Page MenuHomePhabricator

commons file search that includes structured data
Open, MediumPublic

Event Timeline

JanZerebecki raised the priority of this task from to Needs Triage.
JanZerebecki updated the task description. (Show Details)
JanZerebecki added a project: SDC General.
JanZerebecki changed Security from none to None.
JanZerebecki subscribed.
Gilles triaged this task as Medium priority.Nov 26 2014, 4:11 PM
Gilles subscribed.

Here are some things that the Commons wants to be able to search/filter search results by:

  • metadata
  • file type
  • file size
  • dimensions (maybe aspect ratio?)
  • upload date

Here are some things that the Commons wants to be able to search/filter search results by:

  • metadata
  • file type
  • file size
  • dimensions (maybe aspect ratio?)
  • upload date

I think location would be very useful as well and should be pretty easy to do, at least for files that have coordinates attached. Example: find all pictures of bicycles taken in Italy. We'd need a set of geo polygons, probably stored at Commons (which is possible now) that are linked to geographical entities like countries via Wikidata. Such data is easily available from http://naturalearthdata.com/ (CC-0).

Having that geographical data would probably be useful to Wikivoyage, too. I wonder whether other projects could benefit from it. Maybe {{requested photos}} at the Wikipedias? And I could easily imagine some editors at Wikipedia (or other projects) wanting to search first for images from their own region: 'apples grown nearby' vs 'apples'.

Here are some things that the Commons wants to be able to search/filter search results by:

  • metadata
  • file type
  • file size
  • dimensions (maybe aspect ratio?)
  • upload date

Some of these things are searchable already; see https://www.mediawiki.org/wiki/Help:CirrusSearch#File_properties_search. The Structured Data on Commons will add a lot more structured data, which can then be picked up by the search engine.

Having that geographical data would probably be useful to Wikivoyage, too. I wonder whether other projects could benefit from it. Maybe {{requested photos}} at the Wikipedias? And I could easily imagine some editors at Wikipedia (or other projects) wanting to search first for images from their own region: 'apples grown nearby' vs 'apples'.

That kind of data could probably used instead of those location maps on Wikipedias as well. It would need a bit of coordination across the projects (<s>new property at Wikidata</s>, Stuff in the Data: namespace at Commons should probably look more like file description pages and needs better documentation, integration with StructuredData). I've got some ideas for that, but the over-all system should probably be discussed elsewhere, (I'm not sure where, though).

Even media without location data in form of coordinates could be searched for if the location was stored like "this picture was taken in Rome" in StructuredData (current Categories won't really cut it here) and connected to the Wikidata item for Rome (which in turn has the information that it is "located in the administrative territorial entity" of the Province of Rome, which lies in Lazio, which lies in Italy etc.)

Edit: strike the wikidata property: already exists at P3896