LinguaLibre talk

Misleading items

How to correct?

I think it would be valuable to point to a page that explain the different steps to follow in order to correct recordings listed here. Myself, I am not completely sure about how to proceed. Pamputt (talk) 16:55, 16 June 2021 (UTC)

504 timeout

Hello, trying to load this page, I got a 504 error several times before suceeding. Do you have any idea how to avoid this to happen again? Should we create several sub-pages? Pamputt (talk) 17:00, 16 June 2021 (UTC)

Check-green.svg Done I have created a sub-page by set of recordings. Pamputt (talk) 06:04, 4 August 2021 (UTC)

P33

I have started to add type of issue (P33) to all items listed on this page. Thus, it will be easy to query these items to then be able to fix them. Pamputt (talk) 06:05, 4 August 2021 (UTC)

Wrong language and duplicate

I started to tag recording item corresponding to recordings that have been made using a wrong language (example: recording of Esparanto words but with French selected as recording language). That's said, I would like to know what to do next. Especially, what to do with Lepticed7's recordings? For example, oceloto (Q53461) is wrongly tagged as a French word so it needs to be fixed but Lepticed recorded also this word in Esperanto (the correct language) : oceloto (Q102186). So in such case, I think we should just delete item that are duplicate because otherwise we will have a conflict when we will ask for renaming the file on Wikimedia Commons.

So, if we agree, when we have duplicate (same speaker, same transcription, and tag as an erroneous language), I propose to delete the item here and ask for deletion on Wikimedia Commons. Pamputt (talk) 06:59, 5 August 2021 (UTC)

I agree that when the word/phrase/sentence has been recorded again by the speaker with the mistake fixed, we should keep the new item and delete the one with the mistake, and ask for the file's deletion on Commons. In other cases, we should ask for the renaming on Commons (by bot?), then change the filename on the element here, change the language (P4) property and remove type of issue (P33) when its over.
All the best — WikiLucas (🖋️) 19:11, 5 August 2021 (UTC)

New on how to proceed

Hi @Pamputt, Poslovitch, & Titodutta and everyone interested in this topic.

Process

I treated some of the mistakes related to wrong language codes. Let's sum up the steps to go through:

  • Change the file category on Commons (most important) by changing the Qid in the description page (the category is not hard-written, but works thanks to the template)
  • Rename the file on Commons with the correct iso code and Qid. Filemover or Admin status needed.
    • Renaming audio files causes a glitch and prevents the audio to be played. One has to reset the transcode in order to be able to play it again. This is very problematic as it takes a lot of time to complete, a phabricator ticket tracks this issue.
  • Change the language code in the label of the element of the recording on Lingua Libre
  • Change the value of language (P4) with the right language
  • Change the value of recording URL pointer (P3) with the the new filename

Facilitating the process

  • Changing the file category can be done using Auto Wiki Browser or Javascript Wiki Browser on Commons, which is then pretty simple.
  • One can manually move each file, but a script exists and works well (although quite slowly), MassRename. Remember you need filemover or admin rights
  • We need a way to reset the files transcode automatically, maybe via bot?
  • The steps on Lingua Libre could also be automated via bot once the files are correct on Commons

Side question: Do you think it would be possible to use Cat-a-lot or similar tools on Lingua Libre? This could be useful to group problematic items.

We're getting to a world with less problematic files from Lingua Libre! 🥳
All the best — WikiLucas (🖋️) 14:30, 26 August 2021 (UTC)

Discussion

I worked for a bot to do this automatically. I have a working prototype that does the modification both on Wikimedia Commons and Lingua Libre. I am currently waiting for getting the bot status on Commons. Then I may ask for the filemover right for PamputtBot to avoid asking for a lot of rename. Concerning the transcode issue, I was not aware about it so I guess we should wait it is fixed before fixing everything automatically and massively by bot. Pamputt (talk) 15:46, 26 August 2021 (UTC)

Hello @Pamputt , do you happen to have some update about this bot project? Some users are still making this kind of errors (but maybe it would be better to fix the cause (the UI) rather than the consequences). All the best! — WikiLucas (🖋️) 20:12, 22 July 2023 (UTC)
Hi WikiLucas, unfortunately it is stalled. I was busy with other business and I did not work on this bot for a while. Maybe I will find some motivation to finish this work before the end of the year ... Pamputt (talk) 20:28, 28 July 2023 (UTC)