@kensio/pinyinjs 1.5.3 → 1.7.0

package/dist/decode/reading-rules.js

@@ -272,6 +272,162 @@ export const COUNTED_MEASURE = {
             : "keep";
     },
 };
+/**
+ * Whether a syllable is `cháng`, the stative reading of 长.
+ */
+function isChang(syllable) {
+    return (syllable?.initial === "ch" &&
+        syllable.final === "ang" &&
+        syllable.tone === 2);
+}
+/**
+ * Adverbs that measure a quality rather than modify an action.
+ *
+ * The discriminating side of the 长 context, and the reason the set is written
+ * out rather than taken from a tag: jieba calls 很 `zg` and 多 `m`, so the two
+ * commonest members are not adverbs as far as the dictionary is concerned,
+ * while 已经, 也, 却 and 就 — which sit in front of the *verb* 长 — are `d`
+ * along with the ones wanted here. The tag names the wrong set in both
+ * directions, so the set is named directly.
+ *
+ * Everything here is degree or comparison. 还 and 不 are deliberately absent
+ * even though 距离还长 and 那座桥不长 are both `cháng`: they modify verbs as
+ * readily as qualities, and 还长有毛 and 胡子不长在前额上 are the same two
+ * characters around the verb.
+ */
+const DEGREE = new Set([
+    "一样",
+    "十分",
+    "多",
+    "多么",
+    "好",
+    "如此",
+    "尽量",
+    "很",
+    "格外",
+    "极",
+    "极其",
+    "比较",
+    "略",
+    "相当",
+    "真",
+    "稍",
+    "稍微",
+    "越",
+    "越来越",
+    "这么",
+    "这样",
+    "那么",
+    "那样",
+    "更",
+    "最",
+    "有多",
+    "特别",
+    "老",
+    "够",
+    "挺",
+    "太",
+    "非常",
+]);
+/**
+ * What shows a 长 to be growing even with a degree word in front of it.
+ *
+ * {@link DEGREE} already excludes nearly everything that precedes the verb, so
+ * this only has to settle the three shapes reachable from both sets: 真长得很快
+ * and 越长得快, where 得 marks the complement; 长着, where the aspect does; and
+ * 越长越高, where the correlative makes the 长 the thing that increases.
+ *
+ * 了 and 的 are pointedly *not* here, and that is the whole use of conditioning
+ * on the adverb first. Following a bare 长 either would suggest the verb — 长了
+ * 一个大包, 长的很像 — but following an adverb and a 长 they are the sentence
+ * particle and the attributive: 时间太长了 and 很长的道路 are both `cháng`, and
+ * an earlier draft of this rule that guarded on them read both as `zhǎng`.
+ */
+const GROWING = new Set(["得", "着", "越"]);
+/**
+ * What growing produces, as the far half of 越长越X.
+ *
+ * 越长越高, 越长越大, 越长越胖: each names a dimension that increases *because*
+ * something grew, so the 长 in front of it is the verb. The adjectives that go
+ * with a length or a duration are pointedly absent — 时间越长越好, 队伍越长越慢
+ * — because those are the cases the sources already read correctly.
+ */
+const GROWN_INTO = new Set(["高", "大", "胖", "壮", "结实"]);
+/**
+ * Whether an edge is the 越长 of a 越长越X saying something grew.
+ *
+ * This exists because of a defect in the sources rather than a gap in them.
+ * 越长 is a key, read `yuè cháng`, and it is the *only* one of its shape: 越大,
+ * 越高, 越好, 越快 and 越多 are all absent, so 越高越好 decodes as two words and
+ * 越长越高 reaches for a word nothing else in the paradigm has. It carries no
+ * part of speech, which is how the dictionary holds a reading somebody asserted
+ * rather than a word anybody counted — the same category {@link TEACHING_JIAO}
+ * declines to trust — and it comes from one source, `large_pinyin.txt`, with
+ * CC-CEDICT holding no such entry.
+ *
+ * Taking the edge away leaves the character's own default, which is `zhǎng`,
+ * and leaves `cháng` standing as a rival about one bucket dearer. That last
+ * part is the point of forbidding rather than forcing: 越长越X is genuinely
+ * ambiguous — 孩子越长越漂亮 grows and 头发越长越漂亮 lengthens, and only the
+ * subject says which — so the decode should answer with the likelier reading
+ * and still report itself as guessing. A forced edge could not.
+ *
+ * Unlike the rest of this file the shape has no corpus behind it: 越长 occurs
+ * three times in the 88,866 lines, and all three are 越来越长 or 说的越长. The
+ * word list is therefore small on purpose, and everything outside it is left as
+ * the sources have it.
+ */
+function isGrowthCorrelative(context) {
+    const { characters, edge } = context;
+    return (edge.text === "越长" &&
+        edge.partOfSpeech === "" &&
+        characters[edge.to] === "越" &&
+        GROWN_INTO.has(characters[edge.to + 1] ?? ""));
+}
+/**
+ * 长 read as `cháng` where an adverb of degree measures it.
+ *
+ * The character is stored `zhǎng` with `cháng` as an alternate, and the data
+ * says so about the character in isolation: Unihan ranks the readings
+ * `zhǎng(1879) cháng(1179)` by corpus count and names `zhǎng` in `kMandarin`.
+ * That ranking is doing real work — 署长, 团长, 公安局长, 总会长, 审理长 and
+ * the rest reach a bare 长 at the end of a title and read it correctly — so the
+ * default is not the thing to change. What is left over is the 长 that is an
+ * adjective, and nothing in the cost model can prefer an alternate at a
+ * position no word covers: 这篇文章不太长 came out `bú tài zhǎng`.
+ *
+ * The context that separates them is one-sided, as it was for the modal 得. A
+ * 长 that is *growing* is a verb, and a degree adverb cannot modify one — there
+ * is no 很长 meaning it grew a lot — so 很, 太, 最, 多 and their kin in front of
+ * a bare 长 make it the adjective. The other side carries almost nothing: what
+ * follows an adjectival 长 is a noun (很长时间), a particle (最长的大桥), or the
+ * end of the sentence (他的腿很长), which is also what follows half the verbs.
+ *
+ * Measured over the same 88,866 lines of Tatoeba and zh.wikipedia the other
+ * rules were sized against, 260 长 decode as a word of their own and this moves
+ * 75 of them to `cháng`, all 75 correctly. On CPP's 40 hand-labelled 长 it takes
+ * the character from 85.00% to 87.50%, fixing 时间最长 and breaking nothing.
+ *
+ * The other half of the rule is {@link GROWN_INTO}, which pushes the opposite
+ * way on the one shape the sources get wrong.
+ */
+export const ADJECTIVAL_CHANG = {
+    name: "adjectival-chang",
+    verdictFor: (context) => {
+        const { characters, edge } = context;
+        if (isGrowthCorrelative(context)) {
+            return "forbid";
+        }
+        if (edge.text !== "长" || !isChang(edge.reading[0])) {
+            return "keep";
+        }
+        if (GROWING.has(characters[edge.to] ?? "")) {
+            return "keep";
+        }
+        const before = wordEndingAt(context, edge.from);
+        return before !== undefined && DEGREE.has(before) ? "force" : "keep";
+    },
+};
 /**
  * The rules the lattice decode applies, in order.
  */
@@ -280,5 +436,6 @@ export const READING_RULES = [
     TEACHING_JIAO,
     ATTESTED_ERHUA,
     COUNTED_MEASURE,
+    ADJECTIVAL_CHANG,
 ];
 //# sourceMappingURL=reading-rules.js.map

