Page MenuHomePhabricator

ContentTranslation drops all references in first two paragraphs; one page link ends up incorrectly
Open, MediumPublic

Description

I loaded Atazanavir on Spanish Wikipedia into CTX and added the first three paragraphs. You will notice all the references are in place.

I made no other changes as I was testing this tool as some of my volunteers were having issues. Than when I hit publish I get the following with the references missing from the first two paragraphs. Also the Wikipedia page " HIV/AIDS" ends up as " #sida/de VIH "

Details to reproduce the issue

A sample of the problematic Wikitext is captured in this page (quick link to translate)

The example is a reference that is used in multiple parts of a paragraph. All those instances appear in Content Translation when the paragraphs are added to the translation. But some of them are lost after publishing.

Original content:

In Content Translation:

After publishing:

The red circles show the places where the reference "[1]" was expected to appear but failed to do so.

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptMay 26 2018, 3:28 PM
Aklapper renamed this task from Content Translation not handling references appropriately to ContentTranslation drops all references in first two paragraphs; one page link ends up incorrectly.May 28 2018, 6:31 AM

Content translation editing surface has been fragile when dealing with some complex elements such as references at times. We are currently working on version 2 which uses a more solid editing surface provided by VisualEditor. The new version will make dealing with references as reliable as Visual Editor.

Also the Wikipedia page " HIV/AIDS" ends up as " #sida/de VIH "

This other issue may be caused by the automatic translation service. In this case, Apertium may be providing a sub-optimal translation. In addition, since Apertium only works with plain text, Content Translaiton has to recreate the links and references in the right place, resulting on the link being applied on the wrong place in some cases. This is something we cannot do much about since it depends on external translation services.

Pginer-WMF triaged this task as Medium priority.Jun 1 2018, 10:05 AM
Pginer-WMF moved this task from Needs Triage to CX2 on the ContentTranslation board.
Vvjjkkii renamed this task from ContentTranslation drops all references in first two paragraphs; one page link ends up incorrectly to 67baaaaaaa.Jul 1 2018, 1:07 AM
Vvjjkkii raised the priority of this task from Medium to High.
Vvjjkkii updated the task description. (Show Details)
Vvjjkkii removed a subscriber: Aklapper.
CommunityTechBot renamed this task from 67baaaaaaa to ContentTranslation drops all references in first two paragraphs; one page link ends up incorrectly.Jul 2 2018, 1:54 PM
CommunityTechBot lowered the priority of this task from High to Medium.
CommunityTechBot updated the task description. (Show Details)
CommunityTechBot added a subscriber: Aklapper.
Arrbee moved this task from CX2 to Check & Move on the ContentTranslation board.Feb 17 2020, 8:16 AM
Arrbee moved this task from Check & Move to CX2 on the ContentTranslation board.Feb 17 2020, 9:30 AM
Jpita added a comment.Feb 18 2020, 8:59 AM

@Pginer-WMF I translated the Atazanavir article into a draft and everything seems ok, can you double check please?

@Jpita looking at your draft and I see that a bunch of the references have gone missing.

Here every sentence has a reference https://en.wikipedia.org/wiki/Atazanavir

Here I am see at least 5 are missing https://es.wikipedia.org/wiki/Usuario:Jpita23/Atazanavir

Jpita added a comment.Feb 19 2020, 1:59 AM

Thanks @Doc_James , we'll put this in our todo list.

I made a sample of the problematic Wikitext in this page (quick link to translate)

The example is a reference that is used in multiple parts of a paragraph. All those instances appear in Content Translation when the paragraphs are added to the translation. But some of them are lost after publishing.

Original content:

In Content Translation:

After publishing:

The red circles show the places where the reference "[1]" was expected to appear but failed to do so.

Pginer-WMF updated the task description. (Show Details)Feb 19 2020, 9:37 AM
Pginer-WMF removed Jpita as the assignee of this task.Mar 19 2020, 1:42 PM
Pginer-WMF added a subscriber: Jpita.