Wikidata

Wikidata:Property proposal/Flickr Photo ID

< Wikidata:Property proposal

‎Flickr Photo ID

edit

Originally proposed at Wikidata:Property proposal/Creative work

DescriptionPersistent identifier for images on Flickr (Q103204)
Data typeExternal identifier
Domainitem, image (Q478798), video (Q98069877)
Allowed values[0-9]+
Example 1https://commons.wikimedia.org/wiki/File:Il_Duomo_Siena.jpg Flickr Photo ID → 5861347
Example 2https://commons.wikimedia.org/wiki/File:Oedura_lesueurii.jpg Flickr Photo ID → 104409253
Example 3https://commons.wikimedia.org/wiki/File:Gjende_-_panoramio_(1).jpg Flickr Photo ID → 29916169
Sourcehttps://flickr.com
Expected completenessalways incomplete (Q21873886)
Formatter URLhttps://www.flickr.com/photo.gne?id=$1
See alsoFlickr user ID (P3267), image (P18)
Single-value constraintyes

Motivation

edit

We are creating Version 2.0 of the Flickr2Commons tool called Flickypedia. More than 5.4 million files have been uploaded to Wikimedia Commons using Flickr2Commons.

Having specific Wikidata IDs for the Flickr images that Flickypedia uploads to Wikimedia Commons will keep the source image connected using structured data. This has two primary benefits:

  1. Help us check for license washing (which was brought up as a community concern with this type of importing tool).
  2. Allow Wikimedia Commons to more easily check for duplicates which came up on the Community Wishlist for this year.

The Flickr Photo ID can be extracted using complex queries from existing data, but since Flickr images do not exist at a canonical URL (see below) a Flickr Photo ID will greatly simplify this process.

A previous proposal in 2017 was not approved – there were concerns about how to choose the “best” image to represent a given Wikidata entity, and concerns about linking to non-free, non-Commons images. This proposal is different in that we will use this ID for files that have already passed an initial license-check via the Flickypedia software, and are therefore appropriate for Wikimedia Commons.

It would also be possible to programmatically backfill this field, and other structured data, on the 10 million plus existing Flickr images in Wikimedia Commons.

A similar example of item-level identifier usage already exists in Wikidata: YouTube video ID (P1651).

Examples of differing URLs for Flickr images

These are all examples of URL variants in the existing structured data fields:

  • Flickr username, without slash: https://commons.wikimedia.org/wiki/File:Gjende_near_Memurubu.jpg  https://www.flickr.com/photos/sunrise/29916169
  • Flickr username, with slash: https://commons.wikimedia.org/wiki/File:Albufeira,_Portugal_(52557097207).jpg https://www.flickr.com/photos/rickmccharles/52557097207/
  • Numeric user ID, with slash: https://commons.wikimedia.org/wiki/File:Oedura_lesueurii.jpg https://www.flickr.com/photos/55566249@N00/104409253/
  • Numeric user ID, without slash: https://commons.wikimedia.org/wiki/File:Il_Duomo_Siena.jpg  https://www.flickr.com/photos/42033648@N00/5861347

There are more variants in the free text/unstructured data, which are harder to search for (and may be why other bots have yet to backfill a Flickr URL in their structured data):

  • The actual JPEG, rather than the photo page: https://commons.wikimedia.org/wiki/File:Thalia.jpg http://farm4.static.flickr.com/3009/2319323522_20143af300_o_d.jpg
  • A different variant of raw JPEG: https://commons.wikimedia.org/wiki/File:Haus_Wohnung2.jpg http://photos22.flickr.com/29216501_6c9c5e111e_b.jpg
  • Flickr’s short URL: https://commons.wikimedia.org/wiki/File:Lada.riva.akadruid.jpg http://flic.kr/p/dL8zS7 (the last component of this, e.g. dL8zS7, is the base-58-encoded photo ID)

Jessamyn - Flickr Foundation (my talk page) 15:43, 17 October 2023 (UTC)[reply]

Discussion

edit
  •   Support Sounds like a useful property! I added a few links above, and added a note about the fact that the short URL can be derived from the ID (and vice-versa). Also changed the formatter URL to what's commonly used for linking when only the ID is known (although I don't know if there's a more canonical way of doing this; this one isn't mentioned in the API docs, but no alternative is either). Sam Wilson 01:01, 18 October 2023 (UTC)[reply]
  •   Support I have been doing a lot of work to backfill various types of SDC on Commons, and having higher-quality programmatic workflows to attach the IDS in SDC makes sense to me, especially as we think about bringing over more mutli-lingual descriptive depicts and other activities, Sadads (talk) 13:08, 18 October 2023 (UTC)[reply]
  •   Support I have been adding machine readable tags to the Biodiversity Heritage Library images in Flickr. Many of these images have already been uploaded to Wikimedia Commons and so those Flickr tags have not yet been added to the image metadata in Wikimedia commons. If this property is approved we can link the image in Wikimedia Commons to the image in Flickr via structured data in commons statements which will then make it easier to import the Flickr machine readable tags into Wikicommons, improving the metadata describing the image, which in turn will make the image easier to find and reuse.Ambrosia10 (talk) 20:10, 19 October 2023 (UTC)[reply]
  •   Support - very useful identifier to use in sourcing in structured data on Commons. Spinster 💬 07:18, 20 October 2023 (UTC)[reply]
Yep, my bad, removed now. Regards Kirilloparma (talk) 03:32, 25 October 2023 (UTC)[reply]
Thanks! Sam Wilson 06:40, 25 October 2023 (UTC)[reply]
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Property_proposal/Flickr_Photo_ID&oldid=1996722141"