package/dist/decode/reading-rules.js.map

@@ -1 +1 @@
-{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAEvD,OAAO,EAGL,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,SAAS,MAAM,CAAC,QAA8B;IAC5C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,KAAK,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC7E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEhE;;;;;;;GAOG;AACH,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEvD;;GAEG;AACH,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;;;;;GAOG;AACH,SAAS,YAAY,CAAC,OAAoB,EAAE,EAAU;IACpD,IAAI,KAAK,CAAC,OAAO,EAAE,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QACvE,OAAO,KAAK,CAAC;IACf,CAAC;IACD,OAAO,CACL,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5C,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,cAAc,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CACjE,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,aAAa,GAAa;IACrC,IAAI,EAAE,eAAe;IACrB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC;YACtB,OAAO,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC;gBAChE,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,MAAM,CAAC;QACb,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,KAAK,EAAE;YAC7B,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI;YAC3C,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;YAC5B,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC;YAClC,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,aAAa;IACb,cAAc;IACd,eAAe;CAChB,CAAC"}
+{"version":3,"file":"reading-rules.js","sourceRoot":"","sources":["../../src/decode/reading-rules.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAEvD,OAAO,EAGL,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,YAAY,CAAC;AAEpB;;;;;GAKG;AACH,SAAS,KAAK,CAAC,QAA8B;IAC3C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,IAAI,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC5E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE9C;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,eAAe,CAAC;AAEtC;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa;IAChC,IAAI,EAAE,UAAU;IAChB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACjD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;YAC7C,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACzC,CAAC,CAAC,OAAO;YACT,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,CAAC,MAAM,cAAc,GAAa;IACtC,IAAI,EAAE,gBAAgB;IACtB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YACzC,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,wEAAwE;QACxE,MAAM,QAAQ,GAAG,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC;QACvD,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC;QAC7D,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,SAAS,MAAM,CAAC,QAA8B;IAC5C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,KAAK,KAAK,KAAK,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC,CAC7E,CAAC;AACJ,CAAC;AAED;;;;;;GAMG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEhE;;;;;;;GAOG;AACH,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEvD;;GAEG;AACH,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;;;;;GAOG;AACH,SAAS,YAAY,CAAC,OAAoB,EAAE,EAAU;IACpD,IAAI,KAAK,CAAC,OAAO,EAAE,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QACvE,OAAO,KAAK,CAAC;IACf,CAAC;IACD,OAAO,CACL,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5C,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,cAAc,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CACjE,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,aAAa,GAAa;IACrC,IAAI,EAAE,eAAe;IACrB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC;YACtB,OAAO,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC;gBAChE,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,MAAM,CAAC;QACb,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,KAAK,EAAE;YAC7B,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI;YAC3C,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;YAC5B,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC;YAClC,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;GAKG;AACH,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,MAAM,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAC/B,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,OAAO,IAAI,GAAG,CAAC,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACvE,IAAI,EAAE,CAAC;IACT,CAAC;IACD,IAAI,IAAI,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,CAAC,KAAK,OAAO,EAAE,CAAC;QACpD,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACxC,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,CAAC,MAAM,eAAe,GAAa;IACvC,IAAI,EAAE,iBAAiB;IACvB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,yEAAyE;QACzE,yEAAyE;QACzE,yEAAyE;QACzE,WAAW;QACX,IACE,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC;YACvB,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAC3C,CAAC;YACD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,OAAO,GAAG,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;YAC1B,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,OAAO,KAAK,CAAC,OAAO,EAAE,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,OAAO;YACtD,IAAI,CAAC,YAAY,KAAK,OAAO;YAC7B,CAAC,CAAC,QAAQ;YACV,CAAC,CAAC,MAAM,CAAC;IACb,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,SAAS,OAAO,CAAC,QAA8B;IAC7C,OAAO,CACL,QAAQ,EAAE,OAAO,KAAK,IAAI;QAC1B,QAAQ,CAAC,KAAK,KAAK,KAAK;QACxB,QAAQ,CAAC,IAAI,KAAK,CAAC,CACpB,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;GAcG;AACH,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC;IACrB,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,IAAI;IACJ,GAAG;IACH,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,IAAI;IACJ,GAAG;IACH,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,IAAI;IACJ,GAAG;IACH,GAAG;IACH,IAAI;IACJ,GAAG;IACH,KAAK;IACL,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,GAAG;IACH,IAAI;IACJ,IAAI;IACJ,GAAG;IACH,GAAG;IACH,GAAG;IACH,GAAG;IACH,IAAI;CACL,CAAC,CAAC;AAEH;;;;;;;;;;;;;GAaG;AACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEzC;;;;;;;GAOG;AACH,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC;AAEvD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,SAAS,mBAAmB,CAAC,OAAoB;IAC/C,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;IACrC,OAAO,CACL,IAAI,CAAC,IAAI,KAAK,IAAI;QAClB,IAAI,CAAC,YAAY,KAAK,EAAE;QACxB,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,GAAG;QAC3B,UAAU,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAC9C,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAa;IACxC,IAAI,EAAE,kBAAkB;IACxB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACrC,IAAI,mBAAmB,CAAC,OAAO,CAAC,EAAE,CAAC;YACjC,OAAO,QAAQ,CAAC;QAClB,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACnD,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,IAAI,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;YAC3C,OAAO,MAAM,CAAC;QAChB,CAAC;QACD,MAAM,MAAM,GAAG,YAAY,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,OAAO,MAAM,KAAK,SAAS,IAAI,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC;IACvE,CAAC;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,aAAa;IACb,cAAc;IACd,eAAe;IACf,gBAAgB;CACjB,CAAC"}

package/dist/decode/runs.d.ts

@@ -1,3 +1,4 @@
+import type { NumeralContext } from "../numerals/text.js";
 /**
  * A stretch of text that is either all Han or all something else.
  */
@@ -17,4 +18,14 @@ export interface TextRun {
  * conversion can rebuild the text around what it changed.
  */
 export declare function splitRuns(text: string): readonly TextRun[];
+/**
+ * The Han either side of a run, as the two characters a number goes on.
+ *
+ * A number's context, bar one syllable of it. The characters decide how it is
+ * read — 年 makes 1998 a year, 个 makes 3 a count and 2 两, and 第 makes 2 an
+ * ordinal — and characters are all they are, so this is known before anything
+ * has been decoded. Which is what lets the numbers be read first and the decode
+ * of the Han after them see what the digits stood for.
+ */
+export declare function surroundingCharacters(runs: readonly TextRun[], at: number): NumeralContext;
 //# sourceMappingURL=runs.d.ts.map

package/dist/decode/runs.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"runs.d.ts","sourceRoot":"","sources":["../../src/decode/runs.ts"],"names":[],"mappings":"AAEA;;GAEG;AACH,MAAM,WAAW,OAAO;IACtB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,8DAA8D;IAC9D,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC;CACzB;AAUD;;;;;;;;;GASG;AACH,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,OAAO,EAAE,CAsB1D"}
+{"version":3,"file":"runs.d.ts","sourceRoot":"","sources":["../../src/decode/runs.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AAG1D;;GAEG;AACH,MAAM,WAAW,OAAO;IACtB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,8DAA8D;IAC9D,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC;CACzB;AAUD;;;;;;;;;GASG;AACH,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,OAAO,EAAE,CAsB1D;AAED;;;;;;;;GAQG;AACH,wBAAgB,qBAAqB,CACnC,IAAI,EAAE,SAAS,OAAO,EAAE,EACxB,EAAE,EAAE,MAAM,GACT,cAAc,CAUhB"}

package/dist/decode/runs.js

@@ -37,4 +37,23 @@ export function splitRuns(text) {
     }
     return runs;
 }
+/**
+ * The Han either side of a run, as the two characters a number goes on.
+ *
+ * A number's context, bar one syllable of it. The characters decide how it is
+ * read — 年 makes 1998 a year, 个 makes 3 a count and 2 两, and 第 makes 2 an
+ * ordinal — and characters are all they are, so this is known before anything
+ * has been decoded. Which is what lets the numbers be read first and the decode
+ * of the Han after them see what the digits stood for.
+ */
+export function surroundingCharacters(runs, at) {
+    const previous = runs[at - 1];
+    const next = runs[at + 1];
+    return {
+        following: next?.isHan === true ? (toCharacters(next.text)[0] ?? "") : "",
+        preceding: previous?.isHan === true
+            ? (toCharacters(previous.text).at(-1) ?? "")
+            : "",
+    };
+}
 //# sourceMappingURL=runs.js.map

package/dist/decode/runs.js.map

@@ -1 +1 @@
-{"version":3,"file":"runs.js","sourceRoot":"","sources":["../../src/decode/runs.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAWvD;;;;;GAKG;AACH,MAAM,GAAG,GAAG,iBAAiB,CAAC;AAE9B;;;;;;;;;GASG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,MAAM,IAAI,GAAc,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,IAAI,YAAiC,CAAC;IAEtC,KAAK,MAAM,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC;QAC3C,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAClC,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;YAC/B,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;aAAM,IAAI,KAAK,KAAK,YAAY,EAAE,CAAC;YAClC,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;YAClD,OAAO,GAAG,EAAE,CAAC;YACb,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;QACD,OAAO,IAAI,SAAS,CAAC;IACvB,CAAC;IAED,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;QAC/B,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;IACpD,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC"}
+{"version":3,"file":"runs.js","sourceRoot":"","sources":["../../src/decode/runs.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAWvD;;;;;GAKG;AACH,MAAM,GAAG,GAAG,iBAAiB,CAAC;AAE9B;;;;;;;;;GASG;AACH,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,MAAM,IAAI,GAAc,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,IAAI,YAAiC,CAAC;IAEtC,KAAK,MAAM,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC;QAC3C,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAClC,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;YAC/B,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;aAAM,IAAI,KAAK,KAAK,YAAY,EAAE,CAAC;YAClC,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;YAClD,OAAO,GAAG,EAAE,CAAC;YACb,YAAY,GAAG,KAAK,CAAC;QACvB,CAAC;QACD,OAAO,IAAI,SAAS,CAAC;IACvB,CAAC;IAED,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;QAC/B,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;IACpD,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,qBAAqB,CACnC,IAAwB,EACxB,EAAU;IAEV,MAAM,QAAQ,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC9B,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC1B,OAAO;QACL,SAAS,EAAE,IAAI,EAAE,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE;QACzE,SAAS,EACP,QAAQ,EAAE,KAAK,KAAK,IAAI;YACtB,CAAC,CAAC,CAAC,YAAY,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC5C,CAAC,CAAC,EAAE;KACT,CAAC;AACJ,CAAC"}

