unicodedata module

Covers the most-used Unicode Character Database queries. Implemented functions match CPython 3.14 apart from the notes below.

Implemented

category, name, lookup, combining, normalize, is_normalized.

Constants: unidata_version.

Not implemented

decimal, digit, numeric (numeric-value tables), bidirectional, east_asian_width, mirrored, decomposition, and the ucd_3_2_0 object. These need Unicode property data the backing crates do not carry; accessing them raises AttributeError.

Behavioural notes

• Unicode version skew: unidata_version reports "16.0.0" to match CPython 3.14, but the backing crates carry independent data tables whose Unicode versions may differ from each other and from 16.0 (e.g. unicode-normalization currently ships Unicode 17.0 tables). Results for code points assigned, renamed, or recategorised across those versions may therefore diverge from CPython. Long-established code points (ASCII, common Latin/Greek, CJK) are unaffected.
• lookup resolves character names and some aliases, but named sequences that CPython accepts (e.g. "KEYCAP NUMBER SIGN") raise KeyError.
• lookup requires a str argument; CPython also accepts bytes. A non-str argument raises TypeError: "expected string, not <type>" rather than CPython's bytes-oriented message.
• All implemented functions are positional-only, but keyword-argument error wording does not match CPython. CPython raises unicodedata.<fn>() takes no keyword arguments; Monty reports the generated positional arity or unexpected-keyword error.