regex
lxml
python-levenshtein
natsort
multiprocess
tqdm
orjson
black
msgspec
unidecode
lz4
lmdb
numpy~=2.3
numba~=0.63
scikit-learn~=1.8
pandas~=2.3
spacy~=3.8
netaddr

[test]
pytest>=7.0
pytest-timeout>=2.0
pytest-benchmark>=4.0

[transformers]
spacy-transformers~=1.3.8
