Improve section mapping for Section Translation
Open, MediumPublic
Actions

Assigned To

None

Authored By

	Pginer-WMF
	Mar 2 2021, 11:38 AM

Description

Section translation identifies sections potentially missing across two language versions of an article. This is done to encourage users to translate missing sections. In addition, this information can be considered to filter suggestions.

Currently, the mapping approach is based on previous translations made by Content Translation, which has some limitations. We plan to support a combined approach where multiple strategies are used to identify mapping sections in a reliable and effective way. Tickets exploring different approaches are created as sub-tasks of this epic ticket.

Examples of issues with section mapping are compiled in T283817: Compile examples of section mapping issues to use as a benchmark for future improvements

Related Objects
Search...

Status	Subtype	Assigned	Task
Open		None	T279064 Observations from first research study for Section Translation on Bengali Wikipedia
Open		None	T276212 Improve section mapping for Section Translation
Open		None	T270485 Improve section mapping by integrating the section alignment API
Open		None	T276214 Improve section mapping by integrating MT
Resolved		santhosh	T283815 Pre-fill frequent section titles in the section mapping database
Open		None	T283817 Compile examples of section mapping issues to use as a benchmark for future improvements
Resolved		santhosh	T279045 Update the section suggestion database from latest CX corpus dump
Resolved		diego	T293511 Expand section aligment to more languages, and share dumps
Resolved		santhosh	T306963 Integrate new section mapping database
Resolved		Ladsgroup	T309486 Deploy cxserver db password in private puppet repository
Open	BUG REPORT	None	T310568 Section suggestions: Sub sections are not considered for checking if section exists

Event Timeline

Pginer-WMF created this task.Mar 2 2021, 11:38 AM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptMar 2 2021, 11:38 AM

Pginer-WMF triaged this task as Medium priority.Mar 2 2021, 11:39 AM

Pginer-WMF moved this task from Backlog to General infrastructure on the SectionTranslation board.

Pginer-WMF added a subtask: T270485: Improve section mapping by integrating the section alignment API.

Pginer-WMF mentioned this in T276214: Improve section mapping by integrating MT.Mar 2 2021, 11:47 AM

santhosh added a subtask: T279045: Update the section suggestion database from latest CX corpus dump.Apr 1 2021, 9:57 AM

Pginer-WMF added a parent task: T279064: Observations from first research study for Section Translation on Bengali Wikipedia.Apr 1 2021, 2:52 PM

Pginer-WMF mentioned this in T283817: Compile examples of section mapping issues to use as a benchmark for future improvements.May 27 2021, 1:22 PM

Pginer-WMF closed subtask T279045: Update the section suggestion database from latest CX corpus dump as Resolved.Jul 1 2021, 2:23 PM

Pginer-WMF updated the task description. (Show Details)Oct 28 2021, 12:07 PM

Pginer-WMF added a subtask: T293511: Expand section aligment to more languages, and share dumps.Apr 27 2022, 8:29 AM

diego closed subtask T293511: Expand section aligment to more languages, and share dumps as Resolved.Apr 27 2022, 11:11 AM

Pginer-WMF added a project: Language-Team (Language-2022-April-June).Apr 27 2022, 11:43 AM

Pginer-WMF removed a project: Language-Team (Language-2022-April-June).

santhosh changed the status of subtask T306963: Integrate new section mapping database from Open to In Progress.May 10 2022, 11:31 AM

Pginer-WMF added a subtask: T310568: Section suggestions: Sub sections are not considered for checking if section exists.Jun 14 2022, 2:52 PM

Pginer-WMF closed subtask T306963: Integrate new section mapping database as Resolved.Jun 17 2022, 8:23 AM

I don't know if this should be a subtask, or is part of this, but it would be great to have the hability to translate subsections. Currently (it seems) that only h2 sections can be translated, but in many articles this is extremely large. Translating also h3 or h4 would make it easier to present it as small tasks, feasible in a couple of minutes.

In T276212#8014192, @Theklan wrote:

I don't know if this should be a subtask, or is part of this, but it would be great to have the hability to translate subsections. Currently (it seems) that only h2 sections can be translated, but in many articles this is extremely large. Translating also h3 or h4 would make it easier to present it as small tasks, feasible in a couple of minutes.

This is a good suggestions- worth exploring.

Improve section mapping for Section TranslationOpen, MediumPublicActions

Description

Related ObjectsSearch...

Event Timeline

Improve section mapping for Section Translation
Open, MediumPublic
Actions

Related Objects
Search...