Closed parker57 closed 6 years ago
We should probably strip values we're not using from the language json file and just keep keys called name
and iso
or something, which makes scope
value irrelevant
I think it might be useful to keep the extra information and use it to extend the Language
class. E.g. Language("eng").type
I don't think there is much point keeping the other iso values but I am a bit upset don't have language family, that would be neat for analysis and even presentation.
This guy might have better JSON but tragically seems to have stopped at 639-3
"bo": {
"639-1": "bo",
"639-2": "bod",
"639-2/B": "tib",
"family": "Sino-Tibetan",
"name": "Tibetan Standard, Tibetan, Central",
"nativeName": "བོད་ཡིག",
"wikiUrl": "https://en.wikipedia.org/wiki/Standard_Tibetan"
},
...
"ru": {
"639-1": "ru",
"639-2": "rus",
"family": "Indo-European",
"name": "Russian",
"nativeName": "Русский",
"wikiUrl": "https://en.wikipedia.org/wiki/Russian_language"
},
I'm going to strip the unused keys for now, then we can add back other keys (or switch to a different dataset) when we want to expand the language class
noumar/iso639 looks like a good replacement to keeping the data in this project.
Going to close this because the issue is solved - feel free to open a new issue for changing where we source iso639 codes if that's a good idea
Gerard de Melo's readme states:
From Wikipedia for ISO 639-3:
The etywn-relety.json contains the following proto_language references
It's probably best to just add add the relevant JSON and document accordingly, for instance 'p_sla' could be
no idea what to put for scope tbh