package/dist/decode/segment.d.ts

@@ -0,0 +1,66 @@
+import type { Dictionary } from "../dictionary/dictionary.js";
+import type { Syllable } from "../syllable/syllable.js";
+/**
+ * One word of a segmented text, or one stretch that was never Han.
+ */
+export interface Segment {
+    /** The characters, exactly as the text writes them. */
+    readonly text: string;
+    /**
+     * Where it starts, in code points from the start of the text.
+     *
+     * Code points rather than UTF-16 units, so a character outside the basic
+     * plane counts as the one character it is. 𠮷 is a surname and one position.
+     */
+    readonly at: number;
+    /**
+     * The 普通话 reading, or empty for a stretch that was never Han.
+     *
+     * Not one syllable per character: 玩儿 is `wánr`. Where a text needs its
+     * readings written out rather than inspected, `convertPieces` is the one
+     * that lines syllables up with the characters they read.
+     */
+    readonly reading: readonly Syllable[];
+    /** jieba's tag, or the empty string where there is none. */
+    readonly partOfSpeech: string;
+    readonly isProperNoun: boolean;
+    /**
+     * Whether the dictionary holds an entry for exactly these characters.
+     *
+     * Not the same as "is a word": a single character standing on its own is
+     * known, because the dictionary has an entry for the character. What it is
+     * false for is a stretch that was never Han, and a Han character no source
+     * has a reading for at all.
+     */
+    readonly isKnown: boolean;
+}
+/**
+ * Split a text into the words the decoder finds in it.
+ *
+ * Segmentation is not a feature this package added; it is what converting has
+ * always had to do first, because the unit a reading belongs to is the word.
+ * 行 is `xíng`, `háng`, `héng` or `hàng` and nothing about the character says
+ * which — only 银行 and 行长 do. This exposes the answer the decoder already
+ * works out, which was previously thrown away at the boundary: `convert`
+ * returns a string and `convertPieces` returns syllables, and neither says
+ * where the words were.
+ *
+ * ```ts
+ * segment(dictionary, "我要去北京。").map((found) => found.text);
+ * // ["我", "要", "去", "北京", "。"]
+ * ```
+ *
+ * **Every stretch of the text comes back, in order**, including the stretches
+ * that were never Han, so `segment(d, text).map((f) => f.text).join("")` is
+ * the text again. A caller wanting only the words can filter on `isKnown`, and
+ * one wanting to highlight in place has `at`.
+ *
+ * The boundaries are the ones `convert` uses, down to the number in front of a
+ * run: 2个人 is 个 + 人 because the 两 counts with the 个, where 个人 alone is
+ * the word `gèrén`. What is *not* applied is 分词连写, the word spacing GB/T
+ * 16159 wants in written pinyin — 他看了 segments as 他 / 看 / 了 and converts
+ * as `tā kànle`, because attaching an aspect particle to its verb is a fact
+ * about writing pinyin rather than about where the words are.
+ */
+export declare function segment(dictionary: Dictionary, text: string): readonly Segment[];
+//# sourceMappingURL=segment.d.ts.map

package/dist/decode/segment.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"segment.d.ts","sourceRoot":"","sources":["../../src/decode/segment.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAC;AAKxD;;GAEG;AACH,MAAM,WAAW,OAAO;IACtB,uDAAuD;IACvD,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB;;;;;OAKG;IACH,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAC;IACpB;;;;;;OAMG;IACH,QAAQ,CAAC,OAAO,EAAE,SAAS,QAAQ,EAAE,CAAC;IACtC,4DAA4D;IAC5D,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC;IAC9B,QAAQ,CAAC,YAAY,EAAE,OAAO,CAAC;IAC/B;;;;;;;OAOG;IACH,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC;CAC3B;AAgBD;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,wBAAgB,OAAO,CACrB,UAAU,EAAE,UAAU,EACtB,IAAI,EAAE,MAAM,GACX,SAAS,OAAO,EAAE,CAuCpB"}

package/dist/decode/segment.js

@@ -0,0 +1,79 @@
+import { readNumbersIn } from "../numerals/text.js";
+import { toCharacters } from "../script/characters.js";
+import { decodeRun } from "./decode.js";
+import { READING_RULES } from "./reading-rules.js";
+import { splitRuns, surroundingCharacters } from "./runs.js";
+/**
+ * A stretch of text with no reading: punctuation, Latin, digits, whitespace.
+ */
+function unread(text, at) {
+    return {
+        text,
+        at,
+        reading: [],
+        partOfSpeech: "",
+        isProperNoun: false,
+        isKnown: false,
+    };
+}
+/**
+ * Split a text into the words the decoder finds in it.
+ *
+ * Segmentation is not a feature this package added; it is what converting has
+ * always had to do first, because the unit a reading belongs to is the word.
+ * 行 is `xíng`, `háng`, `héng` or `hàng` and nothing about the character says
+ * which — only 银行 and 行长 do. This exposes the answer the decoder already
+ * works out, which was previously thrown away at the boundary: `convert`
+ * returns a string and `convertPieces` returns syllables, and neither says
+ * where the words were.
+ *
+ * ```ts
+ * segment(dictionary, "我要去北京。").map((found) => found.text);
+ * // ["我", "要", "去", "北京", "。"]
+ * ```
+ *
+ * **Every stretch of the text comes back, in order**, including the stretches
+ * that were never Han, so `segment(d, text).map((f) => f.text).join("")` is
+ * the text again. A caller wanting only the words can filter on `isKnown`, and
+ * one wanting to highlight in place has `at`.
+ *
+ * The boundaries are the ones `convert` uses, down to the number in front of a
+ * run: 2个人 is 个 + 人 because the 两 counts with the 个, where 个人 alone is
+ * the word `gèrén`. What is *not* applied is 分词连写, the word spacing GB/T
+ * 16159 wants in written pinyin — 他看了 segments as 他 / 看 / 了 and converts
+ * as `tā kànle`, because attaching an aspect particle to its verb is a fact
+ * about writing pinyin rather than about where the words are.
+ */
+export function segment(dictionary, text) {
+    const runs = [...splitRuns(text)];
+    // A Han run is decoded knowing the number in front of it, which is the one
+    // thing outside a run that moves a boundary inside it. Read per run rather
+    // than for the whole text, for the reason `convert` does: 1998年 is a year
+    // and 3个 is a count, and only the characters either side say which.
+    const said = runs.map((run, at) => run.isHan
+        ? ""
+        : (readNumbersIn(run.text, surroundingCharacters(runs, at)).at(-1)
+            ?.hanzi ?? ""));
+    const segments = [];
+    let at = 0;
+    for (const [index, run] of runs.entries()) {
+        if (!run.isHan) {
+            segments.push(unread(run.text, at));
+            at += toCharacters(run.text).length;
+            continue;
+        }
+        for (const word of decodeRun(dictionary, run.text, READING_RULES, said[index - 1] ?? "")) {
+            segments.push({
+                text: word.text,
+                at,
+                reading: word.reading,
+                partOfSpeech: word.partOfSpeech,
+                isProperNoun: word.isProperNoun,
+                isKnown: word.isKnown,
+            });
+            at += toCharacters(word.text).length;
+        }
+    }
+    return segments;
+}
+//# sourceMappingURL=segment.js.map

package/dist/decode/segment.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"segment.js","sourceRoot":"","sources":["../../src/decode/segment.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AAGvD,OAAO,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AACxC,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,SAAS,EAAE,qBAAqB,EAAE,MAAM,WAAW,CAAC;AAqC7D;;GAEG;AACH,SAAS,MAAM,CAAC,IAAY,EAAE,EAAU;IACtC,OAAO;QACL,IAAI;QACJ,EAAE;QACF,OAAO,EAAE,EAAE;QACX,YAAY,EAAE,EAAE;QAChB,YAAY,EAAE,KAAK;QACnB,OAAO,EAAE,KAAK;KACf,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AACH,MAAM,UAAU,OAAO,CACrB,UAAsB,EACtB,IAAY;IAEZ,MAAM,IAAI,GAAG,CAAC,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;IAClC,2EAA2E;IAC3E,2EAA2E;IAC3E,0EAA0E;IAC1E,oEAAoE;IACpE,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,EAAE,EAAE,CAChC,GAAG,CAAC,KAAK;QACP,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,CAAC,aAAa,CAAC,GAAG,CAAC,IAAI,EAAE,qBAAqB,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;YAC9D,EAAE,KAAK,IAAI,EAAE,CAAC,CACrB,CAAC;IAEF,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,EAAE,GAAG,CAAC,CAAC;IACX,KAAK,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QAC1C,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC;YACf,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;YACpC,EAAE,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;YACpC,SAAS;QACX,CAAC;QACD,KAAK,MAAM,IAAI,IAAI,SAAS,CAC1B,UAAU,EACV,GAAG,CAAC,IAAI,EACR,aAAa,EACb,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CACtB,EAAE,CAAC;YACF,QAAQ,CAAC,IAAI,CAAC;gBACZ,IAAI,EAAE,IAAI,CAAC,IAAI;gBACf,EAAE;gBACF,OAAO,EAAE,IAAI,CAAC,OAAO;gBACrB,YAAY,EAAE,IAAI,CAAC,YAAY;gBAC/B,YAAY,EAAE,IAAI,CAAC,YAAY;gBAC/B,OAAO,EAAE,IAAI,CAAC,OAAO;aACtB,CAAC,CAAC;YACH,EAAE,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;QACvC,CAAC;IACH,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/dictionary/artifact.d.ts

