Ortographic profile

[tresoldi]

A profile for the entirety of IDS is complex, but some languages are low-hanging fruits, such as those with phonemic transcriptions (or phonemic-like, such as languages of Russia with a phonemic Cyrillic).

I'm starting working on it.

[LinguList]

I think it would be very helpful, @tresoldi, if we start with one profile and one class of examples. I have another profile almost ready for a subpart of IDS SEA languages, so if you look into a couple of cyrillic orthographies now and feed back e.g., also with Ilya, so he can have a quick look, I could later add the SEA profiles, and we'd be able to do this step by step. But I think you'll need to adjust the code in the IDS lexibank repo to leave languages unsegmented while segmenting others. Note also that in many cases, words are placed in [] brackets (at least in the SA part) to mark them as borrowings, which is a bit, well, strange, but it means one has to be careful with the form spec, and might even have to intervene here to extract borrowing information with intention, or the like (maybe in a first pass).

[tresoldi]

Great! But depends on lexibank/pylexibank#197

[LinguList]

Yes, I think, the alternative would be to pull out data and make a new dataset, but I rather like to lift particular datasets a bit more, also for ABVD, wehre I'd be keen on doing the SEA part, and one could add the polynesian part by Mary, etc.