@kensio/pinyinjs 1.7.3 → 1.7.5

package/dist/decode/reading-rules.js

@@ -1,4 +1,6 @@
 import { QUANTITY_CHARACTERS } from "../numerals/characters.js";
+import { TEACHING_JIAO } from "./jiao-rule.js";
+export { TEACHING_JIAO } from "./jiao-rule.js";
 import { toCharacters } from "../script/characters.js";
 import { tagOf, wordEndingAt, wordStartingAt, } from "./rules.js";
 /**
@@ -102,93 +104,6 @@ export const ATTESTED_ERHUA = {
         return (reading?.at(-1)?.erhua ?? false) ? "forbid" : "keep";
     },
 };
-/**
- * Whether a syllable is `jiāo`, the verbal reading of 教.
- */
-function isJiao(syllable) {
-    return (syllable?.initial === "j" && syllable.final === "iao" && syllable.tone === 1);
-}
-/**
- * What a 教 that is teaching governs: the thing or the person taught.
- *
- * A pronoun, a common noun, a name or one of the tags jieba gives a subject —
- * 教我, 教英语, 教历史, 教孩子, 教玛丽. `ns` is deliberately absent, since the
- * only 教 with a place tag after it over the corpus is 教青局, an office.
- */
-const TAUGHT_TAGS = new Set(["r", "n", "nz", "nr", "nt", "ng"]);
-/**
- * What follows a 教 that is teaching where no object does: aspect and the
- * complement marker, which only a verb takes.
- *
- * 他教了三年书, 他教过我英语, 她教书教得很快乐. Matched as characters rather
- * than through their tags, because what the dictionary has starting at a 得 is
- * as likely to be 得很 as the particle itself.
- */
-const ASPECT = new Set(["了", "过", "過", "着", "著", "得"]);
-/**
- * The tag prefix jieba gives a particle, which nothing teaching can follow.
- */
-const PARTICLE_TAG = "u";
-/**
- * Whether the 教 at a position is teaching something to somebody.
- *
- * The object is what says so, which is the same shape the modal 得 rule takes:
- * a 教 with a pronoun, a noun or a name after it is governing it, and a 教 with
- * 了, 过, 着 or 得 after it is a verb whatever follows that. A particle in front
- * of it rules it out, since a verb does not follow one.
- */
-function isTeachingAt(context, at) {
-    if (tagOf(context, wordEndingAt(context, at)).startsWith(PARTICLE_TAG)) {
-        return false;
-    }
-    return (ASPECT.has(context.characters[at + 1] ?? "") ||
-        TAUGHT_TAGS.has(tagOf(context, wordStartingAt(context, at + 1))));
-}
-/**
- * 教 read as `jiāo` where it is teaching rather than a religion.
- *
- * The dictionary stores 教 as `jiào` with `jiāo` as an alternate, and `jiào` is
- * right for the compounds — 教育, 教师, 宗教, 主教 — which are words and reach
- * their reading through the word. What is left is the 教 that stands as a word
- * of its own, and that one is the verb: 他在北京大学教了三年书 came out
- * `jiàole sān nián shū`, and 我教英语, 谁教你法语 and 她教我如何游泳 with it.
- * 教书 is a word and was always `jiāo shū`; nothing carried that across a 了.
- *
- * See {@link isTeachingAt} for what decides it. Forcing the single-character
- * edge is not enough on its own, because a reading spanning two characters
- * carries its own 教 into the position: 王老师教我们汉语 read `jiào` off 师教,
- * and 来教我 off 来教. Those are pairs the dictionary holds with no part of
- * speech and at the cost of a word nothing has ever counted — a reading rather
- * than a word — so a *tagged* word ending in 教 is left alone, which is every
- * one that matters: 任教, 宗教, 主教, 佛教, 传教, 执教, 请教, 家教.
- *
- * Measured over 88,866 lines of Tatoeba and zh.wikipedia, 181 教 decode as a
- * word of their own and every one of them read `jiào`. This moves 158 to
- * `jiāo` and is wrong on three: 统一教创始人, 方法教深思 and 做到了教政分离,
- * where a nominal compound takes an object's shape, the last of them because
- * 做到了 is a word and so the 了 in front is not a particle to look at. Of the
- * 21 it leaves, 8 are right — 诸教中, 教外别传, 董教总, 风教, 教青局 twice, a
- * 教室 reached through a bad split, and a 教 used as a bare noun — and 13 are
- * misses, where jieba tags the object something this does not name.
- */
-export const TEACHING_JIAO = {
-    name: "teaching-jiao",
-    verdictFor: (context) => {
-        const { edge } = context;
-        if (edge.text === "教") {
-            return isJiao(edge.reading[0]) && isTeachingAt(context, edge.from)
-                ? "force"
-                : "keep";
-        }
-        return edge.partOfSpeech === "" &&
-            edge.text.endsWith("教") &&
-            edge.reading.length === edge.to - edge.from &&
-            !isJiao(edge.reading.at(-1)) &&
-            isTeachingAt(context, edge.to - 1)
-            ? "forbid"
-            : "keep";
-    },
-};
 /**
  * The tag the dictionary gives a 量词.
  */
@@ -272,284 +187,8 @@ 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";
-    },
-};
-/**
- * Whether a syllable is `tán`, the verbal reading of 弹.
- */
-function isTan(syllable) {
-    return (syllable?.initial === "t" && syllable.final === "an" && syllable.tone === 2);
-}
-/**
- * The character in both scripts, since 彈 is not the shape 弹 is.
- *
- * 得, 教 and 儿 are the same character in either script and need say it once.
- * This one is not, and neither is the 长 above, which is silent on 長 as a
- * result. {@link ASPECT} carries 過 and 著 for the same reason this carries 彈.
- */
-const PLUCKED = new Set(["弹", "彈"]);
-/**
- * Whether the character at a position is 弹, in either script.
- */
-function isPluckedAt(context, at) {