@@ -55,6 +55,16 @@ export declare function buildArtifact(entries: readonly DictionaryEntry[]): Dict
  * reports its key as both. Entries are returned in key order.
  */
 export declare function readArtifact(artifact: DictionaryArtifact): readonly DictionaryEntry[];
+/**
+ * The reading every key in an artifact resolves to, in tone-numbered notation.
+ *
+ * Resolved rather than read off the line, since 87.8% of multi-character lines
+ * store no reading at all and mean "the characters' defaults, in order" — and
+ * those defaults are whatever *this* artifact holds. That is exactly what makes
+ * the comparison between tiers worth doing: a tier missing a character does not
+ * only lose that character, it changes every derived reading that stands on it.
+ */
+export declare function readingsByKey(artifact: DictionaryArtifact): ReadonlyMap<string, string>;
 /**
  * The first key an artifact does not read back the same, or undefined.
  *

package/dist/dictionary/artifact.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"artifact.d.ts","sourceRoot":"","sources":["../../src/dictionary/artifact.ts"],"names":[],"mappings":"AACA,OAAO,EAEL,KAAK,QAAQ,EAEd,MAAM,yBAAyB,CAAC;AACjC,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAuDlD;;;;;;;GAOG;AACH,MAAM,WAAW,kBAAkB;IACjC,8DAA8D;IAC9D,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,sCAAsC;IACtC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,kDAAkD;IAClD,QAAQ,CAAC,WAAW,EAAE,UAAU,CAAC;CAClC;AAED;;;;;;GAMG;AACH,wBAAgB,aAAa,CAAC,SAAS,EAAE,SAAS,QAAQ,EAAE,GAAG,MAAM,CAIpE;AAED;;GAEG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,QAAQ,EAAE,GAAG,SAAS,CAQ3E;AAYD;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,eAAe,GAAG,SAAS,MAAM,EAAE,CAE1E;AA+GD;;;;;;;;;;;;GAYG;AACH,wBAAgB,aAAa,CAC3B,OAAO,EAAE,SAAS,eAAe,EAAE,GAClC,kBAAkB,CAqCpB;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,QAAQ,EAAE,kBAAkB,GAC3B,SAAS,eAAe,EAAE,CA2D5B;AAED;;;;;;;;;GASG;AACH,wBAAgB,oBAAoB,CAClC,OAAO,EAAE,SAAS,eAAe,EAAE,EACnC,QAAQ,EAAE,kBAAkB,GAC3B,MAAM,GAAG,SAAS,CAkBpB"}
+{"version":3,"file":"artifact.d.ts","sourceRoot":"","sources":["../../src/dictionary/artifact.ts"],"names":[],"mappings":"AACA,OAAO,EAEL,KAAK,QAAQ,EAEd,MAAM,yBAAyB,CAAC;AACjC,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAuDlD;;;;;;;GAOG;AACH,MAAM,WAAW,kBAAkB;IACjC,8DAA8D;IAC9D,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,sCAAsC;IACtC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,kDAAkD;IAClD,QAAQ,CAAC,WAAW,EAAE,UAAU,CAAC;CAClC;AAED;;;;;;GAMG;AACH,wBAAgB,aAAa,CAAC,SAAS,EAAE,SAAS,QAAQ,EAAE,GAAG,MAAM,CAIpE;AAED;;GAEG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,QAAQ,EAAE,GAAG,SAAS,CAQ3E;AAYD;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,eAAe,GAAG,SAAS,MAAM,EAAE,CAE1E;AA+GD;;;;;;;;;;;;GAYG;AACH,wBAAgB,aAAa,CAC3B,OAAO,EAAE,SAAS,eAAe,EAAE,GAClC,kBAAkB,CAqCpB;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,QAAQ,EAAE,kBAAkB,GAC3B,SAAS,eAAe,EAAE,CA2D5B;AAED;;;;;;;;GAQG;AACH,wBAAgB,aAAa,CAC3B,QAAQ,EAAE,kBAAkB,GAC3B,WAAW,CAAC,MAAM,EAAE,MAAM,CAAC,CAO7B;AAED;;;;;;;;;GASG;AACH,wBAAgB,oBAAoB,CAClC,OAAO,EAAE,SAAS,eAAe,EAAE,EACnC,QAAQ,EAAE,kBAAkB,GAC3B,MAAM,GAAG,SAAS,CAkBpB"}

package/dist/dictionary/artifact.js

@@ -276,6 +276,23 @@ export function readArtifact(artifact) {
     }
     return entries;
 }
+/**
+ * The reading every key in an artifact resolves to, in tone-numbered notation.
+ *
+ * Resolved rather than read off the line, since 87.8% of multi-character lines
+ * store no reading at all and mean "the characters' defaults, in order" — and
+ * those defaults are whatever *this* artifact holds. That is exactly what makes
+ * the comparison between tiers worth doing: a tier missing a character does not
+ * only lose that character, it changes every derived reading that stands on it.
+ */
+export function readingsByKey(artifact) {
+    const index = KeyIndex.from(artifact.keys);
+    const readings = new Map();
+    for (const [at, entry] of readArtifact(artifact).entries()) {
+        readings.set(index.keyAt(at), encodeReading(entry.readings.cn));
+    }
+    return readings;
+}
 /**
  * The first key an artifact does not read back the same, or undefined.
  *

