feat: add romansh (DEV-867) #193

jnussbaum · 2022-05-24T08:28:08Z

jnussbaum · 2022-05-24T10:01:06Z

@irinaschubert @BalduinLandolt With this PR, I resolved two issues. But I'm not sure if my solution is too superficial, because I only prevented 3-letter-language-codes via the JSON Schema. I don't know if it could/should be prevented on a deeper level.

irinaschubert

Regarding your question, there is a file called language-codes-3b2_csv.csv. I think in there the 3 letter codes should be removed as well. I am not sure if in there only en, de, fr, it and rm should be listed or if we need all of the 2 letter codes. This file is used for the validation of language codes that are used in lists which are created from Excel files.
It was me who added this file initially in #75. But I don't remember if it was us who just added it (I remember that I discussed it with Ivan) or if it was a real user requirement to support more languages in lists.
It would therefore be good to discuss it with PM or Rita if we need more than the 5 languages for list labels.
We should also check in the triplestore if we actually have projects that use other than the 5 languages (at least I hope so). If so, could you please add a sub-task and assign me?

docs/dsp-tools-create.md

Co-authored-by: irinaschubert <irina.schubert@dasch.swiss>

BalduinLandolt

So far so good.

I agree that the 3 letter language codes should be removed from the language-codes-3b2_csv.csv file; and this is only used in knora.dsplib.utild.excel_to_json_lists.py in check_language_code() which then should also be adjusted. (The current implementation checks every field in the CSV, which so far has been 3-letter-code, 2-letter-code and country name; after removing the 3-letter-codes, it would check the 2-letter coded and the country names. But in my opinion it should not check the country names at all.).
I'm just unsure if that's in the scope of this PR.

jnussbaum · 2022-05-30T09:50:52Z

@irinaschubert @BalduinLandolt Following your comments, I suggest removing language-codes-3b2_csv.csv altogether. It doesn't make sense to have tons of language codes that are never used, because everywhere else in the DSP software stack, there are only the hardcoded en, fr, de, it, rm, and nothing else. So I suggest hardcoding it also here. Is that okay for you?

irinaschubert · 2022-06-01T13:03:44Z

@irinaschubert @BalduinLandolt Following your comments, I suggest removing language-codes-3b2_csv.csv altogether. It doesn't make sense to have tons of language codes that are never used, because everywhere else in the DSP software stack, there are only the hardcoded en, fr, de, it, rm, and nothing else. So I suggest hardcoding it also here. Is that okay for you?

Yes, that makes sense to me!