@kensio/pinyinjs 0.1.0-beta.0 → 1.0.0

package/dist/transcription/gwoyeu.js

@@ -0,0 +1,491 @@
+/**
+ * Gwoyeu Romatzyh (國語羅馬字), in both directions.
+ *
+ * GR is the odd one of the five. Every other system here writes a syllable and
+ * then marks its tone — a diacritic, a raised digit, a tone letter — so the
+ * spelling is one mapping over the inventory and the tone rides on top of it.
+ * GR spells the tone *into* the syllable: 山 shān is `shan`, 陝 shǎn is `shaan`,
+ * 善 shàn is `shann`. So it is four mappings rather than one, related by the
+ * tonal rules below, and reading it back is an index keyed per tone.
+ *
+ * The rules come from Wikipedia's *Spelling in Gwoyeu Romatzyh*, which states
+ * them as a rule of thumb per tone plus a handful of clauses. They are
+ * implemented here as stated, with one amendment recorded in
+ * {@link zeroInitial}, and the whole of it is scored against a different page's
+ * four GR columns in [syllabary.test.ts](syllabary.test.ts).
+ *
+ * Three things follow from the tone being in the spelling:
+ *
+ * - **There is no toneless form.** The basic form *is* the first tone, so a
+ *   syllable whose tone was never written cannot be told from one written in
+ *   the first tone. That is bopomofo's shortfall exactly, arrived at from the
+ *   other direction — see `docs/romanization/`.
+ * - **The neutral tone is a dot in front**, and GR keeps the etymological tone
+ *   spelling of the syllable behind it (`perng.yeou` for 朋友). A neutral
+ *   pinyin syllable does not record what that tone was, so this module writes
+ *   the basic form: `.iou` where a GR writer has `.yeou`. Reading is not so
+ *   limited — a dot in front of any tonal spelling reads as the neutral tone.
+ * - **儿化 is the `-l` suffix**, and it is written here as a plain suffix
+ *   rather than as the fusion GR actually performs. See {@link ERHUA_SUFFIX}.
+ */
+import { DICTIONARY_SYLLABLES } from "../syllable/inventory.js";
+import { readSyllable } from "../syllable/syllable.js";
+import { NEUTRAL_TONE } from "../tone/tone.js";
+/**
+ * The four tones, for building the index.
+ */
+const CONTOUR_TONES = [1, 2, 3, 4];
+/**
+ * How each initial is spelled.
+ *
+ * GR uses the voiced letters for the unaspirated series, as pinyin does, so
+ * this half of the table is nearly pinyin's own. The three that differ are the
+ * alveolar affricates — `tz` and `ts` for pinyin z and c — and the palatals,
+ * which GR does not distinguish from the retroflexes at all: `j`, `ch` and `sh`
+ * stand for both series, and it is the following `i` that says which. That is
+ * why pinyin zhu, ju and jiu come out as `ju`, `jiu` and `jiou`.
+ */
+const INITIAL_SPELLINGS = new Map([
+    ["b", "b"],
+    ["p", "p"],
+    ["m", "m"],
+    ["f", "f"],
+    ["d", "d"],
+    ["t", "t"],
+    ["n", "n"],
+    ["l", "l"],
+    ["g", "g"],
+    ["k", "k"],
+    ["h", "h"],
+    ["j", "j"],
+    ["q", "ch"],
+    ["x", "sh"],
+    ["zh", "j"],
+    ["ch", "ch"],
+    ["sh", "sh"],
+    ["r", "r"],
+    ["z", "tz"],
+    ["c", "ts"],
+    ["s", "s"],
+]);
+/**
+ * How each final is spelled in the basic form.
+ *
+ * Written out in full, as the finals table of *Spelling in Gwoyeu Romatzyh*
+ * has them: `iou` and `uei` rather than pinyin's contracted iu and ui, `iu` for
+ * ü, `au` for ao, and `el` for er — the `-r` being spoken for by the second
+ * tone.