package/dist/dictionary/artifact.js.map

@@ -1 +1 @@
-{"version":3,"file":"artifact.js","sourceRoot":"","sources":["../../src/dictionary/artifact.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EACL,YAAY,EAEZ,aAAa,GACd,MAAM,yBAAyB,CAAC;AAEjC,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE1C;;GAEG;AACH,MAAM,MAAM,GAAG,IAAI,CAAC;AAEpB;;GAEG;AACH,MAAM,IAAI,GAAG,IAAI,CAAC;AAElB;;GAEG;AACH,MAAM,SAAS,GAAG,GAAG,CAAC;AAEtB;;GAEG;AACH,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAE7B;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,eAAe,GAAG,0BAA0B,CAAC;AAEnD,MAAM,kBAAkB,GAAG,GAAG,CAAC;AAE/B;;GAEG;AACH,SAAS,gBAAgB,CAAC,KAAa;IACrC,MAAM,EAAE,GAAG,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,IAAI,CAAC,CAAC;IACvD,OAAO,EAAE,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AAC1E,CAAC;AAmBD;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,SAA8B;IAC1D,OAAO,SAAS;SACb,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;SACrD,IAAI,CAAC,GAAG,CAAC,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;QAChB,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC;IACtE,OAAO,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC;QAClC,CAAC,CAAC,SAAS;QACX,CAAC,CAAE,SAAiC,CAAC;AACzC,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,IAAY,EAAE,KAAa;IAC7C,IAAI,IAAI,GAAG,KAAK,EAAE,CAAC;QACjB,OAAO,CAAC,CAAC,CAAC;IACZ,CAAC;IACD,OAAO,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC9B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAAsB;IACrD,OAAO,CAAC,KAAK,CAAC,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC,CAAC;AACrD,CAAC;AAED;;;;;;GAMG;AACH,SAAS,gBAAgB,CACvB,OAAqC,EACrC,KAAsB;IAEtB,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC/B,MAAM,OAAO,GAAG,IAAI,KAAK,KAAK,CAAC,IAAI,IAAI,IAAI,EAAE,IAAI,KAAK,IAAI,CAAC;QAC3D,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;YACzE,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,SAAS,SAAS,CAChB,OAAmC;IAEnC,MAAM,OAAO,GAAG,IAAI,GAAG,EAA2B,CAAC;IAEnD,wEAAwE;IACxE,2EAA2E;IAC3E,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACrC,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC;YAC3D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QACjC,CAAC;IACH,CAAC;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,gBAAgB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IACnC,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;GAEG;AACH,SAAS,iBAAiB,CACxB,OAA6C;IAE7C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,OAAO,EAAE,CAAC;QACnC,IAAI,YAAY,CAAC,GAAG,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnC,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;QACtD,CAAC;IACH,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,WAAW,CAClB,GAAW,EACX,OAAe,EACf,QAAqC;IAErC,MAAM,KAAK,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC;IAChC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACrB,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,KAAK,MAAM,SAAS,IAAI,KAAK,EAAE,CAAC;QAC9B,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QACtC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,KAAK,CAAC;QACf,CAAC;QACD,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACtB,CAAC;IACD,OAAO,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,OAAO,CAAC;AACvC,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,OAA0B;IAC7C,IAAI,GAAG,GAAG,OAAO,CAAC,MAAM,CAAC;IACzB,OAAO,GAAG,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QAC1C,GAAG,EAAE,CAAC;IACR,CAAC;IACD,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5C,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,aAAa,CAC3B,OAAmC;IAEnC,MAAM,OAAO,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IACnC,MAAM,QAAQ,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;IAC5C,0EAA0E;IAC1E,wEAAwE;IACxE,MAAM,OAAO,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,EAAE,EAAE,CACxD,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CACxB,CAAC;IACF,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAEpE,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,OAAO,EAAE,CAAC;QACnC,MAAM,OAAO,GAAG,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QAEjD,KAAK,CAAC,IAAI,CACR,WAAW,CAAC;YACV,WAAW,CAAC,GAAG,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO;YAClD,KAAK,CAAC,QAAQ,CAAC,EAAE,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC;YACvE,KAAK,CAAC,YAAY;YAClB,KAAK,CAAC,YAAY;gBAChB,CAAC,CAAC,gBAAgB;oBAChB,CAAC,KAAK,CAAC,cAAc,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC;gBACvD,CAAC,CAAC,EAAE;YACN,CAAC,KAAK,CAAC,UAAU,IAAI,EAAE,CAAC;iBACrB,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;iBAC5C,IAAI,CAAC,SAAS,CAAC;SACnB,CAAC,CACH,CAAC;QACF,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IACpC,CAAC;IAED,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,SAAS,EAAE;QACvB,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;QACzB,WAAW,EAAE,cAAc,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,SAAS,EAAE;KAC3D,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,QAA4B;IAE5B,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1E,MAAM,WAAW,GAAG,cAAc,CAAC,IAAI,CAAC,QAAQ,CAAC,WAAW,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;IAE1E,4EAA4E;IAC5E,2CAA2C;IAC3C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC;QACvC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC5B,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACzD,IAAI,YAAY,CAAC,GAAG,CAAC,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;YACrD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QAC7B,CAAC;IACH,CAAC;IAED,MAAM,OAAO,GAAsB,EAAE,CAAC;IACtC,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC;QACvC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC5B,MAAM,CACJ,OAAO,GAAG,EAAE,EACZ,MAAM,GAAG,EAAE,EACX,YAAY,GAAG,EAAE,EACjB,KAAK,GAAG,EAAE,EACV,UAAU,GAAG,EAAE,EAChB,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QACpC,MAAM,cAAc,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC;QAE/C,MAAM,MAAM,GACV,OAAO,KAAK,EAAE;YACZ,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC;iBACd,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;iBACjD,IAAI,CAAC,GAAG,CAAC;YACd,CAAC,CAAC,OAAO,CAAC;QACd,MAAM,EAAE,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QACjC,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,MAAM,IAAI,KAAK,CAAC,sBAAsB,GAAG,0BAA0B,CAAC,CAAC;QACvE,CAAC;QACD,MAAM,EAAE,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QAEjC,0EAA0E;QAC1E,0EAA0E;QAC1E,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,GAAG;YACT,IAAI,EAAE,GAAG;YACT,QAAQ,EAAE,EAAE,EAAE,EAAE,GAAG,CAAC,EAAE,KAAK,SAAS,IAAI,EAAE,EAAE,EAAE,CAAC,EAAE;YACjD,SAAS,EAAE,WAAW,CAAC,QAAQ,CAAC,EAAE,CAAC;YACnC,YAAY;YACZ,YAAY,EAAE,KAAK,CAAC,QAAQ,CAAC,gBAAgB,CAAC;YAC9C,GAAG,CAAC,cAAc,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,cAAc,EAAE,CAAC;YACpD,GAAG,CAAC,UAAU,KAAK,EAAE,IAAI;gBACvB,UAAU,EAAE,UAAU;qBACnB,KAAK,CAAC,SAAS,CAAC;qBAChB,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;qBAClC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,KAAK,SAAS,CAAC;aAC9C,CAAC;SACH,CAAC,CAAC;IACL,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,oBAAoB,CAClC,OAAmC,EACnC,QAA4B;IAE5B,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,IAAI,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACpC,MAAM,KAAK,GAAG,IAAI,GAAG,EAA2B,CAAC;IACjD,KAAK,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QACzC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC;IACpC,CAAC;IAED,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,SAAS,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAC7B,IACE,KAAK,KAAK,SAAS;YACnB,CAAC,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,EACpD,CAAC;YACD,OAAO,GAAG,CAAC;QACb,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}
+{"version":3,"file":"artifact.js","sourceRoot":"","sources":["../../src/dictionary/artifact.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC;AACvD,OAAO,EACL,YAAY,EAEZ,aAAa,GACd,MAAM,yBAAyB,CAAC;AAEjC,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE1C;;GAEG;AACH,MAAM,MAAM,GAAG,IAAI,CAAC;AAEpB;;GAEG;AACH,MAAM,IAAI,GAAG,IAAI,CAAC;AAElB;;GAEG;AACH,MAAM,SAAS,GAAG,GAAG,CAAC;AAEtB;;GAEG;AACH,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAE7B;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,eAAe,GAAG,0BAA0B,CAAC;AAEnD,MAAM,kBAAkB,GAAG,GAAG,CAAC;AAE/B;;GAEG;AACH,SAAS,gBAAgB,CAAC,KAAa;IACrC,MAAM,EAAE,GAAG,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,IAAI,CAAC,CAAC;IACvD,OAAO,EAAE,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AAC1E,CAAC;AAmBD;;;;;;GAMG;AACH,MAAM,UAAU,aAAa,CAAC,SAA8B;IAC1D,OAAO,SAAS;SACb,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;SACrD,IAAI,CAAC,GAAG,CAAC,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;QAChB,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,YAAY,CAAC,KAAK,CAAC,CAAC,CAAC;IACtE,OAAO,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC;QAClC,CAAC,CAAC,SAAS;QACX,CAAC,CAAE,SAAiC,CAAC;AACzC,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,IAAY,EAAE,KAAa;IAC7C,IAAI,IAAI,GAAG,KAAK,EAAE,CAAC;QACjB,OAAO,CAAC,CAAC,CAAC;IACZ,CAAC;IACD,OAAO,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC9B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAAsB;IACrD,OAAO,CAAC,KAAK,CAAC,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC,CAAC;AACrD,CAAC;AAED;;;;;;GAMG;AACH,SAAS,gBAAgB,CACvB,OAAqC,EACrC,KAAsB;IAEtB,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC/B,MAAM,OAAO,GAAG,IAAI,KAAK,KAAK,CAAC,IAAI,IAAI,IAAI,EAAE,IAAI,KAAK,IAAI,CAAC;QAC3D,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;YACzE,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,SAAS,SAAS,CAChB,OAAmC;IAEnC,MAAM,OAAO,GAAG,IAAI,GAAG,EAA2B,CAAC;IAEnD,wEAAwE;IACxE,2EAA2E;IAC3E,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACrC,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC;YAC3D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QACjC,CAAC;IACH,CAAC;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,gBAAgB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IACnC,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;GAEG;AACH,SAAS,iBAAiB,CACxB,OAA6C;IAE7C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,OAAO,EAAE,CAAC;QACnC,IAAI,YAAY,CAAC,GAAG,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnC,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;QACtD,CAAC;IACH,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,WAAW,CAClB,GAAW,EACX,OAAe,EACf,QAAqC;IAErC,MAAM,KAAK,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC;IAChC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACrB,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,KAAK,MAAM,SAAS,IAAI,KAAK,EAAE,CAAC;QAC9B,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QACtC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,KAAK,CAAC;QACf,CAAC;QACD,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACtB,CAAC;IACD,OAAO,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,OAAO,CAAC;AACvC,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,OAA0B;IAC7C,IAAI,GAAG,GAAG,OAAO,CAAC,MAAM,CAAC;IACzB,OAAO,GAAG,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QAC1C,GAAG,EAAE,CAAC;IACR,CAAC;IACD,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5C,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,aAAa,CAC3B,OAAmC;IAEnC,MAAM,OAAO,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IACnC,MAAM,QAAQ,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;IAC5C,0EAA0E;IAC1E,wEAAwE;IACxE,MAAM,OAAO,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,EAAE,EAAE,CACxD,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CACxB,CAAC;IACF,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAEpE,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,OAAO,EAAE,CAAC;QACnC,MAAM,OAAO,GAAG,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QAEjD,KAAK,CAAC,IAAI,CACR,WAAW,CAAC;YACV,WAAW,CAAC,GAAG,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO;YAClD,KAAK,CAAC,QAAQ,CAAC,EAAE,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC;YACvE,KAAK,CAAC,YAAY;YAClB,KAAK,CAAC,YAAY;gBAChB,CAAC,CAAC,gBAAgB;oBAChB,CAAC,KAAK,CAAC,cAAc,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC;gBACvD,CAAC,CAAC,EAAE;YACN,CAAC,KAAK,CAAC,UAAU,IAAI,EAAE,CAAC;iBACrB,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;iBAC5C,IAAI,CAAC,SAAS,CAAC;SACnB,CAAC,CACH,CAAC;QACF,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IACpC,CAAC;IAED,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,SAAS,EAAE;QACvB,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;QACzB,WAAW,EAAE,cAAc,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,SAAS,EAAE;KAC3D,CAAC;AACJ,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,QAA4B;IAE5B,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1E,MAAM,WAAW,GAAG,cAAc,CAAC,IAAI,CAAC,QAAQ,CAAC,WAAW,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;IAE1E,4EAA4E;IAC5E,2CAA2C;IAC3C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC;QACvC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC5B,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACzD,IAAI,YAAY,CAAC,GAAG,CAAC,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,KAAK,EAAE,EAAE,CAAC;YACrD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QAC7B,CAAC;IACH,CAAC;IAED,MAAM,OAAO,GAAsB,EAAE,CAAC;IACtC,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC;QACvC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC5B,MAAM,CACJ,OAAO,GAAG,EAAE,EACZ,MAAM,GAAG,EAAE,EACX,YAAY,GAAG,EAAE,EACjB,KAAK,GAAG,EAAE,EACV,UAAU,GAAG,EAAE,EAChB,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QACpC,MAAM,cAAc,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC;QAE/C,MAAM,MAAM,GACV,OAAO,KAAK,EAAE;YACZ,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC;iBACd,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;iBACjD,IAAI,CAAC,GAAG,CAAC;YACd,CAAC,CAAC,OAAO,CAAC;QACd,MAAM,EAAE,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QACjC,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,MAAM,IAAI,KAAK,CAAC,sBAAsB,GAAG,0BAA0B,CAAC,CAAC;QACvE,CAAC;QACD,MAAM,EAAE,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QAEjC,0EAA0E;QAC1E,0EAA0E;QAC1E,OAAO,CAAC,IAAI,CAAC;YACX,IAAI,EAAE,GAAG;YACT,IAAI,EAAE,GAAG;YACT,QAAQ,EAAE,EAAE,EAAE,EAAE,GAAG,CAAC,EAAE,KAAK,SAAS,IAAI,EAAE,EAAE,EAAE,CAAC,EAAE;YACjD,SAAS,EAAE,WAAW,CAAC,QAAQ,CAAC,EAAE,CAAC;YACnC,YAAY;YACZ,YAAY,EAAE,KAAK,CAAC,QAAQ,CAAC,gBAAgB,CAAC;YAC9C,GAAG,CAAC,cAAc,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,cAAc,EAAE,CAAC;YACpD,GAAG,CAAC,UAAU,KAAK,EAAE,IAAI;gBACvB,UAAU,EAAE,UAAU;qBACnB,KAAK,CAAC,SAAS,CAAC;qBAChB,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;qBAClC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,KAAK,SAAS,CAAC;aAC9C,CAAC;SACH,CAAC,CAAC;IACL,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,aAAa,CAC3B,QAA4B;IAE5B,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,KAAK,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,YAAY,CAAC,QAAQ,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC;QAC3D,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;IAClE,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,oBAAoB,CAClC,OAAmC,EACnC,QAA4B;IAE5B,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,IAAI,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACpC,MAAM,KAAK,GAAG,IAAI,GAAG,EAA2B,CAAC;IACjD,KAAK,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;QACzC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC;IACpC,CAAC;IAED,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,SAAS,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAC7B,IACE,KAAK,KAAK,SAAS;YACnB,CAAC,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,EACpD,CAAC;YACD,OAAO,GAAG,CAAC;QACb,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/dist/dictionary/assertions.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"assertions.d.ts","sourceRoot":"","sources":["../../src/dictionary/assertions.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAElD;;GAEG;AACH,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,kEAAkE;IAClE,QAAQ,CAAC,KAAK,EAAE,CAAC,UAAU,EAAE,eAAe,KAAK,MAAM,GAAG,SAAS,CAAC;CACrE;AAED;;GAEG;AACH,qBAAa,eAAe;;IAG1B,QAAQ,CAAC,OAAO,EAAE,SAAS,eAAe,EAAE,CAAC;IAE7C;;;;;OAKG;IACH,YAAY,OAAO,EAAE,SAAS,eAAe,EAAE,EAc9C;IAeD;;OAEG;IACH,GAAG,CAAC,IAAI,EAAE,MAAM,GAAG,eAAe,GAAG,SAAS,CAE7C;IAED;;OAEG;IACH,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAKxC;IAED;;OAEG;IACH,iBAAiB,IAAI,WAAW,CAAC,MAAM,CAAC,CAUvC;IAED;;;;;;;;OAQG;IACH,eAAe,IAAI,WAAW,CAAC,MAAM,CAAC,CAUrC;CACF;AAoBD;;;;;;;;;;;;GAYG;AACH,eAAO,MAAM,gBAAgB,EAAE,SAAS,cAAc,EAwIrD,CAAC;AAEF;;GAEG;AACH,wBAAgB,UAAU,CACxB,OAAO,EAAE,SAAS,eAAe,EAAE,GAClC,SAAS,MAAM,EAAE,CAKnB"}
+{"version":3,"file":"assertions.d.ts","sourceRoot":"","sources":["../../src/dictionary/assertions.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAC;AAElD;;GAEG;AACH,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,kEAAkE;IAClE,QAAQ,CAAC,KAAK,EAAE,CAAC,UAAU,EAAE,eAAe,KAAK,MAAM,GAAG,SAAS,CAAC;CACrE;AAED;;GAEG;AACH,qBAAa,eAAe;;IAG1B,QAAQ,CAAC,OAAO,EAAE,SAAS,eAAe,EAAE,CAAC;IAE7C;;;;;OAKG;IACH,YAAY,OAAO,EAAE,SAAS,eAAe,EAAE,EAc9C;IAeD;;OAEG;IACH,GAAG,CAAC,IAAI,EAAE,MAAM,GAAG,eAAe,GAAG,SAAS,CAE7C;IAED;;OAEG;IACH,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAKxC;IAED;;OAEG;IACH,iBAAiB,IAAI,WAAW,CAAC,MAAM,CAAC,CAUvC;IAED;;;;;;;;OAQG;IACH,eAAe,IAAI,WAAW,CAAC,MAAM,CAAC,CAUrC;CACF;AAoBD;;;;;;;;;;;;GAYG;AACH,eAAO,MAAM,gBAAgB,EAAE,SAAS,cAAc,EAgJrD,CAAC;AAEF;;GAEG;AACH,wBAAgB,UAAU,CACxB,OAAO,EAAE,SAAS,eAAe,EAAE,GAClC,SAAS,MAAM,EAAE,CAKnB"}

