@kensio/pinyinjs 1.7.4 → 1.7.5

package/dist/search/candidates.js

@@ -1,129 +1,10 @@
 import { encodeReading } from "../dictionary/artifact.js";
+import { readQuery } from "./candidate-query.js";
+import { ERHUA, spells } from "./candidate-writings.js";
+import { narrowCandidates, } from "./candidate-narrowing.js";
 import { convertCharacters } from "../script/conversion.js";
 import { toCanonicalGlyphs } from "../script/glyphs.js";
-import { normaliseUmlaut, readSyllable, writeSyllable, } from "../syllable/syllable.js";
-import { applyToneMark, stripToneMarks } from "../tone/tone-mark.js";
-import { TONES } from "../tone/tone.js";
 import { readingKey } from "./reverse-index.js";
-/**
- * What a query may write between syllables, all of which mean nothing here.
- *
- * The same set the matcher accepts, minus its meaning: `match` treats a written
- * boundary as evidence, because a haystack it is filtering can be read more than
- * one way. A reverse index is asked for one reading and answers with the words
- * that have it, so `yin hang` and `yinhang` are the same question.
- */
-const SEPARATORS = /[\s'’‘-]+/gu;
-/**
- * A tone written as a digit, which is the last character of a syllable.
- */
-const TONE_DIGIT = /^[0-5]$/u;
-/**
- * Every tone a syllable that wrote none could be standing for.
- */
-const TONE_DIGITS = ["1", "2", "3", "4", "5"];
-/**
- * The umlaut the readings are written with, and the letter a keyboard has.
- */
-const UMLAUT = "ü";
-const PLAIN_U = "u";
-/**
- * The 儿化 suffix, which is the one letter a reading key keeps and a query may
- * leave off.
- */
-const ERHUA = "r";
-/**
- * Put a typed query into both the forms a search needs.
- *
- * Case, spacing and the ways of writing ü are all noise on the way to the key,
- * and the tone is noise there too — the index is keyed toneless. None of it is
- * noise on the way to the *filter*, which is what honours a tone the query
- * wrote, so the resolved spelling is kept alongside.
- *
- * A tone written as a mark counts here, unlike in `match`. The reason it cannot
- * count there is that a mark says nothing about where its syllable ends and a
- * half-typed query is exactly where that is unsettled; here the candidate's own
- * reading says where every syllable ends, so the mark has somewhere to land.
- */
-function readQuery(query) {
-    const written = normaliseUmlaut(query.normalize("NFC").toLowerCase()).replaceAll(SEPARATORS, "");
-    return { key: readingKey(stripToneMarks(written)), written };
-}
-/**
- * Every way a typist may write one syllable of a stored reading.
- *
- * `shi4` is written `shi`, `shi4` or `shì`; `lü4` adds `lu` and `lu4`, since ü
- * is not on the keyboard; and an 儿化 `wanr2` adds the whole set again with the
- * r left off, because 玩儿 is `wanr` to one typist and `wan` to another.
- *
- * The r comes off only where the syllable really is 儿化. 儿 itself is `er2`,
- * and letting that shed its r would answer `e` with 儿.
- */
-function writingsOf(token) {
-    const digit = token.slice(-1);
-    const toned = TONE_DIGIT.test(digit);
-    const spelling = toned ? token.slice(0, -1) : token;
-    const syllable = readSyllable(token);
-    // A reading with no tone written cannot contradict one, so a query is free to
-    // write any. Every reading the shipped artifacts hold has a tone digit on it,
-    // so this is a case a hand-built dictionary reaches rather than a real one.
-    const tones = toned ? [digit] : TONE_DIGITS;
-    const written = new Set();
-    const add = (form) => {
-        const plain = form.replaceAll(UMLAUT, PLAIN_U);
-        written.add(form);
-        written.add(plain);
-        for (const tone of tones) {
-            written.add(`${form}${tone}`);
-            written.add(`${plain}${tone}`);
-        }
-        if (!toned) {
-            for (const tone of TONES) {
-                written.add(applyToneMark(form, tone));
-            }
-        }
-    };
-    add(spelling);
-    if (syllable !== undefined) {
-        written.add(writeSyllable(syllable, "marks"));
-        if (syllable.erhua === true) {
-            add(spelling.slice(0, -ERHUA.length));
-            written.add(writeSyllable({ ...syllable, erhua: false }, "marks"));
-        }
-    }
-    return [...written];
-}
-/**
- * Whether a query spells a reading, syllable by syllable.
- *
- * Walked as a set of positions the query can have been read up to rather than
- * one cursor, because a syllable can be written more than one way and the ways
- * are different lengths: `wanr2` is `wan` or `wanr`, and only what follows says