+ */
+const FINAL_SPELLINGS = {
+    a: "a",
+    o: "o",
+    e: "e",
+    ê: "è",
+    er: "el",
+    ai: "ai",
+    ei: "ei",
+    ao: "au",
+    ou: "ou",
+    an: "an",
+    en: "en",
+    ang: "ang",
+    eng: "eng",
+    ong: "ong",
+    i: "i",
+    ia: "ia",
+    io: "io",
+    ie: "ie",
+    iao: "iau",
+    iou: "iou",
+    ian: "ian",
+    in: "in",
+    iang: "iang",
+    ing: "ing",
+    iong: "iong",
+    u: "u",
+    ua: "ua",
+    uo: "uo",
+    uai: "uai",
+    uei: "uei",
+    uan: "uan",
+    uen: "uen",
+    uang: "uang",
+    ueng: "ueng",
+    ü: "iu",
+    üe: "iue",
+    üan: "iuan",
+    ün: "iun",
+    m: "m",
+    n: "n",
+    ng: "ng",
+};
+/**
+ * The initials after which the empty rhyme is written `-y`.
+ *
+ * 詩 shī is `shy` and 斯 sī is `sy`, which is the same observation Yale makes
+ * with its `r` and `z`: the i of shi is a continuation of the initial rather
+ * than a vowel. GR spends a letter of its own on it, and that letter behaves
+ * as a vowel for every tonal rule below — `shy`, `shyr`, `shyy`, `shyh`.
+ */
+const EMPTY_RHYME_INITIALS = new Set([
+    "zh",
+    "ch",
+    "sh",
+    "r",
+    "z",
+    "c",
+    "s",
+]);
+/**
+ * The sonorant initials, which swap what the first two tones do.
+ *
+ * A syllable beginning l-, m-, n- or r- takes an `-h-` as its second letter in
+ * the first tone and is left alone in the second: 媽 mā is `mha` and 麻 má is
+ * `ma`. The reason is frequency rather than phonology — these initials carry
+ * far more second tones than first ones — and it is the one place where the
+ * basic form is not the first tone.
+ */
+const SONORANT_INITIALS = new Set(["l", "m", "n", "r"]);
+/**
+ * The letters that count as a vowel for the tonal rules.
+ *
+ * `y` is the empty rhyme and `è` is pinyin's ê; both are the whole of the rime
+ * they appear in, so both take the doubling and the `-r` exactly as a vowel
+ * does.
+ */
+const VOWELS = "aeiouyè";
+/**
+ * The vowels a tone mark can sit on in pinyin, in the order the rule tries
+ * them, which is what "the main vowel" means for the third tone's doubling.
+ */
+const MAIN_VOWELS = ["a", "o", "e"];
+/**
+ * The 儿化 suffix.
+ *
+ * GR marks rhotacisation with `-l`, and this module writes it as a plain
+ * suffix: 花儿 huār is `hual`. **What GR itself does is a fusion rather than a
+ * suffix** — 一點兒 yìdiǎnr is `ideal`, where `dean` has lost its `-n` — and
+ * the fusion collapses syllables that were distinct, which is why *Spelling in
+ * Gwoyeu Romatzyh* notes that `jiel` is either 今兒 or 雞兒. Modelling it needs
+ * a rime-by-rime table of rhotacised forms, which is a phonological claim
+ * rather than a spelling convention, and is not made here — the same line
+ * [ipa.ts](ipa.ts) draws around its suffixed [ɚ].
+ *
+ * The visible consequence is that this module's 儿化 always reads back and
+ * GR's own does not.
+ */
+const ERHUA_SUFFIX = "l";
+/**
+ * The dot written in front of a neutral-tone syllable.
+ *
+ * A full stop, which is how GR writes it in print. Chao's own practice is to
+ * use it on a word's first appearance rather than throughout; this module
+ * always writes it, because it is the only thing in GR that can say "neutral"
+ * rather than leaving it to be inferred.
+ */
+const NEUTRAL_MARK = ".";
+/**
+ * Whether a letter is one the tonal rules treat as a vowel.
+ *
+ * Spelled out rather than written `VOWELS.includes(letter)` because
+ * `"aeiouyè".includes("")` is true, and the letter before the first one of a
+ * spelling is exactly that: the empty string. A rule asking "is the letter
+ * before this one a vowel" would silently say yes at the start of every
+ * syllable.
+ */
+function isVowel(letter) {
+    return letter !== undefined && letter !== "" && VOWELS.includes(letter);
+}
+/**
+ * How a final is spelled after a given initial.