package/dist/dictionary/assertions.js

@@ -135,6 +135,14 @@ export const BUILD_ASSERTIONS = [
     reads("重複", "chóng fù", "one of two 繁體 spellings, both keyed"),
     reads("重覆", "chóng fù", "the other, on the same entry"),
     reads("下麵", "xià miàn", "both spellings of a word CC-CEDICT writes twice"),
+    // The 简体 phrase corpus carries a few 繁體 headwords and reads them as
+    // though the characters were 简体 — 徵 as `zhǐ` rather than as 征, 沈 as the
+    // surname rather than as 沉. Left as entries of their own they outrank the
+    // 繁體 key derived from the 简体 entry, and the word reads one way in a tier
+    // holding the phrase tail and another way in a tier without it.
+    reads("特徵", "tè zhēng", "not tè zhǐ, which is 徵 read as itself"),
+    reads("沈溺", "chén nì", "not shěn nì, which is 沈 read as the surname"),
+    reads("蝨子", "shī zi", "not shī zǐ, which is 子 read with its full tone"),
     // kHanyuPinlu writes 李 as `li(36)`, with no tone mark. Read as 轻声 — which
     // is what an unmarked reading means everywhere else in source data — 李华
     // comes out `Li Huá`. The other Unihan fields all write `lǐ`, and that is