- * which was meant.
- */
-function spells(query, reading, writings) {
-    let reached = new Set([0]);
-    for (const token of reading.split(" ")) {
-        let forms = writings.get(token);
-        if (forms === undefined) {
-            forms = writingsOf(token);
-            writings.set(token, forms);
-        }
-        const next = new Set();
-        for (const at of reached) {
-            for (const form of forms) {
-                if (query.startsWith(form, at)) {
-                    next.add(at + form.length);
-                }
-            }
-        }
-        if (next.size === 0) {
-            return false;
-        }
-        reached = next;
-    }
-    return reached.has(query.length);
-}
 /**
  * The reading the dictionary itself reports for a word, in numbered spelling.
  *
@@ -156,37 +37,6 @@ function rankedPositions(index, key) {
     // are short: the busiest reading in the full dictionary has 805 words in it.
     return found.toSorted((left, right) => index.dictionary.frequencyAt(right) - index.dictionary.frequencyAt(left));
 }
-/**
- * Keep one writing of a word where the group holds both scripts.
- *
- * Paired by 简体 form in both directions rather than by converting toward the
- * preference, because simplification is many-to-one and so the 简体 form is the
- * one both writings agree on. Going the other way needs a reading to tell 發
- * from 髮, and a pair that failed to meet would simply be left as two.
- *
- * The kept writing takes the rank of the better-placed of the two, so
- * collapsing a list never reorders what is left of it.
- */
-function preferScript(words, script) {
-    /** Whether a candidate is the writing the caller asked for. */
-    const isWanted = (word, simplified) => script.prefer === "Hans" ? word === simplified : word !== simplified;
-    const chosen = new Map();
-    const order = [];
-    for (const word of words) {
-        const simplified = convertCharacters(script.tables.toSimplified, word);
-        const held = chosen.get(simplified);
-        if (held === undefined) {
-            chosen.set(simplified, word);
-            order.push(simplified);
-            continue;
-        }
-        if (!isWanted(held, simplified) && isWanted(word, simplified)) {
-            chosen.set(simplified, word);
-        }
-    }
-    /* c8 ignore next -- every entry in `order` was put in `chosen` beside it */
-    return order.map((simplified) => chosen.get(simplified) ?? simplified);
-}
 /**
  * The words a pinyin query could be spelling, likeliest first.
  *
@@ -208,7 +58,8 @@ function preferScript(words, script) {
  * spaces between syllables, and the r of 儿化.
  *
  * Both scripts are dictionary keys, so 银行 and 銀行 both come back unless a
- * {@link ScriptPreference} says which to keep.
+ * {@link import("./candidate-narrowing.js").ScriptPreference} says which to
+ * keep.
  */
 export function candidates(index, query, options = {}) {
     const { key, written } = readQuery(query);
@@ -232,10 +83,7 @@ export function candidates(index, query, options = {}) {
             found.push(word);
         }
     }
-    const preferred = options.script === undefined ? found : preferScript(found, options.script);
-    return options.limit === undefined
-        ? preferred
-        : preferred.slice(0, Math.max(0, options.limit));
+    return narrowCandidates(found, options);
 }
 /**
  * The other words read exactly as this one is, likeliest first.
@@ -250,8 +98,9 @@ export function candidates(index, query, options = {}) {
  * dictionary — 4,659 candidates — that narrowing costs 1.2 ms in total.
  *
  * The word itself is never in its own list. Neither is its other-script writing
- * where a {@link ScriptPreference} is given, since 銀行 is not a homophone of
- * 银行 but the same word spelled for a different reader.
+ * where a {@link import("./candidate-narrowing.js").ScriptPreference} is given,
+ * since 銀行 is not a homophone of 银行 but the same word spelled for a
+ * different reader.
  *
  * Returns nothing for a word the dictionary does not hold, since a word with no
  * reading has no homophones rather than all of them.
@@ -282,9 +131,6 @@ export function homophonesOf(index, word, options = {}) {
             found.push(candidate);
         }
     }
-    const preferred = options.script === undefined ? found : preferScript(found, options.script);
-    return options.limit === undefined
-        ? preferred
-        : preferred.slice(0, Math.max(0, options.limit));
+    return narrowCandidates(found, options);
 }
 //# sourceMappingURL=candidates.js.map

