pchr8 / pymorphy-spacy-disambiguation

A package that picks the correct pymorphy2 morphology analysis based on morphology data from spacy
MIT License
5 stars 0 forks source link

Replace pymorphy2 with pymorphy3 #7

Open mariana-scorp opened 5 months ago

mariana-scorp commented 5 months ago

Pymorphy3 uses a much more recent version of Ukrainian dictionaries (VESUM).

https://github.com/no-plagiarism/pymorphy3

pchr8 commented 5 months ago

Currently it supports using a MorphAnalyzer instance passed during init (the intent behind that was so a pymorphy3 one can be passed if needed), but it does use pymorphy2 under the hood if nothing is specified, and defaults are important. Definitely a good idea to change this

(Or it could detect which pymorphy version is already installed, and prefer pymorphy3 in case both are — either way, defaulting to pymorphy2 is not the right way to do)

Thank you for bringing that up!

pchr8 commented 5 months ago

Or we could do this: https://github.com/explosion/spaCy/blob/master/spacy/lang/uk/lemmatizer.py