Page MenuHomePhabricator

LinguaImporter gadget adds MediaType=audio
Open, Needs TriagePublicBUG REPORT

Description

Hello, I imported the Laari language on Lingua Libre. In the new item that has been created, the property p_24 (MediaType) is present with the value audio. I think this property should not be here.

Related: T233917 , T215056 , T232299

Event Timeline

Pamputt renamed this task from LinguaImporter gadget add MediaType=audio to LinguaImporter gadget adds MediaType=audio.Oct 6 2020, 8:03 PM
Pamputt changed the subtype of this task from "Task" to "Bug Report".

Inspected the code. This is indeed the intended behavior: languages which produce audios files get this tag and value. Sign languages get another value.

Older languages may miss this data tho. 1) Need to check. 2) Create a task to update old Qid language items where Commons category and p_24 (MediaType) are not defined.

Yug updated the task description. (Show Details)
Seb35 added a subscriber: Seb35.

I will solve this task (WMFR contract).

If I understand correctly, it should no more be removed the P⁠24 := audio, but it should be added on every language item (except sign languages where it should be P⁠24 := video).

According to the following request to be executed here, there are currently 561 languages, whose 53 have P⁠24 := audio and 1 have P⁠24 := video.

SELECT ?language ?languageLabel ?code ?type
WHERE {
  ?language prop:P2 entity:Q4 .
  OPTIONAL { ?language prop:P13 ?code } .
  OPTIONAL { ?language rdfs:label ?languageLabel . FILTER( lang(?languageLabel) = "en" ) } .
  OPTIONAL { ?language prop:P24 ?type } .
  #?language prop:P24 ?type .
}

I searched "sign" and there is only Q99628, so I guess all others are audio languages.

Also, I see some specific "languages": Q367 "Judeo-Arabic languages" (macrolanguage), Q364 "Middle English" (historic), Q368 "Egyptian" (extinct), they can be tagged P⁠24 := audio, but it can be debated (collective language or reconstructed pronunciation, it could be used subclasses of Q4 or qualifiers).

If the purpose is to indicate whether selecting this language triggers audio recording or video recording, I think that the P24 property should indeed be added on every language item.
But a more important question is about how we manage languages importation from wikidata, and why we import them individually, please see T284645

Seb35 removed Seb35 as the assignee of this task.Fri, Jul 16, 2:05 PM

I remove myself as assignee, since it is not a technical issue but an editorial one to be decided within the community.