awaforge

unicode

Apache-2.0Release 1L3

Native Unicode helpers — normalization, casefold, collation, grapheme clusters, width, strip diacritics.

Module unicode | Source packages/front/fw/src/io/text/unicode.js | Deps none | Worker-safe yes

Everything delegated to String.prototype and Intl.* — no embedded custom tables.

Resolve

const unicode = runtime.resolve('unicode');
// Returns: { normalize, casefold, compare, collator, graphemes, width, stripDiacritics }

API

Method Signature Returns
normalize (str: string, form?: string) => string NFC/NFD/NFKC/NFKD
casefold (str: string, locale?: string) => string Lowercase locale-aware
compare (a, b, options?) => number One-shot Intl.Collator.compare
collator (options?) => {compare, sort} Reusable collator
graphemes (str: string) => Iterator<string> Grapheme clusters via Intl.Segmenter
width (str: string) => number Visual width (ASCII=1, CJK=2)
stripDiacritics (str: string) => string Removes diacritic marks

Examples

Normalization

const unicode = runtime.resolve('unicode');

unicode.normalize('café', 'NFD').length; // 5 (e + combining)
unicode.normalize('café', 'NFC').length; // 4 (precomposed)

unicode.stripDiacritics('café');    // 'cafe'
unicode.stripDiacritics('Ångström'); // 'Angstrom'

Collation and sorting

const coll = unicode.collator({ numeric: true });
coll.sort(['a10', 'a2', 'a1']); // ['a1', 'a2', 'a10']

unicode.compare('é', 'e', { sensitivity: 'base' }); // 0 (equivalent)

Grapheme clusters

[...unicode.graphemes('👨‍👩‍👧')].length; // 1 (single family unit)
unicode.width('日本');  // 4 (2 CJK × 2)
unicode.width('hello'); // 5

Worker Usage

const worker = fw.createWorker(
    function ({ libs, args }) {
        return libs.unicode.stripDiacritics(args[0]);
    },
    { dependencies: ['unicode'], args: ['café'] }
);

Notes

  • graphemes uses Intl.Segmenter if available; falls back to code points (ZWJ sequences not merged).
  • width is heuristic — for exact console rendering, refer to the official EastAsianWidth tables.
  • stripDiacritics does not translate ß → ss (Unicode decomposition does not decompose ß).
  • normalize throws if str is not a string.

See also

  • str — slug and case conversions (uses unicode)
  • i18n — locale-aware formatting