package/dist/dictionary/assertions.js.map

@@ -1 +1 @@
-{"version":3,"file":"assertions.js","sourceRoot":"","sources":["../../src/dictionary/assertions.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AACzD,OAAO,EAAE,oBAAoB,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAC;AAChF,OAAO,EAAiB,aAAa,EAAE,MAAM,yBAAyB,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AAYjD;;GAEG;AACH,MAAM,OAAO,eAAe;IACjB,MAAM,CAAuC;IAE7C,OAAO,CAA6B;IAE7C;;;;;OAKG;IACH,YAAY,OAAmC;QAC7C,MAAM,KAAK,GAAG,IAAI,GAAG,EAA2B,CAAC;QACjD,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC/B,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3C,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;oBACrB,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;gBACzB,CAAC;YACH,CAAC;QACH,CAAC;QACD,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC;QACpB,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,CAAC,SAAS;QACR,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YACjC,KAAK,CAAC,CAAC;gBACL,KAAK,CAAC,QAAQ,CAAC,EAAE;gBACjB,KAAK,CAAC,QAAQ,CAAC,EAAE,IAAI,EAAE;gBACvB,GAAG,CAAC,KAAK,CAAC,UAAU,IAAI,EAAE,CAAC;aAC5B,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,GAAG,CAAC,IAAY;QACd,OAAO,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,OAAO,CAAC,IAAY;QAClB,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QACpC,OAAO,KAAK,KAAK,SAAS;YACxB,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7E,CAAC;IAED;;OAEG;IACH,iBAAiB;QACf,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;QACpC,KAAK,MAAM,OAAO,IAAI,IAAI,CAAC,SAAS,EAAE,EAAE,CAAC;YACvC,KAAK,MAAM,QAAQ,IAAI,OAAO,EAAE,CAAC;gBAC/B,SAAS,CAAC,GAAG,CACX,aAAa,CAAC,EAAE,GAAG,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAC9D,CAAC;YACJ,CAAC;QACH,CAAC;QACD,OAAO,SAAS,CAAC;IACnB,CAAC;IAED;;;;;;;;OAQG;IACH,eAAe;QACb,MAAM,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;QACrC,KAAK,MAAM,OAAO,IAAI,IAAI,CAAC,SAAS,EAAE,EAAE,CAAC;YACvC,KAAK,MAAM,QAAQ,IAAI,OAAO,EAAE,CAAC;gBAC/B,IAAI,CAAC,cAAc,CAAC,QAAQ,CAAC,EAAE,CAAC;oBAC9B,UAAU,CAAC,GAAG,CAAC,aAAa,CAAC,EAAE,GAAG,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;gBAC/D,CAAC;YACH,CAAC;QACH,CAAC;QACD,OAAO,UAAU,CAAC;IACpB,CAAC;CACF;AAED;;GAEG;AACH,SAAS,KAAK,CAAC,IAAY,EAAE,QAAgB,EAAE,GAAW;IACxD,OAAO;QACL,WAAW,EAAE,GAAG,IAAI,MAAM,QAAQ,KAAK,GAAG,GAAG;QAC7C,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,MAAM,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;gBACzB,OAAO,GAAG,IAAI,iCAAiC,CAAC;YAClD,CAAC;YACD,OAAO,MAAM,KAAK,QAAQ;gBACxB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,GAAG,IAAI,UAAU,MAAM,cAAc,QAAQ,EAAE,CAAC;QACtD,CAAC;KACF,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAA8B;IACzD,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,iCAAiC,CAAC;IACtD,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,+BAA+B,CAAC;IACrD,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,8CAA8C,CAAC;IACnE,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,6BAA6B,CAAC;IACnD,KAAK,CAAC,MAAM,EAAE,gBAAgB,EAAE,yBAAyB,CAAC;IAC1D,KAAK,CAAC,MAAM,EAAE,gBAAgB,EAAE,+BAA+B,CAAC;IAChE,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,oCAAoC,CAAC;IAC3D,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,oBAAoB,CAAC;IAC7C,KAAK,CAAC,IAAI,EAAE,YAAY,EAAE,oBAAoB,CAAC;IAC/C,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,oCAAoC,CAAC;IAC3D,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,oCAAoC,CAAC;IAC5D,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,kDAAkD,CAAC;IACzE,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,qCAAqC,CAAC;IAC9D,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,8BAA8B,CAAC;IACvD,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,iDAAiD,CAAC;IAC1E,0EAA0E;IAC1E,wEAAwE;IACxE,0EAA0E;IAC1E,4EAA4E;IAC5E,uDAAuD;IACvD,KAAK,CAAC,GAAG,EAAE,IAAI,EAAE,yCAAyC,CAAC;IAC3D,KAAK,CAAC,GAAG,EAAE,KAAK,EAAE,+CAA+C,CAAC;IAClE;QACE,WAAW,EAAE,iDAAiD;QAC9D,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE,CACzD,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI;YACzC,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,mCAAmC;KAC1C;IACD;QACE,qEAAqE;QACrE,2EAA2E;QAC3E,WAAW,EAAE,wCAAwC;QACrD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,KAAK,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,MAAM,CAC3C,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI,CACtD,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,2BAA2B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACpD,CAAC;KACF;IACD;QACE,kEAAkE;QAClE,6DAA6D;QAC7D,WAAW,EAAE,mCAAmC;QAChD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,KAAK,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,MAAM,CACvC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI,CACtD,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,2BAA2B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACpD,CAAC;KACF;IACD;QACE,yEAAyE;QACzE,gEAAgE;QAChE,sEAAsE;QACtE,WAAW,EACT,sDAAsD;QACxD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,MAAM,GAAG,CAAC,IAAY,EAAsB,EAAE,CAClD,UAAU;iBACP,GAAG,CAAC,IAAI,CAAC;gBACV,EAAE,QAAQ,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;iBACxD,IAAI,CAAC,GAAG,CAAC,CAAC;YACf,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAChC,IAAI,QAAQ,KAAK,eAAe,EAAE,CAAC;gBACjC,OAAO,cAAc,QAAQ,IAAI,SAAS,mCAAmC,CAAC;YAChF,CAAC;YACD,MAAM,KAAK,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAC3C,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,CACrC,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,iCAAiC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1D,CAAC;KACF;IACD;QACE,WAAW,EAAE,8BAA8B;QAC3C,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,CAAC;YACxC,OAAO,IAAI,KAAK,IAAI;gBAClB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,cAAc,IAAI,IAAI,SAAS,eAAe,CAAC;QACrD,CAAC;KACF;IACD;QACE,4EAA4E;QAC5E,sEAAsE;QACtE,0EAA0E;QAC1E,0EAA0E;QAC1E,4EAA4E;QAC5E,kCAAkC;QAClC,WAAW,EAAE,gDAAgD;QAC7D,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,OAAO,GAAG,CAAC,GAAG,UAAU,CAAC,iBAAiB,EAAE,CAAC,CAAC,MAAM,CACxD,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,oBAAoB,CAAC,GAAG,CAAC,QAAQ,CAAC,CAClD,CAAC;YACF,OAAO,OAAO,CAAC,MAAM,KAAK,CAAC;gBACzB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,gBAAgB,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,uCAAuC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACxG,CAAC;KACF;IACD;QACE,yEAAyE;QACzE,uEAAuE;QACvE,0EAA0E;QAC1E,oEAAoE;QACpE,WAAW,EAAE,sDAAsD;QACnE,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,UAAU,GAAG,CAAC,GAAG,UAAU,CAAC,eAAe,EAAE,CAAC,CAAC;YACrD,OAAO,UAAU,CAAC,MAAM,KAAK,CAAC;gBAC5B,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,gBAAgB,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,uDAAuD,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC9H,CAAC;KACF;IACD;QACE,WAAW,EAAE,0DAA0D;QACvE,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,KAAK,MAAM,KAAK,IAAI,UAAU,CAAC,OAAO,EAAE,CAAC;gBACvC,MAAM,UAAU,GAAG,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;gBAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,CACpC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,KAAK,KAAK,IAAI,CACtC,CAAC,MAAM,CAAC;gBACT,iEAAiE;gBACjE,iEAAiE;gBACjE,IAAI,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,GAAG,KAAK,GAAG,UAAU,EAAE,CAAC;oBAClD,OAAO,GAAG,KAAK,CAAC,IAAI,QAAQ,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,CAAC,kBAAkB,MAAM,CAAC,UAAU,CAAC,aAAa,CAAC;gBAChH,CAAC;YACH,CAAC;YACD,OAAO;QACT,CAAC;KACF;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,UAAU,UAAU,CACxB,OAAmC;IAEnC,MAAM,UAAU,GAAG,IAAI,eAAe,CAAC,OAAO,CAAC,CAAC;IAChD,OAAO,gBAAgB,CAAC,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CACxC,SAAS,CAAC,KAAK,CAAC,UAAU,CAAC,CAC5B,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC;AAC/C,CAAC"}
+{"version":3,"file":"assertions.js","sourceRoot":"","sources":["../../src/dictionary/assertions.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AACzD,OAAO,EAAE,oBAAoB,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAC;AAChF,OAAO,EAAiB,aAAa,EAAE,MAAM,yBAAyB,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,eAAe,CAAC;AAYjD;;GAEG;AACH,MAAM,OAAO,eAAe;IACjB,MAAM,CAAuC;IAE7C,OAAO,CAA6B;IAE7C;;;;;OAKG;IACH,YAAY,OAAmC;QAC7C,MAAM,KAAK,GAAG,IAAI,GAAG,EAA2B,CAAC;QACjD,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAC/B,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3C,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;oBACrB,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;gBACzB,CAAC;YACH,CAAC;QACH,CAAC;QACD,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC;QACpB,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IACzB,CAAC;IAED;;OAEG;IACH,CAAC,SAAS;QACR,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;YACjC,KAAK,CAAC,CAAC;gBACL,KAAK,CAAC,QAAQ,CAAC,EAAE;gBACjB,KAAK,CAAC,QAAQ,CAAC,EAAE,IAAI,EAAE;gBACvB,GAAG,CAAC,KAAK,CAAC,UAAU,IAAI,EAAE,CAAC;aAC5B,CAAC;QACJ,CAAC;IACH,CAAC;IAED;;OAEG;IACH,GAAG,CAAC,IAAY;QACd,OAAO,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAC/B,CAAC;IAED;;OAEG;IACH,OAAO,CAAC,IAAY;QAClB,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QACpC,OAAO,KAAK,KAAK,SAAS;YACxB,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7E,CAAC;IAED;;OAEG;IACH,iBAAiB;QACf,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;QACpC,KAAK,MAAM,OAAO,IAAI,IAAI,CAAC,SAAS,EAAE,EAAE,CAAC;YACvC,KAAK,MAAM,QAAQ,IAAI,OAAO,EAAE,CAAC;gBAC/B,SAAS,CAAC,GAAG,CACX,aAAa,CAAC,EAAE,GAAG,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAC9D,CAAC;YACJ,CAAC;QACH,CAAC;QACD,OAAO,SAAS,CAAC;IACnB,CAAC;IAED;;;;;;;;OAQG;IACH,eAAe;QACb,MAAM,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;QACrC,KAAK,MAAM,OAAO,IAAI,IAAI,CAAC,SAAS,EAAE,EAAE,CAAC;YACvC,KAAK,MAAM,QAAQ,IAAI,OAAO,EAAE,CAAC;gBAC/B,IAAI,CAAC,cAAc,CAAC,QAAQ,CAAC,EAAE,CAAC;oBAC9B,UAAU,CAAC,GAAG,CAAC,aAAa,CAAC,EAAE,GAAG,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;gBAC/D,CAAC;YACH,CAAC;QACH,CAAC;QACD,OAAO,UAAU,CAAC;IACpB,CAAC;CACF;AAED;;GAEG;AACH,SAAS,KAAK,CAAC,IAAY,EAAE,QAAgB,EAAE,GAAW;IACxD,OAAO;QACL,WAAW,EAAE,GAAG,IAAI,MAAM,QAAQ,KAAK,GAAG,GAAG;QAC7C,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,MAAM,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;gBACzB,OAAO,GAAG,IAAI,iCAAiC,CAAC;YAClD,CAAC;YACD,OAAO,MAAM,KAAK,QAAQ;gBACxB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,GAAG,IAAI,UAAU,MAAM,cAAc,QAAQ,EAAE,CAAC;QACtD,CAAC;KACF,CAAC;AACJ,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAA8B;IACzD,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,iCAAiC,CAAC;IACtD,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,+BAA+B,CAAC;IACrD,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,8CAA8C,CAAC;IACnE,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,6BAA6B,CAAC;IACnD,KAAK,CAAC,MAAM,EAAE,gBAAgB,EAAE,yBAAyB,CAAC;IAC1D,KAAK,CAAC,MAAM,EAAE,gBAAgB,EAAE,+BAA+B,CAAC;IAChE,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,oCAAoC,CAAC;IAC3D,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,oBAAoB,CAAC;IAC7C,KAAK,CAAC,IAAI,EAAE,YAAY,EAAE,oBAAoB,CAAC;IAC/C,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,oCAAoC,CAAC;IAC3D,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,oCAAoC,CAAC;IAC5D,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,kDAAkD,CAAC;IACzE,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,qCAAqC,CAAC;IAC9D,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,8BAA8B,CAAC;IACvD,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,iDAAiD,CAAC;IAC1E,oEAAoE;IACpE,wEAAwE;IACxE,0EAA0E;IAC1E,yEAAyE;IACzE,gEAAgE;IAChE,KAAK,CAAC,IAAI,EAAE,UAAU,EAAE,uCAAuC,CAAC;IAChE,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,6CAA6C,CAAC;IACrE,KAAK,CAAC,IAAI,EAAE,QAAQ,EAAE,gDAAgD,CAAC;IACvE,0EAA0E;IAC1E,wEAAwE;IACxE,0EAA0E;IAC1E,4EAA4E;IAC5E,uDAAuD;IACvD,KAAK,CAAC,GAAG,EAAE,IAAI,EAAE,yCAAyC,CAAC;IAC3D,KAAK,CAAC,GAAG,EAAE,KAAK,EAAE,+CAA+C,CAAC;IAClE;QACE,WAAW,EAAE,iDAAiD;QAC9D,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE,CACzD,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI;YACzC,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,mCAAmC;KAC1C;IACD;QACE,qEAAqE;QACrE,2EAA2E;QAC3E,WAAW,EAAE,wCAAwC;QACrD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,KAAK,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,MAAM,CAC3C,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI,CACtD,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,2BAA2B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACpD,CAAC;KACF;IACD;QACE,kEAAkE;QAClE,6DAA6D;QAC7D,WAAW,EAAE,mCAAmC;QAChD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,KAAK,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,MAAM,CACvC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,YAAY,KAAK,IAAI,CACtD,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,2BAA2B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACpD,CAAC;KACF;IACD;QACE,yEAAyE;QACzE,gEAAgE;QAChE,sEAAsE;QACtE,WAAW,EACT,sDAAsD;QACxD,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,MAAM,GAAG,CAAC,IAAY,EAAsB,EAAE,CAClD,UAAU;iBACP,GAAG,CAAC,IAAI,CAAC;gBACV,EAAE,QAAQ,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;iBACxD,IAAI,CAAC,GAAG,CAAC,CAAC;YACf,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;YAChC,IAAI,QAAQ,KAAK,eAAe,EAAE,CAAC;gBACjC,OAAO,cAAc,QAAQ,IAAI,SAAS,mCAAmC,CAAC;YAChF,CAAC;YACD,MAAM,KAAK,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,MAAM,CAC3C,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,CACrC,CAAC;YACF,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC;gBACvB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,iCAAiC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1D,CAAC;KACF;IACD;QACE,WAAW,EAAE,8BAA8B;QAC3C,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,IAAI,CAAC;YACxC,OAAO,IAAI,KAAK,IAAI;gBAClB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,cAAc,IAAI,IAAI,SAAS,eAAe,CAAC;QACrD,CAAC;KACF;IACD;QACE,4EAA4E;QAC5E,sEAAsE;QACtE,0EAA0E;QAC1E,0EAA0E;QAC1E,4EAA4E;QAC5E,kCAAkC;QAClC,WAAW,EAAE,gDAAgD;QAC7D,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,OAAO,GAAG,CAAC,GAAG,UAAU,CAAC,iBAAiB,EAAE,CAAC,CAAC,MAAM,CACxD,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,oBAAoB,CAAC,GAAG,CAAC,QAAQ,CAAC,CAClD,CAAC;YACF,OAAO,OAAO,CAAC,MAAM,KAAK,CAAC;gBACzB,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,gBAAgB,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,uCAAuC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACxG,CAAC;KACF;IACD;QACE,yEAAyE;QACzE,uEAAuE;QACvE,0EAA0E;QAC1E,oEAAoE;QACpE,WAAW,EAAE,sDAAsD;QACnE,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,MAAM,UAAU,GAAG,CAAC,GAAG,UAAU,CAAC,eAAe,EAAE,CAAC,CAAC;YACrD,OAAO,UAAU,CAAC,MAAM,KAAK,CAAC;gBAC5B,CAAC,CAAC,SAAS;gBACX,CAAC,CAAC,gBAAgB,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,uDAAuD,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QAC9H,CAAC;KACF;IACD;QACE,WAAW,EAAE,0DAA0D;QACvE,KAAK,EAAE,CAAC,UAA2B,EAAsB,EAAE;YACzD,KAAK,MAAM,KAAK,IAAI,UAAU,CAAC,OAAO,EAAE,CAAC;gBACvC,MAAM,UAAU,GAAG,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;gBAC9C,MAAM,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,CACpC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,KAAK,KAAK,IAAI,CACtC,CAAC,MAAM,CAAC;gBACT,iEAAiE;gBACjE,iEAAiE;gBACjE,IAAI,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,GAAG,KAAK,GAAG,UAAU,EAAE,CAAC;oBAClD,OAAO,GAAG,KAAK,CAAC,IAAI,QAAQ,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC,MAAM,CAAC,kBAAkB,MAAM,CAAC,UAAU,CAAC,aAAa,CAAC;gBAChH,CAAC;YACH,CAAC;YACD,OAAO;QACT,CAAC;KACF;CACF,CAAC;AAEF;;GAEG;AACH,MAAM,UAAU,UAAU,CACxB,OAAmC;IAEnC,MAAM,UAAU,GAAG,IAAI,eAAe,CAAC,OAAO,CAAC,CAAC;IAChD,OAAO,gBAAgB,CAAC,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CACxC,SAAS,CAAC,KAAK,CAAC,UAAU,CAAC,CAC5B,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC;AAC/C,CAAC"}

