Prevent loading data from Wikipedia pages for specific wikidata pages #29

New Issue

2025-02-13T22:40:41+01:00

jrenslin commented

2025-02-13 22:40:41 +01:00

Some times, a Wikipedia page is trying to be more specific than wikidata. Wikidata links to the target entity in almost all languages, but to a disambiguation page in others. In such cases, the data from Wikidata is altogether better than whatever Wikipedia offers.

jrenslin referenced this issue

2025-02-13 22:42:38 +01:00

Disallow loading disambiguation pages via wikidata fetcher #23

awinkler commented

2025-02-14 07:56:15 +01:00

From a wikidata point of view, the only viable solution seems to be to ask for wikidata uris / QIDs instead of wikipedia urls. Every wikipedia page has a wikidata item attached, but that's not the case the other way round. As to the label (cf #28 ), you can always use the wikipedia labels if need be.

jrenslin commented

2025-02-14 11:01:50 +01:00

Well, point is that we can handle both (and need to be able to). But that's essentially a different issue. Here I'm concerned with cases where one Wikidata page essentially has two Wikipedia pages in one language. It's certainly an edge case, but essentially what I described above: Coming from the Wikidata item, all languages except for one (usually the German Wikipedia) reference a precise match in Wikidata. That single outlier references a disambiguation page.
@stefan

Well, point is that we can handle both (and need to be able to). But that's essentially a different issue. Here I'm concerned with cases where one Wikidata page essentially has two Wikipedia pages in one language. It's certainly an edge case, but essentially what I described above: Coming from the Wikidata item, all languages except for one (usually the German Wikipedia) reference a precise match in Wikidata. That single outlier references a disambiguation page. @stefan

awinkler commented

2025-02-14 12:09:42 +01:00

At least in Wikidata I can't seem to find an item with two wikipedia pages in the same language: https://qlever.cs.uni-freiburg.de/wikidata/Zkub6F

There are, however, 180 cases where more than one Wikipedia pages refer to the same wikidata item (https://qlever.cs.uni-freiburg.de/wikidata/FbVWQW). Is thas what you mean? But maybe @stefan has a concrete example?

At least in Wikidata I can't seem to find an item with two wikipedia pages in the same language: https://qlever.cs.uni-freiburg.de/wikidata/Zkub6F There are, however, 180 cases where more than one Wikipedia pages refer to the same wikidata item (https://qlever.cs.uni-freiburg.de/wikidata/FbVWQW). Is thas what you mean? But maybe @stefan has a concrete example?

stefan commented

2025-02-14 12:19:24 +01:00

Don't know if it is a different problem ...
Wikidata knows that the Wikipedia-Link is one to a disambiguation page in this example: https://www.wikidata.org/wiki/Q487772. It is given in the main box for german, english, french and it is given in the wikipedia-link-box (on the right) NOT for german, but for english and other languages

Don't know if it is a different problem ... Wikidata knows that the Wikipedia-Link is one to a disambiguation page in this example: https://www.wikidata.org/wiki/Q487772. It is given in the main box for german, english, french and it is given in the wikipedia-link-box (on the right) NOT for german, but for english and other languages

jrenslin commented

2025-02-14 12:49:51 +01:00

Don't know if it is a different problem ...
Wikidata knows that the Wikipedia-Link is one to a disambiguation page in this example: https://www.wikidata.org/wiki/Q487772. It is given in the main box for german, english, french and it is given in the wikipedia-link-box (on the right) NOT for german, but for english and other languages

Yeah, that one's the case already solved in #23 , since the whole Wikidata item is marked as a Wikimedia disambiguation entity. Problem is only if there is a Wikipedia disambiguation page linked to a valid Wikidata entity (I chanced upon that only once, so it's really an extreme edge case, but we should keep it in mind).

> Don't know if it is a different problem ... > Wikidata knows that the Wikipedia-Link is one to a disambiguation page in this example: https://www.wikidata.org/wiki/Q487772. It is given in the main box for german, english, french and it is given in the wikipedia-link-box (on the right) NOT for german, but for english and other languages Yeah, that one's the case already solved in #23 , since the whole Wikidata item is marked as a Wikimedia disambiguation entity. Problem is only if there is a Wikipedia disambiguation page linked to a valid Wikidata entity (I chanced upon that only once, so it's really an extreme edge case, but we should keep it in mind).

Sign in to join this conversation.

3 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: museum-digital/MDNodaHelpers#29