Page MenuHomePhabricator

Proof-of-concept for showing a machine translated sections of Wikipedia articles
Open, In Progress, HighPublic

Description

As part of the process to scale the uses of MinT, we want to explore the use of machine translation to provide access to sections missing in other languages. The specific ways in which this will be exposed to users, will depend on the research and design explorations (T340997). However, there are technical services that are expected to be useful regardless of the specific UI used.

This ticket is focused on the technical explorations needed to support the translation of Wikipedia sections (one or several, including the lead section). These may include improvements in the machine translation API, approaches for dealing with content (links, references across sections, templates etc.) and similar technical aspects.


As part of the work in Content Translation, some of the above aspects have been resolved for the editing perspective, and (parts of) these approaches may or may not apply in this context of readers on a Wikipedia article.

Result

By adding the following line in the common.js in wikipedia, you can see the proof of concept

importScript( 'User:Santhosh.thottingal/mint-section-translation.js' );

Example: https://en.wikipedia.org/wiki/User:Santhosh.thottingal/common.js

Screenshot:

image.png (480×451 px, 198 KB)

Event Timeline

santhosh changed the task status from Open to In Progress.Oct 11 2023, 4:54 AM
santhosh claimed this task.

By adding the following line in the common.js in wikipedia, you can see the proof of concept

importScript( 'User:Santhosh.thottingal/mint-section-translation.js' );

Example: https://en.wikipedia.org/wiki/User:Santhosh.thottingal/common.js

Screenshot:

image.png (480×451 px, 198 KB)

A screenshot illustrating reference misplacement with current prototype: From https://en.wikipedia.org/wiki/Polar_bear

image.png (675×864 px, 360 KB)

Another example:

image.png (676×828 px, 277 KB)

Change 1007307 had a related patch set uploaded (by Santhosh; author: Santhosh):

[mediawiki/services/machinetranslation@master] Improvements for rich text adaptation - references

https://gerrit.wikimedia.org/r/1007307

Change 1007568 had a related patch set uploaded (by Santhosh; author: Santhosh):

[mediawiki/services/machinetranslation@master] Improvements for rich text adaptation - repeated annotations

https://gerrit.wikimedia.org/r/1007568

Change 1007307 merged by jenkins-bot:

[mediawiki/services/machinetranslation@master] Improvements for rich text adaptation - references

https://gerrit.wikimedia.org/r/1007307

Change 1007568 merged by jenkins-bot:

[mediawiki/services/machinetranslation@master] Improvements for rich text adaptation - repeated annotations

https://gerrit.wikimedia.org/r/1007568

Change 1012988 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[operations/deployment-charts@master] Update MinT to 2024-03-20-072303-production

https://gerrit.wikimedia.org/r/1012988

Change 1012988 merged by jenkins-bot:

[operations/deployment-charts@master] Update MinT to 2024-03-20-072303-production

https://gerrit.wikimedia.org/r/1012988