package/dist/search/candidates.js.map

@@ -1 +1 @@
-{"version":3,"file":"candidates.js","sourceRoot":"","sources":["../../src/search/candidates.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,2BAA2B,CAAC;AAC1D,OAAO,EAAE,iBAAiB,EAAqB,MAAM,yBAAyB,CAAC;AAC/E,OAAO,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AAExD,OAAO,EACL,eAAe,EACf,YAAY,EACZ,aAAa,GACd,MAAM,yBAAyB,CAAC;AACjC,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACrE,OAAO,EAAE,KAAK,EAAE,MAAM,iBAAiB,CAAC;AACxC,OAAO,EAAE,UAAU,EAAqB,MAAM,oBAAoB,CAAC;AAEnE;;;;;;;GAOG;AACH,MAAM,UAAU,GAAG,aAAa,CAAC;AAEjC;;GAEG;AACH,MAAM,UAAU,GAAG,UAAU,CAAC;AAE9B;;GAEG;AACH,MAAM,WAAW,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC;AAE9C;;GAEG;AACH,MAAM,MAAM,GAAG,GAAG,CAAC;AAEnB,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;GAGG;AACH,MAAM,KAAK,GAAG,GAAG,CAAC;AAmDlB;;;;;;;;;;;;GAYG;AACH,SAAS,SAAS,CAAC,KAAa;IAC9B,MAAM,OAAO,GAAG,eAAe,CAC7B,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CACrC,CAAC,UAAU,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IAC7B,OAAO,EAAE,GAAG,EAAE,UAAU,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC;AAC/D,CAAC;AAED;;;;;;;;;GASG;AACH,SAAS,UAAU,CAAC,KAAa;IAC/B,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACrC,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;IACpD,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IACrC,8EAA8E;IAC9E,8EAA8E;IAC9E,4EAA4E;IAC5E,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC;IAE5C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,MAAM,GAAG,GAAG,CAAC,IAAY,EAAQ,EAAE;QACjC,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;QAC/C,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAClB,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACnB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,OAAO,CAAC,GAAG,CAAC,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC,CAAC;YAC9B,OAAO,CAAC,GAAG,CAAC,GAAG,KAAK,GAAG,IAAI,EAAE,CAAC,CAAC;QACjC,CAAC;QACD,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;IACH,CAAC,CAAC;IACF,GAAG,CAAC,QAAQ,CAAC,CAAC;IACd,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;QAC3B,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC;QAC9C,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,EAAE,CAAC;YAC5B,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;YACtC,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,EAAE,GAAG,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,OAAO,CAAC,CAAC,CAAC;QACrE,CAAC;IACH,CAAC;IACD,OAAO,CAAC,GAAG,OAAO,CAAC,CAAC;AACtB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,MAAM,CACb,KAAa,EACb,OAAe,EACf,QAAwC;IAExC,IAAI,OAAO,GAAG,IAAI,GAAG,CAAS,CAAC,CAAC,CAAC,CAAC,CAAC;IACnC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QACvC,IAAI,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAChC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;YAC1B,QAAQ,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;QAC7B,CAAC;QACD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,KAAK,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;YACzB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,IAAI,KAAK,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,EAAE,CAAC;oBAC/B,IAAI,CAAC,GAAG,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC;gBAC7B,CAAC;YACH,CAAC;QACH,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACpB,OAAO,KAAK,CAAC;QACf,CAAC;QACD,OAAO,GAAG,IAAI,CAAC;IACjB,CAAC;IACD,OAAO,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;AACnC,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,SAAS,CAAC,KAAmB,EAAE,IAAY;IAClD,MAAM,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC5C,4EAA4E;IAC5E,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AACjE,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,eAAe,CAAC,KAAmB,EAAE,GAAW;IACvD,MAAM,KAAK,GAAG;QACZ,GAAG,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC;QAC1B,GAAG,KAAK,CAAC,YAAY,CAAC,GAAG,GAAG,GAAG,KAAK,EAAE,CAAC;KACxC,CAAC;IACF,8EAA8E;IAC9E,4EAA4E;IAC5E,6EAA6E;IAC7E,OAAO,KAAK,CAAC,QAAQ,CACnB,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,KAAK,CAAC,UAAU,CAAC,WAAW,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC,WAAW,CAAC,IAAI,CAAC,CAC3E,CAAC;AACJ,CAAC;AAED;;;;;;;;;;GAUG;AACH,SAAS,YAAY,CACnB,KAAwB,EACxB,MAAwB;IAExB,+DAA+D;IAC/D,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,UAAkB,EAAW,EAAE,CAC7D,MAAM,CAAC,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,IAAI,KAAK,UAAU,CAAC,CAAC,CAAC,IAAI,KAAK,UAAU,CAAC;IAEvE,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,UAAU,GAAG,iBAAiB,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC;QACvE,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;QACpC,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACvB,MAAM,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YACvB,SAAS;QACX,CAAC;QACD,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,UAAU,CAAC,IAAI,QAAQ,CAAC,IAAI,EAAE,UAAU,CAAC,EAAE,CAAC;YAC9D,MAAM,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IACD,4EAA4E;IAC5E,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,UAAU,CAAC,CAAC;AACzE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;GAsBG;AACH,MAAM,UAAU,UAAU,CACxB,KAAmB,EACnB,KAAa,EACb,OAAO,GAAqB,EAAE;IAE9B,MAAM,EAAE,GAAG,EAAE,OAAO,EAAE,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAC1C,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;QACf,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA6B,CAAC;IACtD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,eAAe,