Hankan/lib/hangul.js

/* 한글 — decomposition, composition, and a 두벌식 input method.
   No dependencies. Lifted from the artifact; the IME is unit-tested
   against 먹어 · 왔어 · 읽어 · 괜찮아 · 값 · 의사 · 뭐야 and backspace. */

export const CHO  = "ㄱㄲㄴㄷㄸㄹㅁㅂㅃㅅㅆㅇㅈㅉㅊㅋㅌㅍㅎ";
export const JUNG = "ㅏㅐㅑㅒㅓㅔㅕㅖㅗㅘㅙㅚㅛㅜㅝㅞㅟㅠㅡㅢㅣ";
export const JONG = " ㄱㄲㄳㄴㄵㄶㄷㄹㄺㄻㄼㄽㄾㄿㅀㅁㅂㅄㅅㅆㅇㅈㅊㅋㅌㅍㅎ";

const VJOIN = {"ㅗㅏ":"ㅘ","ㅗㅐ":"ㅙ","ㅗㅣ":"ㅚ","ㅜㅓ":"ㅝ","ㅜㅔ":"ㅞ","ㅜㅣ":"ㅟ","ㅡㅣ":"ㅢ"};
const FJOIN = {"ㄱㅅ":"ㄳ","ㄴㅈ":"ㄵ","ㄴㅎ":"ㄶ","ㄹㄱ":"ㄺ","ㄹㅁ":"ㄻ","ㄹㅂ":"ㄼ",
               "ㄹㅅ":"ㄽ","ㄹㅌ":"ㄾ","ㄹㅍ":"ㄿ","ㄹㅎ":"ㅀ","ㅂㅅ":"ㅄ"};
const FSPLIT = Object.fromEntries(Object.entries(FJOIN).map(([k,v]) => [v, [k[0], k[1]]]));

/** [initialIndex, medialIndex, finalIndex] or null if not a syllable block. */
export function decompose(ch) {
  const c = ch.charCodeAt(0) - 0xAC00;
  if (c < 0 || c > 11171) return null;
  return [Math.floor(c / 588), Math.floor((c % 588) / 28), c % 28];
}
export function compose(i, m, f = 0) {
  return String.fromCharCode(0xAC00 + (i * 21 + m) * 28 + f);
}
export const isJamo = {
  initial: c => CHO.includes(c),
  medial:  c => JUNG.includes(c),
  final:   c => c && JONG.indexOf(c) > 0,
};

/* ── 두벌식 IME ───────────────────────────────────────────
   Hold one composing buffer per input. Feed jamo with key(),
   punctuation and spaces with text(), and Backspace with back().
   Each call returns the full new value. */
export class Composer {
  constructor() { this.reset(); }
  reset() { this.cho = null; this.jung = null; this.jong = null; }
  get empty() { return !this.cho && !this.jung && !this.jong; }

  /** The syllable currently being assembled, as text. */
  render() {
    if (this.cho && this.jung) {
      const i = CHO.indexOf(this.cho), m = JUNG.indexOf(this.jung);
      const f = this.jong ? JONG.indexOf(this.jong) : 0;
      if (i >= 0 && m >= 0 && f >= 0) return compose(i, m, f);
    }
    return (this.cho || "") + (this.jung || "") + (this.jong || "");
  }
  /** Strip the composing tail off a value so it can be rebuilt. */
  _base(value) {
    const cur = this.render();
    return cur && value.endsWith(cur) ? value.slice(0, -cur.length) : value;
  }

  key(value, j) {
    let base = this._base(value);
    if (isJamo.medial(j)) {
      if (this.jong) {                       // final splits off to start a new block
        const parts = FSPLIT[this.jong];
        let moved;
        if (parts) { this.jong = parts[0]; moved = parts[1]; }
        else { moved = this.jong; this.jong = null; }
        base += this.render();
        this.cho = moved; this.jung = j; this.jong = null;
      } else if (this.jung) {
        const join = VJOIN[this.jung + j];
        if (join) this.jung = join;
        else { base += this.render(); this.reset(); this.jung = j; }
      } else this.jung = j;
    } else {
      if (this.cho && this.jung) {
        if (this.jong) {
          const join = FJOIN[this.jong + j];
          if (join) this.jong = join;
          else { base += this.render(); this.reset(); this.cho = j; }
        } else if (isJamo.final(j)) this.jong = j;
        else { base += this.render(); this.reset(); this.cho = j; }
      } else {
        if (!this.empty) base += this.render();
        this.reset(); this.cho = j;
      }
    }
    return base + this.render();
  }

  back(value) {
    let base = this._base(value);
    if (this.jong) {
      const parts = FSPLIT[this.jong];
      this.jong = parts ? parts[0] : null;
    } else if (this.jung) {
      let peeled = null;
      for (const [k, v] of Object.entries(VJOIN)) if (v === this.jung) peeled = k[0];
      this.jung = peeled;
    } else if (this.cho) {
      this.cho = null;
    } else {                                  // pull a finished block back in
      const last = base.slice(-1);
      base = base.slice(0, -1);
      const d = last ? decompose(last) : null;
      if (d) {
        this.cho = CHO[d[0]]; this.jung = JUNG[d[1]];
        this.jong = d[2] ? JONG[d[2]] : null;
        if (this.jong) { const p = FSPLIT[this.jong]; this.jong = p ? p[0] : null; }
        else this.jung = null;
      }
    }
    return base + this.render();
  }

  /** Commit the buffer and append literal text (space, punctuation). */
  text(value, t) {
    const base = this._base(value) + this.render();
    this.reset();
    return base + t;
  }
}

/** Standard 두벌식 layout, top row first. Shift gives the tense pairs. */
export const KEYBOARD = {
  rows: [
    ["ㅂ","ㅈ","ㄷ","ㄱ","ㅅ","ㅛ","ㅕ","ㅑ","ㅐ","ㅔ"],
    ["ㅁ","ㄴ","ㅇ","ㄹ","ㅎ","ㅗ","ㅓ","ㅏ","ㅣ"],
    ["ㅋ","ㅌ","ㅊ","ㅍ","ㅠ","ㅜ","ㅡ"],
  ],
  shift: {"ㅂ":"ㅃ","ㅈ":"ㅉ","ㄷ":"ㄸ","ㄱ":"ㄲ","ㅅ":"ㅆ","ㅐ":"ㅒ","ㅔ":"ㅖ"},
};