+ */
+function rimeSpelling(initial, final) {
+    return final === "i" && EMPTY_RHYME_INITIALS.has(initial)
+        ? "y"
+        : FINAL_SPELLINGS[final];
+}
+/**
+ * The basic form of a syllable: the spelling every tonal rule starts from.
+ */
+function basicForm(syllable) {
+    const { initial, final } = syllable;
+    const spelt = INITIAL_SPELLINGS.get(initial) ?? "";
+    return `${spelt}${rimeSpelling(initial, final)}`;
+}
+/**
+ * Where the rime's vowels start and end.
+ *
+ * The five syllables with no vowel letter at all — the syllabic nasals 呣, 唔,
+ * 嗯 and their aspirated pair — fall back to the last letter, so that the rules
+ * asking for "the vowel" have something to work on. GR has no attested
+ * spelling for any of them; see `docs/romanization/`.
+ */
+function vowelSpan(form) {
+    let start = 0;
+    while (start < form.length && !isVowel(form[start])) {
+        start += 1;
+    }
+    if (start === form.length) {
+        return [form.length - 1, form.length - 1];
+    }
+    let end = start;
+    while (isVowel(form[end + 1])) {
+        end += 1;
+    }
+    return [start, end];
+}
+/**
+ * Tone 1: the basic form, with `-h-` inserted after a sonorant initial.
+ */
+function firstTone(form, initial) {
+    return SONORANT_INITIALS.has(initial)
+        ? `${form.slice(0, 1)}h${form.slice(1)}`
+        : form;
+}
+/**
+ * The i or u a glide can be written on, turned into y or w.
+ *
+ * The rule is `NiV → NyV` and `NuV → NwV`, where N is a non-vowel or nothing at
+ * all: the i or u has to be the medial rather than part of a diphthong, which
+ * is what the "not preceded by a vowel" test says. The `+ -i if final` clause
+ * covers the rimes that are nothing but the medial — `i` becomes `yi` and not
+ * a bare `y`.
+ */
+function glided(form) {
+    for (const [letter, glide] of [
+        ["i", "y"],
+        ["u", "w"],
+    ]) {
+        const at = form.indexOf(letter);
+        if (at !== -1 && !isVowel(form[at - 1])) {
+            const rest = form.slice(at + 1);
+            return `${form.slice(0, at)}${glide}${rest === "" ? letter : rest}`;
+        }
+    }
+    return undefined;
+}
+/**
+ * Tone 2: i/u to y/w, or an `-r` after the vowels.
+ */
+function secondTone(form, initial) {
+    if (SONORANT_INITIALS.has(initial)) {
+        return form;
+    }
+    const glide = glided(form);
+    if (glide !== undefined) {
+        return glide;
+    }
+    const [, end] = vowelSpan(form);
+    return `${form.slice(0, end + 1)}r${form.slice(end + 1)}`;
+}
+/**
+ * The i or u a third tone can be swapped for e or o.
+ *
+ * `Vi`/`iV` becomes `Ve`/`eV` and `Vu`/`uV` becomes `Vo`/`oV`, so the letter
+ * has to sit beside a vowel; where both an i and a u are available it is
+ * whichever comes *first* that changes, which is why 交 jiǎo is `jeau` and not
+ * `jiao`. The swap is abandoned where it would produce `ee` or `oo`, both of
+ * which are the doubling below and would collide with it.
+ */
+function swapped(form) {
+    const found = [
+        ["i", "e"],
+        ["u", "o"],
+    ]
+        .map(([letter, swap]) => [form.indexOf(letter), swap])
+        .filter(([at]) => at !== -1)
+        .toSorted((one, other) => one[0] - other[0])[0];
+    if (found === undefined) {
+        return undefined;
+    }
+    const [at, swap] = found;
+    const before = form[at - 1];
+    const after = form[at + 1];
+    const isBeside = isVowel(before) || isVowel(after);
+    const isDoubles = before === swap || after === swap;
+    return isBeside && !isDoubles
+        ? `${form.slice(0, at)}${swap}${form.slice(at + 1)}`
+        : undefined;
+}
+/**
+ * Tone 3, where the swap does not apply: double the main vowel.
+ *
+ * The main vowel is the one a pinyin tone mark would sit on — a, then o or e,
+ * then the last vowel there is — so 請 qǐng is `chiing` and 給 gěi is `geei`.