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE,CAAC;QACnD,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC/C,yEAAyE;QACzE,2EAA2E;QAC3E,sEAAsE;QACtE,wEAAwE;QACxE,0EAA0E;QAC1E,6EAA6E;QAC7E,IAAI,iBAAiB,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;YACrC,SAAS;QACX,CAAC;QACD,IAAI,MAAM,CAAC,OAAO,EAAE,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,QAAQ,CAAC,EAAE,CAAC;YACtD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACnB,CAAC;IACH,CAAC;IAED,MAAM,SAAS,GACb,OAAO,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;IAC7E,OAAO,OAAO,CAAC,KAAK,KAAK,SAAS;QAChC,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;AACrD,CAAC;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAmB,EACnB,IAAY,EACZ,OAAO,GAAqB,EAAE;IAE9B,MAAM,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC5C,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC;IACZ,CAAC;IACD,MAAM,OAAO,GAAG,aAAa,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC7C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IACzC,IAAI,OAAO,CAAC,MAAM,KAAK,SAAS,EAAE,CAAC;QACjC,IAAI,CAAC,GAAG,CAAC,iBAAiB,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC,CAAC;IACxE,CAAC;IAED,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,eAAe,CAAC,KAAK,EAAE,UAAU,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QACnE,MAAM,SAAS,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACpD,IAAI,iBAAiB,CAAC,SAAS,CAAC,KAAK,SAAS,EAAE,CAAC;YAC/C,SAAS;QACX,CAAC;QACD,MAAM,OAAO,GACX,OAAO,CAAC,MAAM,KAAK,SAAS;YAC1B,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,iBAAiB,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,EAAE,SAAS,CAAC,CAAC;QACvE,IAAI,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAC7C,SAAS;QACX,CAAC;QACD,IAAI,SAAS,CAAC,KAAK,EAAE,SAAS,CAAC,KAAK,OAAO,EAAE,CAAC;YAC5C,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACxB,CAAC;IACH,CAAC;IAED,MAAM,SAAS,GACb,OAAO,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;IAC7E,OAAO,OAAO,CAAC,KAAK,KAAK,SAAS;QAChC,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;AACrD,CAAC"}
+{"version":3,"file":"candidates.js","sourceRoot":"","sources":["../../src/search/candidates.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,2BAA2B,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,sBAAsB,CAAC;AACjD,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,yBAAyB,CAAC;AACxD,OAAO,EACL,gBAAgB,GAEjB,MAAM,0BAA0B,CAAC;AAClC,OAAO,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAC5D,OAAO,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AACxD,OAAO,EAAE,UAAU,EAAqB,MAAM,oBAAoB,CAAC;AAOnE;;;;;;;;GAQG;AACH,SAAS,SAAS,CAAC,KAAmB,EAAE,IAAY;IAClD,MAAM,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC5C,4EAA4E;IAC5E,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AACjE,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,eAAe,CAAC,KAAmB,EAAE,GAAW;IACvD,MAAM,KAAK,GAAG;QACZ,GAAG,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC;QAC1B,GAAG,KAAK,CAAC,YAAY,CAAC,GAAG,GAAG,GAAG,KAAK,EAAE,CAAC;KACxC,CAAC;IACF,8EAA8E;IAC9E,4EAA4E;IAC5E,6EAA6E;IAC7E,OAAO,KAAK,CAAC,QAAQ,CACnB,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,KAAK,CAAC,UAAU,CAAC,WAAW,CAAC,KAAK,CAAC,GAAG,KAAK,CAAC,UAAU,CAAC,WAAW,CAAC,IAAI,CAAC,CAC3E,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,MAAM,UAAU,UAAU,CACxB,KAAmB,EACnB,KAAa,EACb,OAAO,GAAqB,EAAE;IAE9B,MAAM,EAAE,GAAG,EAAE,OAAO,EAAE,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAC1C,IAAI,GAAG,KAAK,EAAE,EAAE,CAAC;QACf,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA6B,CAAC;IACtD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,eAAe,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE,CAAC;QACnD,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC/C,yEAAyE;QACzE,2EAA2E;QAC3E,sEAAsE;QACtE,wEAAwE;QACxE,0EAA0E;QAC1E,6EAA6E;QAC7E,IAAI,iBAAiB,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC;YACrC,SAAS;QACX,CAAC;QACD,IAAI,MAAM,CAAC,OAAO,EAAE,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,QAAQ,CAAC,EAAE,CAAC;YACtD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACnB,CAAC;IACH,CAAC;IAED,OAAO,gBAAgB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;AAC1C,CAAC;AAED;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,UAAU,YAAY,CAC1B,KAAmB,EACnB,IAAY,EACZ,OAAO,GAAqB,EAAE;IAE9B,MAAM,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC5C,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,OAAO,EAAE,CAAC;IACZ,CAAC;IACD,MAAM,OAAO,GAAG,aAAa,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC7C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IACzC,IAAI,OAAO,CAAC,MAAM,KAAK,SAAS,EAAE,CAAC;QACjC,IAAI,CAAC,GAAG,CAAC,iBAAiB,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC,CAAC;IACxE,CAAC;IAED,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,eAAe,CAAC,KAAK,EAAE,UAAU,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QACnE,MAAM,SAAS,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACpD,IAAI,iBAAiB,CAAC,SAAS,CAAC,KAAK,SAAS,EAAE,CAAC;YAC/C,SAAS;QACX,CAAC;QACD,MAAM,OAAO,GACX,OAAO,CAAC,MAAM,KAAK,SAAS;YAC1B,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,iBAAiB,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,EAAE,SAAS,CAAC,CAAC;QACvE,IAAI,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAC7C,SAAS;QACX,CAAC;QACD,IAAI,SAAS,CAAC,KAAK,EAAE,SAAS,CAAC,KAAK,OAAO,EAAE,CAAC;YAC5C,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACxB,CAAC;IACH,CAAC;IAED,OAAO,gBAAgB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;AAC1C,CAAC"}

