Page MenuHomePhabricator

Newly supported languages in Google Translate
Open, HighPublic

Description

Recently Google Translate added machine translate supports in 24 new languages, 21 of which have a Wikipedia project:

  1. Assamese (as)
  2. Aymara (ay)
  3. Bambara (bm)
  4. Bhojpuri (bh)
  5. Ewe (ee)
  6. Guarani (gn)
  7. Iloko (ilo)
  8. Konkani (kon)
  9. Sorani (ckb)
  10. Lingala (ln)
  11. Luganda (lg)
  12. Maithili (mai)
  13. Maldivian (dv)
  14. Meitei (mni)
  15. Northern Sotho (nso)
  16. Oromo (om)
  17. Quechua (qu)
  18. Sanskrit (sa)
  19. Tigrinya (ti)
  20. Tsonga (ts)
  21. Twi (tw) (Google API seems to be using "ak" code)

This should be enabled in ContentTranslation.

Event Timeline

Pginer-WMF triaged this task as Medium priority.May 27 2022, 8:32 AM
Pginer-WMF moved this task from Needs Triage to MT on the ContentTranslation board.

Checking the Google API it seems the newly listed languages are already available. I also noticed that Akan was also supported but was not enabled yet in Content Translation and it may be the code used for Twi. As well as others for which there is no Wikipedia yet (Hmong, and Krio/Sierra Leonean Creole)

Pginer-WMF updated the task description. (Show Details)
Pginer-WMF updated the task description. (Show Details)
Pginer-WMF raised the priority of this task from Medium to High.Wed, Jul 27, 8:22 AM

Change 817734 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[mediawiki/services/cxserver@master] Add newly supported languages in Google Translate

https://gerrit.wikimedia.org/r/817734

Change 817734 merged by jenkins-bot:

[mediawiki/services/cxserver@master] Add newly supported languages in Google Translate

https://gerrit.wikimedia.org/r/817734

Change 817931 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[operations/deployment-charts@master] Update cxserver to 2022-07-27-220330-production

https://gerrit.wikimedia.org/r/817931

Change 817931 merged by jenkins-bot:

[operations/deployment-charts@master] Update cxserver to 2022-07-27-220330-production

https://gerrit.wikimedia.org/r/817931

Mentioned in SAL (#wikimedia-operations) [2022-07-28T08:57:31Z] <kart_> Updated cxserver to 2022-07-27-220330-production (T308248)

Change 818087 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[mediawiki/services/cxserver@master] MT: Map Google MT language code with Wikipedia

https://gerrit.wikimedia.org/r/818087

Change 818087 merged by jenkins-bot:

[mediawiki/services/cxserver@master] MT: Map Google MT language code with Wikipedia

https://gerrit.wikimedia.org/r/818087

Change 819045 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[mediawiki/services/cxserver@master] MT: Add mapping for mni for Google

https://gerrit.wikimedia.org/r/819045

Change 819045 merged by jenkins-bot:

[mediawiki/services/cxserver@master] MT: Add mapping for mni for Google

https://gerrit.wikimedia.org/r/819045

Change 820075 had a related patch set uploaded (by KartikMistry; author: KartikMistry):

[operations/deployment-charts@master] Update cxserver to 2022-08-03-082610-production

https://gerrit.wikimedia.org/r/820075

Change 820075 merged by jenkins-bot:

[operations/deployment-charts@master] Update cxserver to 2022-08-04-022612-production

https://gerrit.wikimedia.org/r/820075

Mentioned in SAL (#wikimedia-operations) [2022-08-04T05:49:40Z] <kart_> Updated cxserver to 2022-08-04-022612-production (T313296, T308248)

@Pginer-WMF The following languages seem not working with Google MT.

  1. Bambara (bm)

Error:

"Translation with Google failed for en > bm. Error: 503 : The service is unavailable at this time."
  1. Manipuri Meitei (mni) (Google MT code: mni-Mtei)

Error:

"Translation with Google failed for en > mni-Mtei. Error: 503 : The service is unavailable at this time."
  1. Goan Konkani (gom)

Error:

"Translation with Google failed for en > kon. Error: 400 : Invalid Value"
  1. Goan Konkani (gom)

Error:

"Translation with Google failed for en > kon. Error: 400 : Invalid Value"

It seems Google is using 'gom' code only, but with that too,

"Translation with Google failed for en > gom. Error: 503 : The service is unavailable at this time."

I tested Iloko (ilo) and it seems not to be working. Note that testing Iloko on the tool may require to switch the MT service to Google in the drop-down since it is not the current default.

  1. Goan Konkani (gom)

Error:

"Translation with Google failed for en > kon. Error: 400 : Invalid Value"

It seems Google is using 'gom' code only, but with that too,

"Translation with Google failed for en > gom. Error: 503 : The service is unavailable at this time."

Should note here that the ISO 639-3 code for Konkani is kok, not kon (kon is the 3-letter code for Kongo, which has two-letter code kg). In any case, gom is the code that Konkani Wikipedia uses, so Google Translate only using gom shouldn't be an issue.