+ */
+function doubled(form) {
+    const [start, end] = vowelSpan(form);
+    const rime = form.slice(start, end + 1);
+    const main = MAIN_VOWELS.map((vowel) => rime.indexOf(vowel)).find((at) => at !== -1);
+    const at = start + (main ?? rime.length - 1);
+    return `${form.slice(0, at + 1)}${form.slice(at, at + 1)}${form.slice(at + 1)}`;
+}
+/**
+ * Tone 3: i/u to e/o beside a vowel, or the main vowel doubled.
+ */
+function thirdTone(form) {
+    return swapped(form) ?? doubled(form);
+}
+/**
+ * Tone 4: change or double the last letter, or add `-h`.
+ *
+ * The `⇏iw` clause is why 去 qù is `chiuh` and not `chiw`: the u of `iu` is the
+ * ü rather than the second half of a diphthong, and turning it into a w would
+ * say the wrong thing about the vowel.
+ */
+function fourthTone(form) {
+    const last = form.slice(-1);
+    const before = form.slice(-2, -1);
+    if (last === "i" && isVowel(before)) {
+        return `${form.slice(0, -1)}y`;
+    }
+    if (last === "u" && isVowel(before) && before !== "i") {
+        return `${form.slice(0, -1)}w`;
+    }
+    if (form.endsWith("ng")) {
+        return `${form.slice(0, -2)}nq`;
+    }
+    if (last === "n" || last === "l") {
+        return `${form}${last}`;
+    }
+    return `${form}h`;
+}
+/**
+ * Put the y- or w- on a syllable that has no initial, in tones 3 and 4.
+ *
+ * **This is the one place where the published rules needed amending, and the
+ * amendment is a correction of one word.** They say that a basic form starting
+ * i- or u- has that letter *replaced* by y- or w- (tone 4), or changed to e- or
+ * o- with a y- or w- added (tone 3). Applied literally that deletes a vowel the
+ * attested spellings keep: 一 yī's basic form is `i`, and tone 4 comes out `yh`
+ * where the source has `yih`. Nine cells of the syllabary fail that way, all of
+ * them a rime that is bare or closed by a consonant — `yii`, `yih`, `wuu`,
+ * `wuh`, `yiin`, `yinn`, `yiing`, `yinq`, `yuh`.
+ *
+ * The rime table of *Spelling in Gwoyeu Romatzyh* says what actually happens,
+ * and it is one rule rather than two: **the letter replaced is a medial**, so
+ * it goes only when a different vowel follows it. `iuh` has a u after the i and
+ * becomes `yuh`; `ih` has an h and becomes `yih`; `ii` has the same vowel
+ * doubled, which is not a following vowel at all, and becomes `yii`. Tone 3
+ * needs no separate clause once that is said, because its own swap has already
+ * eaten the medial wherever there was one.
+ */
+function zeroInitial(form, basic) {
+    const onset = basic.slice(0, 1);
+    if (onset !== "i" && onset !== "u") {
+        return form;
+    }
+    const glide = onset === "i" ? "y" : "w";
+    const next = form[1];
+    const isMedial = form.slice(0, 1) === onset && isVowel(next) && next !== onset;
+    return `${glide}${isMedial ? form.slice(1) : form}`;
+}
+/**
+ * Spell a syllable in a given tone, which is the whole of GR's tonal rules.
+ */
+function toneForm(syllable, tone) {
+    const { initial } = syllable;
+    const basic = basicForm(syllable);
+    if (tone === 1) {
+        return firstTone(basic, initial);
+    }
+    if (tone === 2) {
+        return secondTone(basic, initial);
+    }
+    const changed = tone === 3 ? thirdTone(basic) : fourthTone(basic);
+    return initial === "" ? zeroInitial(changed, basic) : changed;
+}
+/**
+ * Write a syllable in Gwoyeu Romatzyh: 就 jiù becomes `jiow`.
+ *
+ * An unwritten tone is written as the basic form, because GR has no other
+ * form to write — which means it reads back as a first tone, exactly as
+ * bopomofo's unmarked syllable does. The neutral tone takes the dot in front,
+ * and the syllable behind it is the basic form rather than the etymological
+ * spelling GR would use, for the reason the module comment gives.
