Page MenuHomePhabricator

Allow more filter options in GapFinder
Open, HighPublic

Description

Allow filtering to help translators identify articles that are already in a better shape in the source language and ready to be translated.

Filter options:

The current production instance is somewhat limited in what other information can be gathered without significant performance implications, but moving to pre-computing the predictions could also add the extra feature gathering for more performant filtering.

Details

Related Gerrit Patches:
mediawiki/services/recommendation-api : masterAdd byte size filter

Event Timeline

leila created this task.Jun 28 2017, 3:11 PM

The Translation Studies MSc course at the University of Edinburgh would like to continue to offering the 4000 word Wikipedia translation assignment in future years but one particular area of concern has been how hard it has been for the students to select a suitable article to translate.
The chosen article must:

  1. Be 4000 words long approx.
  2. Be Well-referenced / sufficient quality or importance.
  3. Not exist in the target Wikipedia.
  4. Be of sufficient language challenge (evaluated by the tutor)
  5. Be on a suitable subject matter / be of interest to them.

Meeting all 5 criteria is difficult and students/tutors have raised concerns. Therefore if we can remove hurdles and streamline the process by adding a word count filter and a filter for article quality / its referencing then I think they will continue with the project as a longterm proposition as it is win-win for both sides. Aberystwyth University and the Open University are looking to embark on similar translation projects now too so it would be useful to help rollout translation assignments elsewhere too.

Echoing what Stinglehammer said, this would be a big improvement to the user experience, and it would also be used at other institutions - Aberystwythand the Open University are two likely partners, and anyone who wants to run translation projects would really benefit.

Richard_Nevell_WMUK rescinded a token.
Richard_Nevell_WMUK awarded a token.
leila assigned this task to schana.Jul 7 2017, 6:18 PM
leila raised the priority of this task from Normal to High.

Thanks @Richard_Nevell_WMUK and @Stinglehammer for your comments. They are very helpful for us understanding how we can prioritize tasks on our end. I've changed the priority of this task from Normal to High which means that in the absence of major bugs or tasks in other boards, we will prioritize this feature(s) over others. Please stay tuned.

leila moved this task from Backlog to Next Up on the GapFinder board.Jul 7 2017, 6:20 PM

Filtering in GapFinder will need to be supported through the backend, otherwise there likely won't be enough results for GapFinder alone to perform the filtering. I've started working on what a backend query may look like in WDQS to include the relevant information for future filtering, but this will likely need amended since it doesn't include some of the information.

wdqs query link

I would like to add my support to these changes as i am also working with Translation students and this would be very useful.

Change 373338 had a related patch set uploaded (by Nschaaf; owner: Nschaaf):
[mediawiki/services/recommendation-api@master] Add byte size filter

https://gerrit.wikimedia.org/r/373338

schana updated the task description. (Show Details)Aug 24 2017, 1:51 PM

@bmansurov If there are more people interested as part of GSoC, you can consider assigning this task. It would be great to have filters on top of recommend.wmflabs.org in the interface. Pau has done some mock-ups for the front-end development it as well, so if you get both backend and frontend expertise, they can pick this up.

@leila I'm afraid I cannot manage more than two students. According to The DOs and DON’Ts of Google Summer of Code: Mentor Edition, even in my second year of mentoring, it's not advised to mentor more than one student.

leila added a comment.Mar 13 2019, 6:04 PM

@bmansurov understood. Just have it on your radar for the next time an opportunity arises.