-    return PLUCKED.has(context.characters[at] ?? "");
-}
-/**
- * What a 弹 that is playing governs: the instrument, the piece or its composer.
- *
- * 弹吉他, 弹竖琴, 弹肖邦, 弹一曲, 给我们弹一手. The name tags are in because
- * jieba calls 吉他 `ns` and 管风琴, 班卓琴 and 古琴 `nr` — instruments read as
- * places and people — and `m` because a piece is as often counted as named.
- */
-const PLAYED_TAGS = new Set(["n", "nz", "nr", "ns", "nt", "ng", "m"]);
-/**
- * Whether a word is an object rather than the far half of a compound.
- *
- * The one place this rule departs from {@link isTeachingAt}, and it is what
- * makes it usable on technical prose. 弹 joins bound morphemes into nouns far
- * more readily than 教 does — 着弹点, 掷弹兵, 供弹爪, 底排弹时, 弹洞 — and every
- * one of those puts a single tagged character after the 弹 where an object
- * would go. What the verb actually governs is a word: 吉他, 竖琴, 管风琴, 电子琴,
- * 一曲, 肖邦. Requiring two characters keeps the compounds out at the cost of
- * 反手持法去弹班卓琴, where the decode reaches 班 rather than 班卓琴 anyway.
- *
- * 教 could not take the same guard, since 教我 and 教你 are the commonest shape
- * it has. Nobody plays a pronoun, so `r` is out of {@link PLAYED_TAGS} too.
- */
-function isPlayedObject(word) {
-    return word !== undefined && toCharacters(word).length > 1;
-}
-/**
- * Whether the 弹 at a position is playing something.
- *
- * Much the same shape as {@link isTeachingAt}, and for the same reason: the
- * object is what says so. A 弹 with an instrument, a tune or a name after it is
- * governing it, and a 弹 with 了, 过, 着 or 得 after it is a verb whatever
- * follows that. A particle in front of it rules it out, since a verb does not
- * follow one.
- */
-function isPlayingAt(context, at) {
-    if (tagOf(context, wordEndingAt(context, at)).startsWith(PARTICLE_TAG)) {
-        return false;
-    }
-    if (ASPECT.has(context.characters[at + 1] ?? "")) {
-        return true;
-    }
-    const played = wordStartingAt(context, at + 1);
-    return isPlayedObject(played) && PLAYED_TAGS.has(tagOf(context, played));
-}
-/**
- * 弹 read as `tán` where it is playing rather than a projectile.
- *
- * The character is stored `dàn` with `tán` as an alternate, and the sources are
- * agreed about it: Unihan's `kHanyuPinlu` counts `dàn(313)` against `tán(50)`
- * and `kMandarin` names `dàn`. That is a fact about a corpus in which nearly
- * every 弹 is ammunition — 子弹, 炮弹, 导弹, 原子弹, 手榴弹 — and every one of
- * those is a word that carries its own reading. What the default is left
- * deciding is the 弹 that stands as a word of its own, and in running text that
- * one is the verb: 他会弹一点儿古筝 came out `dàn`, and 弹吉他, 弹竖琴, 我在弹肖邦
- * and 他钢琴弹得很好 with it.
- *
- * See {@link isPlayingAt} for what decides it. Forcing the single-character
- * edge is not enough on its own, for the reason {@link TEACHING_JIAO} gives: a
- * reading spanning two characters carries its own 弹 into the position, and
- * 我的爱好是开车和弹吉他 read `dàn` off 和弹, a pair the dictionary holds with
- * no part of speech and at the cost of a reading nobody counted. A *tagged*
- * word ending in 弹 is left alone, which is every one that matters — 子弹, 炸弹,
- * 导弹, 反弹, 手榴弹. Nothing has to be done about the other side, since the
- * nominal compounds starting with 弹 are all listed: 弹匣, 弹坑, 弹壳, 弹片,
- * 弹药, 弹道 and 弹头 reach their reading through the word, so a bare 弹 with a
- * noun after it is not the front of one that got away.
- *
- * The 88,866 lines of Tatoeba and zh.wikipedia the other rules were sized
- * against are the wrong corpus for this one on their own: 38 弹 decode as a
- * word of their own in them, and 31 are somebody playing something. So CPP's
- * 20,147 sentences are measured with them as plain text — the benchmark is
- * drawn from zh.wikipedia's military articles, which is precisely where the
- * shapes this rule can break live. Over the 109,013 lines together, 60 弹
- * decode as a word of their own, every one read `dàn`, and 34 of the 60 are
- * wrong.
- *
- * This moves 30 of them, of which 29 are right: the whole of 弹吉他, 弹一曲肖邦
- * and 钢琴弹得很好, and none of 南部弹, 毛瑟弹, 掷弹兵, 供弹爪, 着弹点, 弹洞 or
- * 底排弹时. The one it breaks is 拆弹专家, where a compound has an object's
- * shape on both sides. Five it leaves alone: 开始弹, 四手联弹 and 弹起三次, which
- * have no object to see, 用那种指法弹不会觉得费力, and 反手持法去弹班卓琴, where
- * the decode reaches 班 rather than 班卓琴. On CPP's 40 labelled 弹 nothing
- * moves at all, every one of them having been right through a word.
- */
-export const PLAYING_TAN = {
-    name: "playing-tan",
-    verdictFor: (context) => {
-        const { edge } = context;
-        if (PLUCKED.has(edge.text)) {
-            return isTan(edge.reading[0]) && isPlayingAt(context, edge.from)
-                ? "force"
-                : "keep";
-        }
-        return edge.partOfSpeech === "" &&
-            isPluckedAt(context, edge.to - 1) &&
-            edge.reading.length === edge.to - edge.from &&
-            !isTan(edge.reading.at(-1)) &&
-            isPlayingAt(context, edge.to - 1)
-            ? "forbid"
-            : "keep";
-    },
-};
+import { ADJECTIVAL_CHANG, PLAYING_TAN } from "./polyphone-rules.js";
+export { ADJECTIVAL_CHANG, PLAYING_TAN } from "./polyphone-rules.js";
 /**
  * The rules the lattice decode applies, in order.
  */