+ */
+export function writeGwoyeu(syllable) {
+    const { tone } = syllable;
+    const spelt = toneForm(syllable, tone === undefined || tone === NEUTRAL_TONE ? 1 : tone);
+    const suffix = syllable.erhua === true ? ERHUA_SUFFIX : "";
+    const dot = tone === NEUTRAL_TONE ? NEUTRAL_MARK : "";
+    return `${dot}${spelt}${suffix}`;
+}
+/**
+ * Write a word in Gwoyeu Romatzyh, one syllable after another.
+ *
+ * Solid, as GR writes a word: 北京 is `Beeijing`, which is the convention
+ * pinyin inherited. GR uses an apostrophe where the join is ambiguous — its own
+ * name for pinyin, `Pin'in`, is the standard example — and that is not written
+ * here, for the same reason `readGwoyeu` takes one syllable at a time: knowing
+ * where the boundary is ambiguous means being able to split the word, and
+ * nothing here splits a GR word.
+ */
+export function writeGwoyeuWord(syllables) {
+    return syllables.map((syllable) => writeGwoyeu(syllable)).join("");
+}
+/**
+ * Every syllable each spelling stands for, keyed per tone.
+ *
+ * The other readers here index the inventory once and read the tone off the
+ * marks; GR has no marks, so the index is built four times over and each entry
+ * carries the tone its spelling means. That is the whole practical difference
+ * between reading GR and reading the rest.
+ */
+function indexTonalForms() {
+    const index = new Map();
+    for (const pinyin of DICTIONARY_SYLLABLES) {
+        const syllable = readSyllable(pinyin);
+        /* c8 ignore next 3 -- inventory.test.ts holds the parser to the inventory */
+        if (syllable === undefined) {
+            continue;
+        }
+        for (const tone of CONTOUR_TONES) {
+            const toned = { ...syllable, tone };
+            const key = toneForm(syllable, tone);
+            const found = index.get(key);
+            if (found === undefined) {
+                index.set(key, [toned]);
+            }
+            else {
+                found.push(toned);
+            }
+        }
+    }
+    return index;
+}
+const INDEX = indexTonalForms();
+/**
+ * Look a spelling up, allowing for the 儿化 suffix on the end of it.
+ *
+ * Tried second and separately, as it is for Yale: `ell` is 二 èr, and it is also
+ * the same rime in the first tone with the suffix on the end. Both readings are
+ * real, and it is the caller's to narrow.
+ */
+function lookUp(spelling) {
+    const base = spelling.endsWith(ERHUA_SUFFIX) ? spelling.slice(0, -1) : "";
+    const suffixed = base === ""
+        ? []
+        : (INDEX.get(base) ?? []).map((syllable) => ({
+            ...syllable,
+            erhua: true,
+        }));
+    return [...(INDEX.get(spelling) ?? []), ...suffixed];
+}
+/**
+ * Read a Gwoyeu Romatzyh syllable: `jiow` becomes 就 jiù.
+ *
+ * A leading dot is the neutral tone, and it wins over whatever tone the
+ * spelling behind it carries — `.yeou` is a neutral 友, which is what GR means
+ * by it. Returns every syllable the spelling stands for, and nothing at all for
+ * a spelling no syllable of the inventory writes.