package/dist/search/match-chunks.d.ts

@@ -0,0 +1,19 @@
+/**
+ * How much of a query one whole reading can account for.
+ *
+ * The smallest step the search takes, and the only part of it that looks at the
+ * query's letters rather than at the haystack's characters. Kept apart from the
+ * walk because it is a pure question about two strings — no dictionary, no
+ * memo, no score — which is what makes the walk above it readable as a walk.
+ */
+import type { Syllable } from "../syllable/syllable.js";
+import { type QueryChunk } from "./query.js";
+/**
+ * Every position the query can be read up to by one whole reading.
+ *
+ * A reading is usually one syllable, and is not always: 瓩 is `qiānwǎ`, one
+ * character read as two. The query accounts for them in order, and may run
+ * out partway through, which is a query still being typed.
+ */
+export declare function chunksFor(query: string, from: number, reading: readonly Syllable[]): readonly QueryChunk[];
+//# sourceMappingURL=match-chunks.d.ts.map

package/dist/search/match-chunks.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-chunks.d.ts","sourceRoot":"","sources":["../../src/search/match-chunks.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AACH,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AACxD,OAAO,EAAE,KAAK,UAAU,EAAmC,MAAM,YAAY,CAAC;AAE9E;;;;;;GAMG;AACH,wBAAgB,SAAS,CACvB,KAAK,EAAE,MAAM,EACb,IAAI,EAAE,MAAM,EACZ,OAAO,EAAE,SAAS,QAAQ,EAAE,GAC3B,SAAS,UAAU,EAAE,CAoBvB"}

package/dist/search/match-chunks.js

@@ -0,0 +1,30 @@
+import { readQueryChunks, skipSeparators } from "./query.js";
+/**
+ * Every position the query can be read up to by one whole reading.
+ *
+ * A reading is usually one syllable, and is not always: 瓩 is `qiānwǎ`, one
+ * character read as two. The query accounts for them in order, and may run
+ * out partway through, which is a query still being typed.
+ */
+export function chunksFor(query, from, reading) {
+    let reached = [{ next: from, isFull: true }];
+    for (const syllable of reading) {
+        const found = [];
+        for (const chunk of reached) {
+            const start = skipSeparators(query, chunk.next);
+            if (start === query.length) {
+                found.push({ next: start, isFull: chunk.isFull });
+                continue;
+            }
+            for (const one of readQueryChunks(query, start, syllable)) {
+                found.push({
+                    next: one.next,
+                    isFull: chunk.isFull && one.isFull,
+                });
+            }
+        }
+        reached = found;
+    }
+    return reached;
+}
+//# sourceMappingURL=match-chunks.js.map

package/dist/search/match-chunks.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-chunks.js","sourceRoot":"","sources":["../../src/search/match-chunks.ts"],"names":[],"mappings":"AASA,OAAO,EAAmB,eAAe,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAE9E;;;;;;GAMG;AACH,MAAM,UAAU,SAAS,CACvB,KAAa,EACb,IAAY,EACZ,OAA4B;IAE5B,IAAI,OAAO,GAA0B,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;IACpE,KAAK,MAAM,QAAQ,IAAI,OAAO,EAAE,CAAC;QAC/B,MAAM,KAAK,GAAiB,EAAE,CAAC;QAC/B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,MAAM,KAAK,GAAG,cAAc,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;YAChD,IAAI,KAAK,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC;gBAC3B,KAAK,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC;gBAClD,SAAS;YACX,CAAC;YACD,KAAK,MAAM,GAAG,IAAI,eAAe,CAAC,KAAK,EAAE,KAAK,EAAE,QAAQ,CAAC,EAAE,CAAC;gBAC1D,KAAK,CAAC,IAAI,CAAC;oBACT,IAAI,EAAE,GAAG,CAAC,IAAI;oBACd,MAAM,EAAE,KAAK,CAAC,MAAM,IAAI,GAAG,CAAC,MAAM;iBACnC,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QACD,OAAO,GAAG,KAAK,CAAC;IAClB,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC"}