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,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,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,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEpC;;GAEG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,OAAO,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC;AACnD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC;AAEtE;;;;;;;;;;;;;GAaG;AACH,SAAS,cAAc,CAAC,IAAwB;IAC9C,OAAO,IAAI,KAAK,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;AAC7D,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,WAAW,CAAC,OAAoB,EAAE,EAAU;IACnD,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,IAAI,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QACjD,OAAO,IAAI,CAAC;IACd,CAAC;IACD,MAAM,MAAM,GAAG,cAAc,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;IAC/C,OAAO,cAAc,CAAC,MAAM,CAAC,IAAI,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC;AAC3E,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AACH,MAAM,CAAC,MAAM,WAAW,GAAa;IACnC,IAAI,EAAE,aAAa;IACnB,UAAU,EAAE,CAAC,OAAoB,EAAE,EAAE;QACnC,MAAM,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC;QACzB,IAAI,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3B,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC;gBAC9D,CAAC,CAAC,OAAO;gBACT,CAAC,CAAC,MAAM,CAAC;QACb,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,KAAK,EAAE;YAC7B,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC;YACjC,IAAI,CAAC,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI;YAC3C,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3B,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC;YACjC,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;IACf,gBAAgB;IAChB,WAAW;CACZ,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,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,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,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,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE,OAAO,EAAE,gBAAgB,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAErE;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAwB;IAChD,QAAQ;IACR,aAAa;IACb,cAAc;IACd,eAAe;IACf,gBAAgB;IAChB,WAAW;CACZ,CAAC"}