+ */
+export function readGwoyeu(text) {
+    const written = text.trim().normalize("NFC").toLowerCase();
+    const isNeutral = written.startsWith(NEUTRAL_MARK);
+    const found = lookUp(isNeutral ? written.slice(NEUTRAL_MARK.length) : written);
+    return isNeutral
+        ? found.map((syllable) => ({ ...syllable, tone: NEUTRAL_TONE }))
+        : found;
+}
+//# sourceMappingURL=gwoyeu.js.map

package/dist/transcription/gwoyeu.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"gwoyeu.js","sourceRoot":"","sources":["../../src/transcription/gwoyeu.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AACH,OAAO,EAAE,oBAAoB,EAAE,MAAM,0BAA0B,CAAC;AAEhE,OAAO,EAAE,YAAY,EAAiB,MAAM,yBAAyB,CAAC;AACtE,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAU/C;;GAEG;AACH,MAAM,aAAa,GAA2B,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;AAE3D;;;;;;;;;GASG;AACH,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAkB;IACjD,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,IAAI,CAAC;IACX,CAAC,GAAG,EAAE,IAAI,CAAC;IACX,CAAC,IAAI,EAAE,GAAG,CAAC;IACX,CAAC,IAAI,EAAE,IAAI,CAAC;IACZ,CAAC,IAAI,EAAE,IAAI,CAAC;IACZ,CAAC,GAAG,EAAE,GAAG,CAAC;IACV,CAAC,GAAG,EAAE,IAAI,CAAC;IACX,CAAC,GAAG,EAAE,IAAI,CAAC;IACX,CAAC,GAAG,EAAE,GAAG,CAAC;CACX,CAAC,CAAC;AAEH;;;;;;;GAOG;AACH,MAAM,eAAe,GAA0B;IAC7C,CAAC,EAAE,GAAG;IACN,CAAC,EAAE,GAAG;IACN,CAAC,EAAE,GAAG;IACN,CAAC,EAAE,GAAG;IACN,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,CAAC,EAAE,GAAG;IACN,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,IAAI;IACR,IAAI,EAAE,MAAM;IACZ,GAAG,EAAE,KAAK;IACV,IAAI,EAAE,MAAM;IACZ,CAAC,EAAE,GAAG;IACN,EAAE,EAAE,IAAI;IACR,EAAE,EAAE,IAAI;IACR,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,IAAI,EAAE,MAAM;IACZ,IAAI,EAAE,MAAM;IACZ,CAAC,EAAE,IAAI;IACP,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,MAAM;IACX,EAAE,EAAE,KAAK;IACT,CAAC,EAAE,GAAG;IACN,CAAC,EAAE,GAAG;IACN,EAAE,EAAE,IAAI;CACT,CAAC;AAEF;;;;;;;GAOG;AACH,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAAU;IAC5C,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,GAAG;IACH,GAAG;IACH,GAAG;CACJ,CAAC,CAAC;AAEH;;;;;;;;GAQG;AACH,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAU,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEjE;;;;;;GAMG;AACH,MAAM,MAAM,GAAG,SAAS,CAAC;AAEzB;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAU,CAAC;AAE7C;;;;;;;;;;;;;;GAcG;AACH,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;;;;;GAOG;AACH,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;;;;;;GAQG;AACH,SAAS,OAAO,CAAC,MAA0B;IACzC,OAAO,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,EAAE,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;AAC1E,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,OAAgB,EAAE,KAAY;IAClD,OAAO,KAAK,KAAK,GAAG,IAAI,oBAAoB,CAAC,GAAG,CAAC,OAAO,CAAC;QACvD,CAAC,CAAC,GAAG;QACL,CAAC,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,SAAS,CAAC,QAAkB;IACnC,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,GAAG,QAAQ,CAAC;IACpC,MAAM,KAAK,GAAG,iBAAiB,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;IACnD,OAAO,GAAG,KAAK,GAAG,YAAY,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC;AACnD,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,SAAS,CAAC,IAAY;IAC7B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QACpD,KAAK,IAAI,CAAC,CAAC;IACb,CAAC;IACD,IAAI,KAAK,KAAK,IAAI,CAAC,MAAM,EAAE,CAAC;QAC1B,OAAO,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5C,CAAC;IACD,IAAI,GAAG,GAAG,KAAK,CAAC;IAChB,OAAO,OAAO,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;QAC9B,GAAG,IAAI,CAAC,CAAC;IACX,CAAC;IACD,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;AACtB,CAAC;AAED;;GAEG;AACH,SAAS,SAAS,CAAC,IAAY,EAAE,OAAgB;IAC/C,OAAO,iBAAiB,CAAC,GAAG,CAAC,OAAO,CAAC;QACnC,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE;QACxC,CAAC,CAAC,IAAI,CAAC;AACX,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,MAAM,CAAC,IAAY;IAC1B,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI;QAC5B,CAAC,GAAG,EAAE,GAAG,CAAC;QACV,CAAC,GAAG,EAAE,GAAG,CAAC;KACF,EAAE,CAAC;QACX,MAAM,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAChC,IAAI,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;YACxC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;YAChC,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,KAAK,GAAG,