package/dist/search/match-ranking.d.ts

@@ -0,0 +1,39 @@
+/**
+ * Choosing between the places a query matched.
+ *
+ * The walk in `match-search.ts` answers *where* and *how* the query reads; this
+ * answers *which of those is the match*. The two are separate because they weigh
+ * different things: the walk scores a reading against the query as it goes, and
+ * this scores a finished path against the haystack it lies in — how the text is
+ * actually read, whether the match starts a word, how early it falls.
+ */
+import type { Dictionary } from "../dictionary/dictionary.js";
+import { type MatchRange, type Path } from "./match-scoring.js";
+/**
+ * What a query matched, and how well.
+ */
+export interface PinyinMatch {
+    /**
+     * The stretches of the haystack the query spelled, in order.
+     *
+     * Usually one. There is more than one where the query ran over something
+     * with no reading of its own: 北京·大学 is matched by `bjdx` as two ranges
+     * with the separator between them left out, so that highlighting them marks
+     * what was matched rather than what lay between.
+     */
+    readonly ranges: readonly MatchRange[];
+    /**
+     * How good the match is, where higher is better.
+     *
+     * For ordering the haystacks one query matched — sort descending and the
+     * best is first. Comparable within a query and not across queries: what it
+     * weighs is described at {@link CONTEXT_READING}, and none of it is a
+     * probability.
+     */
+    readonly score: number;
+}
+/**
+ * The best of the paths a query read, or undefined where it read none.
+ */
+export declare function bestMatch(dictionary: Dictionary, haystack: string, query: string, found: readonly Path[]): PinyinMatch | undefined;
+//# sourceMappingURL=match-ranking.d.ts.map

package/dist/search/match-ranking.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-ranking.d.ts","sourceRoot":"","sources":["../../src/search/match-ranking.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAG9D,OAAO,EAIL,KAAK,UAAU,EACf,KAAK,IAAI,EAIV,MAAM,oBAAoB,CAAC;AAE5B;;GAEG;AACH,MAAM,WAAW,WAAW;IAC1B;;;;;;;OAOG;IACH,QAAQ,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,CAAC;IACvC;;;;;;;OAOG;IACH,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;CACxB;AAgCD;;GAEG;AACH,wBAAgB,SAAS,CACvB,UAAU,EAAE,UAAU,EACtB,QAAQ,EAAE,MAAM,EAChB,KAAK,EAAE,MAAM,EACb,KAAK,EAAE,SAAS,IAAI,EAAE,GACrB,WAAW,GAAG,SAAS,CAoBzB"}

package/dist/search/match-ranking.js

@@ -0,0 +1,51 @@
+import { readQueryChunks } from "./query.js";
+import { CONTEXT_READING, contextOf, EARLINESS, rangesOf, WORD_START, } from "./match-scoring.js";
+/**
+ * How much of a match was written the way the text is actually read.
+ *
+ * Asked of what the query wrote rather than of the reading the search happened
+ * to take it by: 西 is stored with a neutral-toned `xi` and a first-tone `xī`,
+ * and a query writing `xi` has picked neither of them over the other. What it
+ * has done is write something the settled reading can account for, which is the
+ * question — and the one `yx` over 银行 answers differently from `yh`.
+ */
+function agreementOf(query, steps, preferred) {
+    let comparable = 0;
+    let agreeing = 0;
+    for (const step of steps) {
+        const settled = preferred[step.at];
+        if (!step.isRead || settled === undefined) {
+            continue;
+        }
+        comparable++;
+        const written = readQueryChunks(query, step.from, settled);
+        if (written.some((chunk) => chunk.next === step.next)) {
+            agreeing++;
+        }
+    }
+    return comparable === 0 ? 0 : agreeing / comparable;
+}
+/**
+ * The best of the paths a query read, or undefined where it read none.
+ */
+export function bestMatch(dictionary, haystack, query, found) {
+    const context = contextOf(dictionary, haystack);
+    let best;
+    for (const path of found) {
+        const ranges = rangesOf(path.steps);
+        /* c8 ignore next 3 -- a match always starts on a character it read */
+        if (ranges[0] === undefined) {
+            continue;
+        }
+        const start = ranges[0].at;
+        const score = CONTEXT_READING * agreementOf(query, path.steps, context.preferred) +
+            (context.wordStarts.has(start) ? WORD_START : 0) +
+            EARLINESS / (1 + start);
+        // Strictly better, so that two matches worth the same keep the earlier.
+        if (best === undefined || score > best.score) {
+            best = { ranges, score };
+        }
+    }
+    return best;
+}
+//# sourceMappingURL=match-ranking.js.map

