issue #63 include canonical language-code support with backward-compatible aliases

[georgeslabreche]

Added canonical language-code support with backward-compatible aliases:

• Canonical: be, bg, cnr, el, mk, mn, ru, sr, tg, uk.
• Legacy aliases still work: by, me, rs, tj, ua.
• ISO 639 alpha-3 aliases also work: bel, bul, ell, mkd, mon, rus, srp, tgk, ukr.

Key changes:

• Added language-code normalization in cyrtranslit/mapping/init.py.
• API normalizes codes before lookup and special Latin digraph handling in cyrtranslit/init.py.
• supported() now returns canonical codes by default, and supported(include_aliases=True) returns every accepted code.
• CLI validation now accepts canonical and alias codes.
• Added TDD coverage for canonical codes, ISO alpha-3 aliases, legacy aliases, uppercase code handling, and CLI canonical input in tests.py.
• Bumped project version to 1.3.0.