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACtE,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,IAAY,EAAE,OAAgB;IAChD,IAAI,iBAAiB,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;QACnC,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,CAAC,EAAE,GAAG,CAAC,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IAChC,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC;AAC5D,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,OAAO,CAAC,IAAY;IAC3B,MAAM,KAAK,GACT;QACE,CAAC,GAAG,EAAE,GAAG,CAAC;QACV,CAAC,GAAG,EAAE,GAAG,CAAC;KAEb;SACE,GAAG,CAAC,CAAC,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,IAAI,CAAU,CAAC;SAC9D,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC;SAC3B,QAAQ,CAAC,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAClD,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,CAAC,EAAE,EAAE,IAAI,CAAC,GAAG,KAAK,CAAC;IACzB,MAAM,MAAM,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC5B,MAAM,KAAK,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC3B,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,OAAO,CAAC,KAAK,CAAC,CAAC;IACnD,MAAM,SAAS,GAAG,MAAM,KAAK,IAAI,IAAI,KAAK,KAAK,IAAI,CAAC;IACpD,OAAO,QAAQ,IAAI,CAAC,SAAS;QAC3B,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE;QACpD,CAAC,CAAC,SAAS,CAAC;AAChB,CAAC;AAED;;;;;GAKG;AACH,SAAS,OAAO,CAAC,IAAY;IAC3B,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IACrC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC;IACxC,MAAM,IAAI,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAC/D,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,CAClB,CAAC;IACF,MAAM,EAAE,GAAG,KAAK,GAAG,CAAC,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC7C,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;AAClF,CAAC;AAED;;GAEG;AACH,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,OAAO,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;AACxC,CAAC;AAED;;;;;;GAMG;AACH,SAAS,UAAU,CAAC,IAAY;IAC9B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAClC,IAAI,IAAI,KAAK,GAAG,IAAI,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QACpC,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;IACjC,CAAC;IACD,IAAI,IAAI,KAAK,GAAG,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,MAAM,KAAK,GAAG,EAAE,CAAC;QACtD,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;IACjC,CAAC;IACD,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC;QACxB,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;IAClC,CAAC;IACD,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;QACjC,OAAO,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC;IAC1B,CAAC;IACD,OAAO,GAAG,IAAI,GAAG,CAAC;AACpB,CAAC;AAED;;;;;;;;;;;;;;;;;;;GAmBG;AACH,SAAS,WAAW,CAAC,IAAY,EAAE,KAAa;IAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IAChC,IAAI,KAAK,KAAK,GAAG,IAAI,KAAK,KAAK,GAAG,EAAE,CAAC;QACnC,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,KAAK,GAAG,KAAK,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;IACxC,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;IACrB,MAAM,QAAQ,GACZ,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,KAAK,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,KAAK,CAAC;IAChE,OAAO,GAAG,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;AACtD,CAAC;AAED;;GAEG;AACH,SAAS,QAAQ,CAAC,QAAkB,EAAE,IAAiB;IACrD,MAAM,EAAE,OAAO,EAAE,GAAG,QAAQ,CAAC;IAC7B,MAAM,KAAK,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAC;IAClC,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC;QACf,OAAO,SAAS,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACnC,CAAC;IACD,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC;QACf,OAAO,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACpC,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;IAClE,OAAO,OAAO,KAAK,EAAE,CAAC,CAAC,CAAC,WAAW,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC;AAChE,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,WAAW,CAAC,QAAkB;IAC5C,MAAM,EAAE,IAAI,EAAE,GAAG,QAAQ,CAAC;IAC1B,MAAM,KAAK,GAAG,QAAQ,CACpB,QAAQ,EACR,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,YAAY,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CACvD,CAAC;IACF,MAAM,MAAM,GAAG,QAAQ,CAAC,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,CAAC;IAC3D,MAAM,GAAG,GAAG,IAAI,KAAK,YAAY,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,CAAC;IACtD,OAAO,GAAG,GAAG,GAAG,KAAK,GAAG,MAAM,EAAE,CAAC;AACnC,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,eAAe,CAAC,SAA8B;IAC5D,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACrE,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,eAAe;IACtB,MAAM,KAAK,GAAG,IAAI,GAAG,EAAsB,CAAC;IAC5C,KAAK,MAAM,MAAM,IAAI,oBAAoB,EAAE,CAAC;QAC1C,MAAM,QAAQ,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;QACtC,6EAA6E;QAC7E,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,SAAS;QACX,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;YACjC,MAAM,KAAK,GAAG,EAAE,GAAG,QAAQ,EAAE,IAAI,EAAE,CAAC;YACpC,MAAM,GAAG,GAAG,QAAQ,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;YACrC,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAC7B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;gBACxB,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;YAC1B,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACpB,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,KAAK,GAAG,eAAe,EAAE,CAAC;AAEhC;;;;;;GAMG;AACH,SAAS,MAAM,CAAC,QAAgB;IAC9B,MAAM,IAAI,GAAG,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC1E,MAAM,QAAQ,GACZ,IAAI,KAAK,EAAE;QACT,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;YACzC,GAAG,QAAQ;YACX,KAAK,EAAE,IAAI;SACZ,CAAC,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,EAAE,GAAG,QAAQ,CAAC,CAAC;AACvD,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,UAAU,CAAC,IAAY;IACrC,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;IAC3D,MAAM,SAAS,GAAG,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,CAAC;IACnD,MAAM,KAAK,GAAG,MAAM,CAClB,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CACzD,CAAC;IACF,OAAO,SAAS;QACd,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,QAAQ,EAAE,IAAI,EAAE,YAAY,EAAE,CAAC,CAAC;QAChE,CAAC,CAAC,KAAK,CAAC;AACZ,CAAC"}

package/dist/transcription/ipa.d.ts

@@ -0,0 +1,40 @@
+import type { Syllable } from "../syllable/syllable.js";
+/**
+ * How an IPA syllable writes its tone.
+ *
+ * Chao's tone letters are the standard notation and the default; `numbers`
+ * writes his pitch numerals instead (55, 35, 214, 51), which is what a plain
+ * ASCII field wants.
+ */
+export interface IpaOptions {
+    readonly tones?: "letters" | "numbers" | "none";
+}
+/**
+ * Transcribe a syllable without its tone: 就 jiù becomes `tɕiou`.
+ */
+export declare function writeIpaSymbols(syllable: Syllable): string;
+/**
+ * Transcribe a syllable in IPA: 就 jiù becomes `tɕiou˥˩`.
+ *
+ * Without brackets, since whether it is phonemic or phonetic is the caller's
+ * claim to make and not this table's.
+ */
+export declare function writeIpa(syllable: Syllable, options?: IpaOptions): string;
+/**
+ * Transcribe a word, one syllable after another.
+ *
+ * Run together, as a transcription of connected speech is written. The tone
+ * letters mark where each syllable ends, so nothing is lost by it — except for
+ * the neutral tone, which writes nothing and so leaves a boundary unmarked.
+ */
+export declare function writeIpaWord(syllables: readonly Syllable[], options?: IpaOptions): string;
+/**
+ * Read an IPA transcription back: `tɕiou˥˩` becomes 就 jiù.
+ *
+ * An index over the inventory, as everything here is, so a transcription of
+ * something Mandarin does not have reads as nothing rather than as a syllable
+ * nobody says. Returns an array for the same reason the other readers do,
+ * though IPA has less to be ambiguous about than any of them.
+ */
+export declare function readIpa(text: string): readonly Syllable[];
+//# sourceMappingURL=ipa.d.ts.map

package/dist/transcription/ipa.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ipa.d.ts","sourceRoot":"","sources":["../../src/transcription/ipa.ts"],"names":[],"mappings":"AAwBA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AAqIxD;;;;;;GAMG;AACH,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,KAAK,CAAC,EAAE,SAAS,GAAG,SAAS,GAAG,MAAM,CAAC;CACjD;AA0BD;;GAEG;AACH,wBAAgB,eAAe,CAAC,QAAQ,EAAE,QAAQ,GAAG,MAAM,CAI1D;AAED;;;;;GAKG;AACH,wBAAgB,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,OAAO,GAAE,UAAe,GAAG,MAAM,CAW7E;AAED;;;;;;GAMG;AACH,wBAAgB,YAAY,CAC1B,SAAS,EAAE,SAAS,QAAQ,EAAE,EAC9B,OAAO,GAAE,UAAe,GACvB,MAAM,CAER;AAkCD;;;;;;;GAOG;AACH,wBAAgB,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,QAAQ,EAAE,CAGzD"}