Re-normalize with nijsa 0.2.0 (bundled lexicon, typography fixes)

v2.1: add dah00_ref (book number in Дахкильгов 2000, fuzzy-matched)

card: valid task category (text-generation)

v2: normalize orthography (palochka U+04C0, Latin look-alikes), restore ru notes, fix RU typos, add group_id, English dataset card, license: other

Update README.md

Upload README.md with huggingface_hub

Upload data/train-00000-of-00001-b1684951ab370c73.parquet with huggingface_hub

initial commit