Community Wishlist Survey 2019/Multimedia and Commons/Support for IIIF Presentation API in WikiCommons

From Meta, a Wikimedia project coordination wiki

Support for IIIF Presentation API in WikiCommons

  • Problem: IIIF was created to allow the sharing and interoperability of high quality zoomable images both for viewing but also for annotating. Version 3 of the presentation API now also supports audio and video. This proposal is for WikiCommons to support the IIIF Presentation API by exposing IIIF manifests and coining static Canvas IDs. A IIIF Manifest is a JSON-LD metadata representation of a Book, Painting, Archive or Video that can be imported into various tools and viewers. A Manifest contains a number of Canvases, one for each page. A static Canvas ID will mean items in Wikimedia Commons can be annotated using the W3C Annotation Framework and existing IIIF tools.

    Presentations on IIIF are available on the IIIF youtube channel.

    Images in WikiCommons are difficult to annotate as the image URLs are not static and access is limited to a number of pre-generated sizes. The interface on WikiCommons could be improved for ‘book’ like objects.

  • Who would benefit: All Commons users
  • Proposed solution: Expose WikiCommons data using the IIIF standard and in particular mint static Canvas Ids and Manifest Ids.
  • More comments: This would help with the Book Reader wish as the Internet Archive Book Reader is IIIF compatible.

    Tom Crane has the following proof of concept: https://phabricator.wikimedia.org/D1122 which takes Wikidata items which have the P2677 relation (relative position within image) and turns them into IIIF manifests and annotation lists. These manfiests can be viewed and compared in a IIIF viewer like Mirador. This is an extension to a tool called tool-wd-image-positions written by Lucas Werkmeister and as such will only work with items that have the P2677 property.

    Tom also has the following project: wikipedia-to-iiif which works with any content in Wikimedia Commons and creates a IIIF manifest. This uses the Wikimedia API to create manifests using static images provided by Wikimedia Commons. The benefit of taking this work and embedding it into the core Wikimedia Commons code is that it would give static URIs for manifests and canvas ids allowing annotation. This is further discussed in a gist.

  • Phabricator tickets:
  • Proposer: Glenrobson (talk) 16:06, 9 November 2018 (UTC) with Andy Mabbett, Tom Crane[reply]

Discussion

It's important to note that implementation of the IIIF Presentation API is not dependent on implementation of the IIIF Image API. You can share, annotate and remix images without the tiled deep zoom provided by an image server. Implementing the IIIF Presentation API could also mean that, as well as their current formats, Wikidata query results could be available as IIIF Collections. I could attempt a POC for this too if it would be useful. As an example, this IIIF Collection: Pencil works, Wellcome Library is available for visual re-use, such as here: http://tomcrane.github.io/wellcome-today/today.html?collection=https://wellcomelibrary.org/service/collections/genres/Pencil%20works/

By exposing result sets as IIIF Collections, those result sets can be consumed by IIIF-aware applications, such as a crowdsourcing tool or an image analysis project; IIIF removes the need for a bespoke mapping, and new data generated by those clients addresses the digital object in a standard format, W3C Web Annotations.

Tom Crane

Voting