package/dist/search/match-ranking.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-ranking.js","sourceRoot":"","sources":["../../src/search/match-ranking.ts"],"names":[],"mappings":"AAWA,OAAO,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAC7C,OAAO,EACL,eAAe,EACf,SAAS,EACT,SAAS,EAGT,QAAQ,EAER,UAAU,GACX,MAAM,oBAAoB,CAAC;AA0B5B;;;;;;;;GAQG;AACH,SAAS,WAAW,CAClB,KAAa,EACb,KAAsB,EACtB,SAA4C;IAE5C,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1C,SAAS;QACX,CAAC;QACD,UAAU,EAAE,CAAC;QACb,MAAM,OAAO,GAAG,eAAe,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;QAC3D,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACtD,QAAQ,EAAE,CAAC;QACb,CAAC;IACH,CAAC;IACD,OAAO,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,GAAG,UAAU,CAAC;AACtD,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,SAAS,CACvB,UAAsB,EACtB,QAAgB,EAChB,KAAa,EACb,KAAsB;IAEtB,MAAM,OAAO,GAAG,SAAS,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAChD,IAAI,IAA6B,CAAC;IAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpC,sEAAsE;QACtE,IAAI,MAAM,CAAC,CAAC,CAAC,KAAK,SAAS,EAAE,CAAC;YAC5B,SAAS;QACX,CAAC;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3B,MAAM,KAAK,GACT,eAAe,GAAG,WAAW,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,OAAO,CAAC,SAAS,CAAC;YACnE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;YAChD,SAAS,GAAG,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC;QAC1B,wEAAwE;QACxE,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC;YAC7C,IAAI,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;QAC3B,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/search/match-readings.d.ts

@@ -0,0 +1,43 @@
+/**
+ * What a character in the haystack can be read as, asked once per position.
+ *
+ * Kept apart from the search itself because it is a different question: the
+ * search asks how a query walks over a text, and this asks only what each
+ * character offers it to walk on.
+ */
+import type { Dictionary } from "../dictionary/dictionary.js";
+import type { Syllable } from "../syllable/syllable.js";
+import type { Candidate } from "./match-scoring.js";
+export declare class CharacterReadings {
+    #private;
+    constructor(dictionary: Dictionary, characters: readonly string[]);
+    /**
+     * Every reading a character offers, likeliest first, asked for once.
+     *
+     * Canonicalised here rather than left to the dictionary, so that a Hong Kong
+     * glyph form is matched by the readings of the Taiwan-standard key it folds
+     * to rather than by nothing.
+     *
+     * The 國語 reading goes on the end where it differs, so that 圾 answers to
+     * `se` as well as to `ji`. A search box is not the place to insist on one
+     * standard: somebody typing what they say is not typing it wrongly, and the
+     * decoder's own reading is what ranks the two afterwards.
+     */
+    at(at: number): readonly (readonly Syllable[])[];
+    /**
+     * Every reading the query could be accounting for at a character.
+     *
+     * One per reading the character has — and one more for each of them where an
+     * 儿 follows, carrying the r suffix that character stands for. 玩儿 is
+     * `wánr`, one syllable over two characters, and somebody looking for it types
+     * `wanr`: the r is part of how the word is said, so it is part of how the
+     * word is searched for.
+     *
+     * Offered wherever an 儿 follows rather than only where the dictionary
+     * attests the 儿化, because the query is what decides which of them was
+     * meant. 女儿 is `nǚ'ér` and not `nǚr`, and typing `nvr` still finds it —
+     * below `nver`, since the reading the text takes is what ranks them.
+     */
+    candidatesAt(at: number): readonly Candidate[];
+}
+//# sourceMappingURL=match-readings.d.ts.map

package/dist/search/match-readings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-readings.d.ts","sourceRoot":"","sources":["../../src/search/match-readings.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AACH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAI9D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AACxD,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAEpD,qBAAa,iBAAiB;;IAK5B,YAAY,UAAU,EAAE,UAAU,EAAE,UAAU,EAAE,SAAS,MAAM,EAAE,EAGhE;IAED;;;;;;;;;;;OAWG;IACH,EAAE,CAAC,EAAE,EAAE,MAAM,GAAG,SAAS,CAAC,SAAS,QAAQ,EAAE,CAAC,EAAE,CAgB/C;IAED;;;;;;;;;;;;;OAaG;IACH,YAAY,CAAC,EAAE,EAAE,MAAM,GAAG,SAAS,SAAS,EAAE,CA2B7C;CACF"}

