Page MenuHomePhabricator

Media Settings "Uploaded by..." not consistently accurate
Closed, ResolvedPublic1 Estimated Story Points

Description

checked in test2

  1. Open VE -> Insert->Media
  2. type in search criteria: Leonardo da Vinci and select Mona Lisa image. Media Settings box will display "Uploaded by Leonardo da Vinci' .
  3. Go to 'More information' - there is a user name in File history and it's not Leonardo da Vinci.

Should "Upload by..." be read from File history->User and not from Artist, Author and other easily changed/edited categories?
'Uploaded by...' often gives incorrect information - try Rembrandt, for example

Screen_Shot_2015-02-06_at_12.17.26_PM.png (619×1 px, 300 KB)

Event Timeline

Etonkovidova raised the priority of this task from to Medium.
Etonkovidova updated the task description. (Show Details)
Etonkovidova added a project: VisualEditor.
Etonkovidova added a subscriber: Etonkovidova.

Okay, we need to figure out what to do with the labling of the information we're receiving from the API.

From the info we're getting from the api, here's what's relevant to user --

  1. Artist (extmetadata)
  2. Credit (extmetadata)
  3. Uploader (imageinfo iiprop user)

#1 and #2 are user-generated from templates, and are somewhat inconsistent. The Mona Lisa image above is marked to be "Artist" da Vinci (which makes sense) but other pictures usually have the picture-taker as Artist (like this one of a building in fifth avenue)

#3 is the name of the user who uploaded th image, but it behaves a bit awkwardly with images that have revisions. It will only show the latest uploader, and not the first uploader. (There's a bug/feature request about that in phabricator.)

We need to figure out which field we label as what. As a first step I recommend we change "Uploaded by" to something more generic; The media viewer just has the name of the Artist without a label. We could also add "by " before, or something similar.

Then we can decide which of the information we want to present at all.

Maybe add "Uploader" from imageinfo iiprop user with the current label, and add "Artist" as the label for the existing field?

gerritbot added a subscriber: gerritbot.

Change 190031 had a related patch set uploaded (by Mooeypoo):
Display image uploader and artist correctly in the media dialog

https://gerrit.wikimedia.org/r/190031

Patch-For-Review

Artist is the author, Credit is the source. Changing them to reasonable names is T71914.

Both are parsed fromm templates (almost all extmetadata fields are) and as a result can contain all kinds of crap, like a bullet list of people when there are multiple authors, or a table of various information about the author (like with the {{creator}} template - CommonsMetadata can extract information from that one, but there might be others we are not aware of). Media Viewer removes all tags except <a> while keeping their content, and that gives a reasonable result in most of the cases.

When there are multiple author-ish persons (e.g. a painter and a photographer) and they are given in separate templates so CommonsMetadata can recognize them as multiple authors, it will only return the photographer. extmetadata has a flat API format so there is no way to return multiple values (T59259).

Uploader is less problematic, but if you want the first uploader, not last, you need an extra imageinfo request with iiprop=user and iilimit=lots. T65097 is about providing a workaround in CommonsMetadata (extmetadata).

Change 190031 merged by jenkins-bot:
Display image uploader and artist correctly in the media dialog

https://gerrit.wikimedia.org/r/190031