Page MenuHomePhabricator

machine translation in ContentTranslation adds space before links before which there is an apostrophe
Closed, InvalidPublic

Description

Hi, contenttranslation add spaces before intern links predated by an apostrophe. For example :

à l' [[Université de Reading]]

meet on "Lindy Grant" translate en->fr

(what is the good phabricator tag for contenttranslation issue that need triage ?)

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptAug 25 2016, 6:48 PM

Hi @Framawiki. For future reference, please associate at least one project when creating a task, in order to allow others to see this task when looking at the tasks in corresponding project(s). Thanks! Adding ContentTranslation now :)

(thanks @Aklapper)

It is probably a bug from Yandex, but it can be very useful to correct this.

Amire80 moved this task from Needs Triage to Bugs on the ContentTranslation board.Sep 7 2016, 8:25 AM
Amire80 renamed this task from contenttranslation: it add space before links predated by an apostrophe to machine translation in ContentTranslation adds space before links before which there is an apostrophe.Oct 2 2016, 5:29 AM
Amire80 triaged this task as Medium priority.
Amire80 added a project: WorkType-Maintenance.
Amire80 added subscribers: santhosh, Amire80.

@santhosh, can we fix it in our annotation processing, or is it totally upstream?

Arrbee moved this task from Bugs to Check & Move on the ContentTranslation board.Jul 22 2019, 1:56 PM
Jpita added a subscriber: Jpita.Jul 22 2019, 2:16 PM

QA NOTES:
Try to find specific examples from each MT engine:

  • google
  • yandex
  • apertium
Arrbee assigned this task to Jpita.Jul 22 2019, 2:18 PM
Jpita added a subscriber: Pginer-WMF.

the case specified in the description has the space issue.
@Pginer-WMF where should this go now?

the case specified in the description has the space issue.
@Pginer-WMF where should this go now?

Can you confirm whether this happens for a specific machine translation service or several of them?

Jpita added a comment.Aug 1 2019, 8:27 PM

yandex and google translate, apertium is not available for the en fr pair

Pginer-WMF closed this task as Invalid.Aug 2 2019, 4:08 PM

This seems to be an issue on how Google Translate (example) and Yandex Translate (example) deal with HTML tags for these cases.

Since this is not an issue that is introduced by Content translation and it seems language-specific, we expect the developer teams of these services to improve them. Especially considering that bots are fixing these cases in French Wikipedia currently.

Since this is not an issue that is introduced by Content translation and it seems language-specific, we expect the developer teams of these services to improve them. Especially considering that bots are fixing these cases in French Wikipedia currently.

For the record, all entries entitled espace superflu in that table are concerned.