package/dist/search/match-readings.js

@@ -0,0 +1,83 @@
+import { isSameReading } from "../dictionary/entry.js";
+import { isErCharacter, withErhua } from "../dictionary/erhua.js";
+import { toCanonicalGlyphs } from "../script/glyphs.js";
+export class CharacterReadings {
+    #dictionary;
+    #characters;
+    #readings = [];
+    constructor(dictionary, characters) {
+        this.#dictionary = dictionary;
+        this.#characters = characters;
+    }
+    /**
+     * Every reading a character offers, likeliest first, asked for once.
+     *
+     * Canonicalised here rather than left to the dictionary, so that a Hong Kong
+     * glyph form is matched by the readings of the Taiwan-standard key it folds
+     * to rather than by nothing.
+     *
+     * The 國語 reading goes on the end where it differs, so that 圾 answers to
+     * `se` as well as to `ji`. A search box is not the place to insist on one
+     * standard: somebody typing what they say is not typing it wrongly, and the
+     * decoder's own reading is what ranks the two afterwards.
+     */
+    at(at) {
+        const held = this.#readings[at];
+        if (held !== undefined) {
+            return held;
+        }
+        /* c8 ignore next -- only ever asked about a character that is there */
+        const character = toCanonicalGlyphs(this.#characters[at] ?? "");
+        const found = this.#dictionary.readingsOf(character);
+        const taiwan = this.#dictionary.lookup(character)?.taiwanReading;
+        const readings = taiwan === undefined ||
+            found.some((reading) => isSameReading(reading, taiwan))
+            ? found
+            : [...found, taiwan];
+        this.#readings[at] = readings;
+        return readings;
+    }
+    /**
+     * Every reading the query could be accounting for at a character.
+     *
+     * One per reading the character has — and one more for each of them where an
+     * 儿 follows, carrying the r suffix that character stands for. 玩儿 is
+     * `wánr`, one syllable over two characters, and somebody looking for it types
+     * `wanr`: the r is part of how the word is said, so it is part of how the
+     * word is searched for.
+     *
+     * Offered wherever an 儿 follows rather than only where the dictionary
+     * attests the 儿化, because the query is what decides which of them was
+     * meant. 女儿 is `nǚ'ér` and not `nǚr`, and typing `nvr` still finds it —
+     * below `nver`, since the reading the text takes is what ranks them.
+     */
+    candidatesAt(at) {
+        const readings = this.at(at);
+        const plain = readings.map((reading, rank) => ({
+            reading,
+            rank,
+            characters: 1,
+        }));
+        if (!isErCharacter(this.#characters[at + 1] ?? "")) {
+            return plain;
+        }
+        return [
+            ...plain,
+            ...readings.flatMap((reading, rank) => {
+                const last = reading.at(-1);
+                /* c8 ignore next 3 -- a stored reading is never empty or already r-ed */
+                if (last === undefined || last.erhua === true) {
+                    return [];
+                }
+                return [
+                    {
+                        reading: [...reading.slice(0, -1), withErhua(last)],
+                        rank,
+                        characters: 2,
+                    },
+                ];
+            }),
+        ];
+    }
+}
+//# sourceMappingURL=match-readings.js.map

package/dist/search/match-readings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"match-readings.js","sourceRoot":"","sources":["../../src/search/match-readings.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,aAAa,EAAE,MAAM,wBAAwB,CAAC;AACvD,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,wBAAwB,CAAC;AAClE,OAAO,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AAIxD,MAAM,OAAO,iBAAiB;IACnB,WAAW,CAAa;IACxB,WAAW,CAAoB;IAC/B,SAAS,GAAyC,EAAE,CAAC;IAE9D,YAAY,UAAsB,EAAE,UAA6B;QAC/D,IAAI,CAAC,WAAW,GAAG,UAAU,CAAC;QAC9B,IAAI,CAAC,WAAW,GAAG,UAAU,CAAC;IAChC,CAAC;IAED;;;;;;;;;;;OAWG;IACH,EAAE,CAAC,EAAU;QACX,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;QAChC,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;YACvB,OAAO,IAAI,CAAC;QACd,CAAC;QACD,uEAAuE;QACvE,MAAM,SAAS,GAAG,iBAAiB,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC;QAChE,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC;QACrD,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,SAAS,CAAC,EAAE,aAAa,CAAC;QACjE,MAAM,QAAQ,GACZ,MAAM,KAAK,SAAS;YACpB,KAAK,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,aAAa,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YACrD,CAAC,CAAC,KAAK;YACP,CAAC,CAAC,CAAC,GAAG,KAAK,EAAE,MAAM,CAAC,CAAC;QACzB,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC;QAC9B,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED;;;;;;;;;;;;;OAaG;IACH,YAAY,CAAC,EAAU;QACrB,MAAM,QAAQ,GAAG,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;QAC7B,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC;YAC7C,OAAO;YACP,IAAI;YACJ,UAAU,EAAE,CAAC;SACd,CAAC,CAAC,CAAC;QACJ,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;YACnD,OAAO,KAAK,CAAC;QACf,CAAC;QACD,OAAO;YACL,GAAG,KAAK;YACR,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE;gBACpC,MAAM,IAAI,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC5B,yEAAyE;gBACzE,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,KAAK,IAAI,EAAE,CAAC;oBAC9C,OAAO,EAAE,CAAC;gBACZ,CAAC;gBACD,OAAO;oBACL;wBACE,OAAO,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,SAAS,CAAC,IAAI,CAAC,CAAC;wBACnD,IAAI;wBACJ,UAAU,EAAE,CAAC;qBACd;iBACF,CAAC;YACJ,CAAC,CAAC;SACH,CAAC;IACJ,CAAC;CACF"}