)
    • Pywikibot support, hopefully (T223820)
    • Pattypan support, hopefully (T181057)

Would anybody be interested in co-authoring this post?

9 Likes

I definitely like the idea. I’m happy to help review and edit, I’m not so sure about co-authoring as the topic is a little technically beyond me. @SandraF_WMF perhaps you might be interested in helping out here? I know your were thinking about writing something similar (if not so technical).

4 Likes

I can help review it, but deep technicalities will be beyond me - I’m not a coder :slight_smile:

QuickStatements has very rudimentary support already? How does that work? I’d love to see an example of that! :smiley:

2 Likes

QuickStatements support really is very rudimentary – I’m not sure if it should even be mentioned, but here’s the current state:

  • It kinda works, if you select “Commons [does not work yet]” as the target wiki and enter MediaInfo entity IDs in the first column (M123…) – example syntax
  • Background mode only makes about ¼ to ⅓ of edits, for no reason I could discern, avoid it
  • Foreground mode makes all edits, but does not check for duplicate statements like it’s supposed to, you’ll have to clean those up yourself (I still have to do that for my test batch, because the edits from background mode are now duplicates)

You can see all the QuickStatements edits in the recent changes (so far it’s only me).

If you have suggestions for things I’ve missed, that would also be welcome – for example, earlier today I happened to see a Phabricator task for deleted depicts not working in ISA (T229801), from which I conclude that ISA has some structured data support and I should add it to the “what’s already possible” list :D

2 Likes

ISA is under very active development at this moment and the developers are ironing out the last bugs :smile: … it would be absolutely great if you could mention it! https://commons.wikimedia.org/wiki/Commons:ISA_Tool

1 Like

Oh, and I think the Wikimedia Commons Android app also has support for editing structured data. I don’t own an Android phone, so have been unable to test this myself.

This is a great idea :+1:

I don’t see depicts or any SDC features in v2.10.2 of the Commons App, but I see there is good active development on GitHub.

I would so love to add “depicts” statements in batch mode when I am already adding a new category to search results using cat-a-lot. Usually I am doing this because I have just added a new concept to Wikidata, so I know nothing depicts it yet (for example “sauce tureen” just last week).

2 Likes

Alright, I’ve written a first version of the post and submitted it for review – I think you can also see it as post #1035 if you’re logged in? (Otherwise I can copy it to Google Docs or wherever so you can still take a look.)

@PKM The final paragraph kind of addresses your request – the user script I mentioned in the original post (published by now: AC/DC) can be used to add statements to a set of files, and I’m hoping to add better integration with Cat-a-lot soon.

2 Likes

Thanks, @LucasWerkmeister, you rock! I’ll try out AC/DC.

1 Like

@LucasWerkmeister, the draft looks nice. I appreciate you writing it.

Before we schedule to publish, do you want to pick a photo to accompany the article? Feel free to add one if you’d like. Regardless I think we can schedule this to publish when you’re ready.

@Ckoerner-WMF thanks!

I don’t have any great idea for a photo… the editorial guidelines tell me to “[l]eave screenshots, graphs, and logos for the middle of the post, please”, so I guess it’ll have to be some kind of symbolic image for “working with structure”? I found a few candidates on Commons:

The last image doesn’t quite have the minimal width of 1024px (it’s 902×599), but it’s still my favorite – it’s much closer to the subject than the others (though I do like the cluttered one as well, as a metaphor :wink:). Any thoughts?

(Also, I wrote most of this post en route to Wikimania, with little to no internet – now that I’m in a better place, I could probably add screenshots of some of the things mentioned in it. Would that be useful?)

Apart from the images, the post is ready to go as far as I’m concerned, and given that it currently speaks of Wikimania as a future thing, it would be great if we could publish it as soon as possible, otherwise I’ll have to update those parts. (I should’ve started the draft earlier, sorry.)

Ha, I suppose we should be more clear there. We want authors so suggest a primary photo. Screenshots, graphs, and logos to be in the body of the image, but not used as a primary “eye catching” photo. I’ll see if I can update that to make it more clear.

Let’s go with your favorite, It’s pretty closet to the minimal width and should look fine for our purposes. I added it to the draft. If you have no complaints I can publish.

(The cluttered one is funny, but I think we want folks to feel like working with Structured Data is a little more organized. :slight_smile: )

Well, it’s too late for a “before Wikimania” post now and I also want to incorporate some things mentioned in Sandra’s presentations, so I’ll rework the blog post over the next week and get back to you.

I’ve added paragraphs on the Wikipedia Android app and upcoming Lua support and updated some other bits as well – the post is now a bit more timeless, and again ready for publication IMHO :)

Hey @LucasWerkmeister! We had a look at your post in the Editorial Board meeting today and it looks great and ready to go. It’s scheduled for publication on Tuesday and will go out then, unless you have another preference for timing. Just let me know!

2 Likes

Great, thanks a lot! Tuesday is fine by me.

The post is live on space and discuss-space \o/

3 Likes