Page MenuHomePhabricator

Fix Machine Translation Service code names to be consistent
Closed, ResolvedPublic

Description

When reviewing recent translation data for the analysis in T310773, I noticed that there are a couple of machine translation services that have inconsistent internal code names in the data (source: cx_corpora cxc_origin field.). See details below:

Google and Google Translate
154 total translations tagged with Google Translate instead of Google. Details below:

  • First Google Translate instance occurred on 13 July 2022
  • Last Google Translate instance occurs on 04 August 2022
  • Since 04 August 2022, all Google translations have been correctly tagged as Google so far.

Yandex and Yandex.Translate
Only 1 translation was tagged with Yandex.Translate. Details of the translation below:

  • translation start time: 28 July 2022.
  • Source Language: en
  • Target Language: mr

Can you please review to confirm if the code names for each Machine Translation service are consistent?

Event Timeline

Pginer-WMF triaged this task as Medium priority.Sep 28 2022, 1:50 PM

Following our latest work for the user-facing names of the MT services (T317293), I think that this task is no longer relevant, as we are only using user-facing names like "Google Translate" for the UI and the proper MT names (e.g. "Google") for all the underlying functionality, including cx corpora translation units. However, this task cannot easily be verified by QA.

Pginer-WMF claimed this task.
Pginer-WMF subscribed.

Thanks Nik.
@MNeisler, this is expected to be resolved. Let us know if you find similar inconsistencies in the future.