Sentence Collector Localization Update

Contributions Agreement template (and Wikipedia link etc) are in English only.

Formally translating the Agreement for 150+ languages will not be possible I think. Also, although Wikipedia link has several translations, it is not in every language.

Thanks, so how about:
0 No sentences
one 1 sentence
other a total of { $totalSentences }

https://commonvoice.mozilla.org/sentence-collector/#/en/stats isn’t showing this on my profile.

The keyboard shortcuts are localized here: https://commonvoice.mozilla.org/cy/speak

Please just hit “Statistics” from the top menu in Sentence Collector. Most of these 3 option translations are from the Statistics page. They all are indicated with context:
CONTEXT sc-lang-info-*

image

Thanks, what I’m seeing is

Statistics

Last Update: ⁨18/11/2021 14:54:27⁩

The Common Voice Sentence Collector has collected ⁨0⁩ sentences in ⁨0⁩ languages!

What you’ve shown looks more useful :slight_smile:

1 Like

Oh, now I see! You should have contributed some :slight_smile:
Currently “zero” cases are controlled in the SW and no output is shown. But we added them in case the SW changes in this respect.

There should be some already - we have 134 hour of recordings!
Or do I need to add some now?

Sentence Collector is not in the same repo, it is maintained separately. So current code does not take localized keyboard shortcuts yet. You can file a feature request for it of course :slight_smile:

Not likely, but @mkohler can answer that one more precisely.

As indicated it is in a different repo, but it has been decided to add the sentences to the main CV Pontoon, thus into CV repo. Although CV has a staging server in Pontoon, as SC does not have a separate project in Pontoon and it is not directly part of CV a staging server is not directly applicable.

It is possible to install SC locally and test it somehow but it will require lots of work. I think we will see it online and correct possible mistakes afterward.

This cannot be true…

Statistics

Last Update: ⁨11/18/2021, 5:30:17 PM⁩

The Common Voice Sentence Collector has collected ⁨4,553,176⁩ sentences in ⁨131⁩ languages!

That should be the same with your output…

And you already should have ~15000 sentences in “cy”, in addition to PR’d ones…

Are you using a separate installation? Don’t you use this one?

https://commonvoice.mozilla.org/sentence-collector/

I tried another computer with Firefox 94.01 on it and Firefox ESR 91.3.0 and it came though the same as yours. :slight_smile:

But Firefox Nightly 96.0a1 (2021-11-18) is still showing:

Statistics

Last Update: ⁨18/11/2021 14:54:27⁩
The Common Voice Sentence Collector has collected ⁨0⁩ sentences in ⁨0⁩ languages!
:man_shrugging:

Mystery solved :slight_smile: Nightly means alpha…

@rprys thanks for reporting all these questions! And thanks to @bozden for already answering so many questions here.

A few notes from my side:

As Bülent has already mentioned this is for now hardcoded and can’t be localized. I filed https://github.com/common-voice/sentence-collector/issues/547 to track this enhancement. Until then, I’ll add a localization note to clarify this.

There is already a staging site: commonvoice.allizom.org/sentence-collector/
However there are no automatic deployments right now, but definitely would be nice to have. @rprys From a localizer’s point of view, how often would you need that website to update with the up to date translations? Is once a day enough for example?

I think from a user experience perspective it would be great if it would convey the message of it being the total amount of sentences for that language in Sentence Collector, no matter if they are reviewed or not. So I’m not sure if removing “total” conveys that.

No no, it’s not that easy. :wink:
@rprys I’d like to figure out what is going wrong in your case. I’m gonna send you a private message here on Discourse.

2 Likes

Yes, thanks @bozden bozden for you contribution :+1:

From my point of view it’s about how the translation looks and reads on the screen. Sometimes it’s good to be able to play around with the content. So 2-3 times a day or at least once a day.

*I couldn’t see the original as the Statistics page did not show the content in Nightly.
And as I said I was confused by the use of total, was it related to the number of sentences or that the sentences were correct ones. Personally, I think not using total is clearer :slightly_smiling_face:

Thanks, I have filed https://github.com/common-voice/sentence-collector/issues/557 for this.

Could the localization update process be started?
I’ve already got some corrections in there… :wink:
Thanks

On the staging website to test it we deploy automatically every 6 hours: https://commonvoice.allizom.org/sentence-collector/#/cy/

Or did you mean the production website? If so, I’m most likely not gonna trigger releases more often than every 2 weeks (same as Common Voice), except if there are fixes or new features developed (there are some features waiting to be released, probably gonna do that next Monday or so).

Thanks, none of the changes I made in Pontoon on 21/11 have appeared yet on https://commonvoice.allizom.org/sentence-collector/#/cy/

Right. I think I found and fixed the bug. Will keep an eye on it! Thanks for reporting this.