package/dist/dictionary/corpus-mass.d.ts

@@ -0,0 +1,68 @@
+import type { CedictEntry } from "../sources/cedict.js";
+import type { JiebaEntry } from "../sources/jieba.js";
+import { type ReadingField } from "../sources/unihan.js";
+import type { Syllable } from "../syllable/syllable.js";
+import type { TraditionalTable } from "./traditional.js";
+/**
+ * The sources a character's corpus mass is counted over.
+ *
+ * The same three the merge already holds: the words, their readings, and how
+ * often jieba's corpus saw each word.
+ */
+export interface MassSources {
+    readonly phrase: ReadonlyMap<string, readonly string[]>;
+    readonly cedict: readonly CedictEntry[];
+    readonly jieba: ReadonlyMap<string, JiebaEntry>;
+}
+/**
+ * How much corpus each of a character's readings carries, keyed by reading.
+ */
+export type CorpusMass = ReadonlyMap<string, ReadonlyMap<string, number>>;
+/**
+ * Count how much of jieba's corpus each reading of each character carries.
+ *
+ * Every word the dictionary knows votes for the reading it gives each of its
+ * characters, weighted by how often jieba's corpus saw that word. It is the one
+ * piece of evidence about a polyphone that is neither a lexicographer's pick nor
+ * a dictionary's page order: 行 is `xíng` because 进行, 行政 and 行为 outweigh
+ * 行业 and 银行, and nothing had to say so.
+ *
+ * **Single characters do not vote.** A one-character entry's reading *is* the
+ * default being ranked — it comes from Unihan, or from a source that picked one
+ * reading for the character — so counting it would have the ranking confirm
+ * itself. Only a word puts a character in a context that chose between its
+ * readings.
+ *
+ * **繁體 is counted through the script pairing.** jieba's corpus is 简体 and the
+ * phrase corpus is 简体, so a traditional character appears in none of the words
+ * counted here and would rank on nothing at all. Each vote is therefore also
+ * cast for whichever 繁體 form the word's reading picks — 髮 gets 头发's `fà` and
+ * 發 gets 发送's `fā` — which is the same evidence reaching the same character
+ * under the other script.
+ */
+export declare function countCorpusMass(sources: MassSources, traditional: TraditionalTable): CorpusMass;
+/**
+ * Re-rank one character's readings by how much corpus each carries.
+ *
+ * The sort is stable and the mass of an unseen reading is zero, so a character
+ * no word covers keeps Unihan's order exactly, and so does the tail of readings
+ * below the ones words attest. The corpus only ever reorders what it has
+ * evidence about.
+ *
+ * **Only where `kHanyuPinlu` did not already rank the character**, which is the
+ * whole of what this instrument is for. `kMandarin` names one reading and orders
+ * nothing, so where it decides the default there is nothing behind it, and that
+ * is what the corpus supplies. Where `kHanyuPinlu` lists several readings there
+ * already is a corpus ranking, and it is the better one: **this mass counts a
+ * character only inside words, so a reading used mostly as a bare character is
+ * invisible to it.** 呢 is 呢子 and 呢绒 in the dictionary and 呢 on the end of a
+ * question everywhere else; 都 is 首都 and 都市 against a bare 都 that is simply
+ * `dōu`; 将, 与, 处, 喝, 蒙 and 吗 are the same shape. `kHanyuPinlu` counts
+ * occurrences rather than words and has seen exactly what this cannot.
+ *
+ * Measured over CPP, that division is the whole difference between the rule
+ * paying and not: where `kHanyuPinlu` is silent or alone the mass fixes 438 and
+ * breaks 256, and where it ranks the character it fixes 201 and breaks 263.
+ */
+export declare function rankByCorpusMass(character: string, readings: readonly Syllable[], fields: ReadonlyMap<ReadingField, readonly string[]>, mass: CorpusMass): readonly Syllable[];
+//# sourceMappingURL=corpus-mass.d.ts.map