Page MenuHomePhabricator

Add limit for maximum audios per page
Open, LowPublicFeature

Description

Within the function which check existing audios of a wikipages, add parameters such as...

  • limit [number]: a limit number for maximum audios per page.
    • default : 10
  • local, [string] : define a language code corresponding to one of the 300 mediawiki-suported languages
    • default: en
  • wikitype, [string] : make it specific to each wikis/wiktionaries.
    • default: wikt

Example :

checkExistingAudio('apple','en','wikt',3)

Event Timeline

Yug triaged this task as Low priority.Feb 19 2021, 8:59 PM
Yug created this task.

Could you describe a little bit morewhy you want to add a limit? This is policy that is decided one a project basis. For example, for now, the French Wiktionary community decided not to limit the number of audio pronunciations. If one day, the community think there are too many pronunciations, it will decide what to do. It is not really relevant to add an function to disable addition of audio on wiki page. For example, if we set a limit to 10 and there are 10 pronunciation from the same region, it would be interesting to keep only 2 audio recordings from this region to allow recording from other regions. To be useful, this job can only be managed by humans, so that the community, Wiktionary project by Wiktionary project (because the rule may be different from one project to another.

Yug updated the task description. (Show Details)

First, the limit parameter should follow local wiktionaries practices. On fr:wikt, it's large, let set it to 10~15 as of now.

Second, as for representativity of various accents, I agree. I think some decision must be made on diversity in order to display a representative set of recordings. Such heuristic should be discussed by the developer in T275245.

The Wiktionaries we currently support (fr and oc) do not enforce a limit on how many recordings are to be added. Most Wiktionaries, however, do. The English or the German Wiktionaries only showcase a single audio (or maybe two). They're handpicked so far - and this obviously raises the question of which recording to choose, but it's out of the scope of this task.
On these Wiktionaries, the bot could be used to populate pages automatically with up to a few recordings while avoiding situations we've seen on the Wiktionnaire with entries having about 20 recordings and counting. This is a tradeoff most Wiktionaries are going to request from the bot, and that's why I support this feature request.

If we want to reach more Wiktionaries, we'll have to abide to more of their rules on audio recordings. And this starts by implementing a Wiktionary-specific per-entry audio recordings limit. So that, to them, it does not appear the bot is going wild.

@Yug There's no need to add limits on Wiktionaries which have not requested them. If at some point the French WIktionary decides it must enforce a limit, at least we'll have the ability to enforce that in the bot's code.

This comment was removed by Yug.
Poslovitch changed the subtype of this task from "Task" to "Feature Request".Feb 22 2021, 7:02 PM