Can not localize Collabora Code

Hi all.

I am using collabora/code: in a container environment.

Two questions here:

1- How to add a UI language and make it default? I can’t find it in the docs.

2- My instance has the necessary dictionary files out of the box for my desired language, but spellchecking doesn’t list that language on the UI. How to make it available an already installed language? And how to update those language files later? Is only downloading newer .aff and .dic files and overwriting enough?

Hey @TurabG, thanks for the questions + welcome to the forum. :slight_smile:

Which language are you using?

I explained how to change the UI to “Portuguese (Brazilian)” in this post:

At the end of your URL, you have to stick:

  • &lang=pt-BR

where pt-BR is whatever UI language you are trying to use.

On Default UI… see answer earlier in that same thread too:

Did you see the Collabora Online SDK:

Method #1: Docker Environment Variable

At startup, you can pass your docker an environment variable:

  • dictionaries

Method #2: coolwsd.xml File

The list of languages + how to expand them is also described in the Collabora Online SDK:

If you look inside the file:

  • coolwsd.xml

you should see a big list of:

<allowed_languages desc="List of supported languages of Writing Aids (spell checker, grammar checker, thesaurus, hyphenation) on this instance. Allowing too many has negative effect on startup performance." default="de_DE en_GB en_US es_ES fr_FR it nl pt_BR pt_PT ru">de_DE en_GB en_US es_ES fr_FR it nl pt_BR pt_PT ru</allowed_languages>

You can adjust those languages as needed.

Thank you @Tex, I enabled the desired language for spellchecking with the environment variable as you pointed out.

But I don’t understand how changing URL parameters will work through Collabora Online docker image? I am using Nextcloud for cloud storage and set it to use CODE server to edit files online which doesn’t involve direct URLs of Collabora. How to apply it?

Secondly, would overriding dictionary files with new versions be enough for updating them?

1 Like

Yes, I would suspect so.

What were the dictionary files you were hoping to expand? Do you have your own custom wordlist or something?

Hmmm, I’m unsure.

If you’re using Nextcloud, you might want to look into “NextCloud Office” instead.

It’s a custom fork that’s specifically tuned for Nextcloud:

I know that they automatically handle the UI language based on user. So:

  • A German user would automatically get the German UI.
  • An English user would automatically get the English UI.

I did a search through their code for the word “language”, and I think this might be the files (url.js + index.js) where they work their URL-substitution magic:

Thank you for you time. About the UI language, I think I will need to ask it in Nextcloud forums.

About spell-checking: how to set a default language system wide? Everytime I open a document, English (USA) is the default one. Because my Nextcloud UI is already a different language but Collabora is still displayed in English.

No, only for updating I asked. Because the dictionary versions shipped with Collabora is a bit old.

Oh? Which language?

And do you have a URL to an updated dictionary?

We could probably get that change submitted over to Github and merged for the upcoming versions. :slight_smile:

Side Note: On another forum + another open-source software I use, a Spanish user just appeared and informed them the Spanish dictionary was way out of date. Turns out, the same maintainer split off and created:

which was up-to-date Spanish dictionaries.

I’m unsure.

I’m suspecting once you install Nextcloud Office, and the Nextcloud UI flips, your new documents will start using that as the document language too.

For example, see this Nextcloud Office fix earlier this year that adjusted en-AU (Australian) defaults:

It is Turkish. I think Collabora includes a 2014 or 2015 version of this dictionary:

It has a much newer release in 2022.

Yes, it looks like the LibreOffice version is from here:

Hunspell dictionary for Turkish
Version 1.2.0 (Apr 4, 2015)

This dictionary is released under Mozilla Public License Version 2.0.

And the one you linked to is:

TDD hunspell-tr
Version 1.1.0 (Jan 7, 2022)

MIT License
Copyright (c) 2021 Turkish Data Depository

QUESTION: Do you have a LibreOffice Bugzilla account?

  • If yes, could you submit an enhancement request on their Bugzilla?
  • If no, that’s okay. (I could submit one of your behalf too.)

Side Note: Some more dictionary info can also be found here:

and the exact files can be seen in:

Once we get that updated Turkish dictionary merged into LibreOffice, everyone can benefit from it in future versions of LibreOffice/Collabora. :slight_smile:

Well, newer/bigger doesn’t always mean better. :stuck_out_tongue:

But yes, in this case, looking at their comparison stats, it looks like it might be more accurate + give more accurate Right-Click suggestions.

What are some of the example words you found missing?

Yes quite right. I suggested this because as you noticed, comparison stats are better and it is already a fork and updated version.

To be honest, I haven’t tested thoroughly, I am just keen on updated versions :slight_smile: Because you know, languages are alive and continue evolving. And the one which LibreOffice uses right now actually seems abandoned.

I did, thank you for the heads up. 158573 – Turkish spell check dictionary update

Fantastic. I’ll follow it now + responded there. :slight_smile:

Heh. A few months ago, I actually gave a talk all about spellchecking:

And that was based on these in-depth posts I wrote:

There’s actually a balancing act:

You can include every single “valid word” under the sun = less red squigglies… But then you’d be missing more ACTUAL typos.

(So you’d need the upgraded grammar level on top of that too.) :stuck_out_tongue:

(See the above talk/posts for much more details.)

Side Note: Personally, I prefer SCOWL’s methods—splitting dictionaries based on actual usage/popularity/frequency:

  • Words that exist in every single dictionary?
    • Yes, those can go in the default spellcheck lists.
  • Obscure medical/scientific/biology terms that no mere human is going to be using in their everyday lives?
    • Those can go into an expanded dictionary that you can optionally download instead. :slight_smile:

In my posts/talk, I actually mention “clothes” vs. “clotes”.

99.999% of the time, people will be speaking about the shirt on their backs! The other EXTREMELY RARE word would’ve just been a typo.

Thank you for taking time to share so much on this.

I totally agree.

1 Like

Hey @TurabG ,

The latest Turkish dictionary you recommended just got merged a few days ago:

It should make its way into:

  • LibreOffice 24.8 (coming out later this year)

and Collabora Online soon. :slight_smile:

Thanks again for letting us know about the spellchecking issue. :slight_smile:

Yeah, I saw it from a notification e-mail from the bug tracker. Thank you for escalating it there so it could help more.

1 Like