Release 0.2.3

Hi everyone! After a long while, I now have some time to do some maintenance release for calamanCy!
Basically, this release changes how models will be downloaded, and just modernizes how we install this library.

Previously, calamancy.load() installed model wheels via pip at runtime, which failed in environments without pip, notably uv-managed projects. So uv is now our dependency manager for both installation and dev.
Also, models are now fetched with huggingface_hub and loaded straight from your local Hugging Face cache.

import calamancy
nlp = calamancy.load("tl_calamancy_md")  # bare names resolve to the latest version

Finally, stay tuned for 0.3.0! I'm hoping to train better models given that the ecosystem has improved since last year.
This might also be the first version where we'll train dedicated models for *wink* *wink* other major Philippine languages!