package/dist/decode/rules.d.ts

@@ -1,5 +1,6 @@
 import type { Dictionary } from "../dictionary/dictionary.js";
 import type { Lattice, LatticeEdge } from "./lattice.js";
+export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordStartingAt, } from "./edge-context.js";
 /**
  * What a rule can say about one edge.
  *
@@ -33,23 +34,6 @@ export interface EdgeRule {
     readonly name: string;
     readonly verdictFor: (context: EdgeContext) => EdgeVerdict;
 }
-/**
- * The word ending at a position, longest first, or undefined.
- *
- * What a rule needs when it asks about the word *before* an edge. The decode
- * has not run yet, so there is no settled answer; the longest dictionary word
- * ending here is the closest thing available, and it is what the decode will
- * usually choose, since the reading charge prefers fewer words.
- */
-export declare function wordEndingAt(context: EdgeContext, at: number, longest?: number): string | undefined;
-/**
- * The word starting at a position, longest first, or undefined.
- */
-export declare function wordStartingAt(context: EdgeContext, at: number, longest?: number): string | undefined;
-/**
- * The part of speech the dictionary gives a word, or the empty string.
- */
-export declare function tagOf(context: EdgeContext, word: string | undefined): string;
 /**
  * Apply the rules to a lattice, returning the lattice they leave.
  *

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

@@ -1 +1 @@
-{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD;;;;;;;GAOG;AACH,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,UAAU,EAAE,UAAU,CAAC;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC;CAC5B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,WAAW,KAAK,WAAW,CAAC;CAC5D;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAYpB;AAED;;GAEG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,WAAW,EACpB,EAAE,EAAE,MAAM,EACV,OAAO,SAAI,GACV,MAAM,GAAG,SAAS,CAUpB;AAED;;GAEG;AACH,wBAAgB,KAAK,CAAC,OAAO,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,MAAM,CAI5E;AA+DD;;;;;;;;;;;GAWG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,UAAU,EACtB,KAAK,EAAE,SAAS,QAAQ,EAAE,GACzB,OAAO,CAsCT"}
+{"version":3,"file":"rules.d.ts","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEzD,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,mBAAmB,CAAC;AAE3B;;;;;;;GAOG;AACH,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;AAEtD;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,UAAU,EAAE,UAAU,CAAC;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC;CAC5B;AAED;;;;;;;GAOG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,WAAW,KAAK,WAAW,CAAC;CAC5D;AA+DD;;;;;;;;;;;GAWG;AACH,wBAAgB,cAAc,CAC5B,OAAO,EAAE,OAAO,EAChB,UAAU,EAAE,UAAU,EACtB,KAAK,EAAE,SAAS,QAAQ,EAAE,GACzB,OAAO,CAsCT"}

package/dist/decode/rules.js

@@ -1,46 +1,4 @@
-/**
- * The word ending at a position, longest first, or undefined.
- *
- * What a rule needs when it asks about the word *before* an edge. The decode
- * has not run yet, so there is no settled answer; the longest dictionary word
- * ending here is the closest thing available, and it is what the decode will
- * usually choose, since the reading charge prefers fewer words.
- */
-export function wordEndingAt(context, at, longest = 4) {
-    for (let length = longest; length > 0; length--) {
-        const from = at - length;
-        if (from < 0) {
-            continue;
-        }
-        const text = context.characters.slice(from, at).join("");
-        if (context.dictionary.lookup(text) !== undefined) {
-            return text;
-        }
-    }
-    return undefined;
-}
-/**
- * The word starting at a position, longest first, or undefined.
- */
-export function wordStartingAt(context, at, longest = 4) {
-    for (let length = longest; length > 0; length--) {
-        // Nothing starts at the end of a run, which is where a 得 closing a
-        // sentence asks from.
-        const text = context.characters.slice(at, at + length).join("");
-        if (text !== "" && context.dictionary.lookup(text) !== undefined) {
-            return text;
-        }
-    }
-    return undefined;
-}
-/**
- * The part of speech the dictionary gives a word, or the empty string.
- */
-export function tagOf(context, word) {
-    return word === undefined
-        ? ""
-        : (context.dictionary.lookup(word)?.partOfSpeech ?? "");
-}
+export { ASPECT, PARTICLE_TAG, tagOf, wordEndingAt, wordStartingAt, } from "./edge-context.js";
 /**
  * Whether every position can still be left, and the end still reached.
  *

package/dist/decode/rules.js.map

@@ -1 +1 @@
-{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAuCA;;;;;;;GAOG;AACH,MAAM,UAAU,YAAY,CAC1B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,MAAM,IAAI,GAAG,EAAE,GAAG,MAAM,CAAC;QACzB,IAAI,IAAI,GAAG,CAAC,EAAE,CAAC;YACb,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzD,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAClD,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAoB,EACpB,EAAU,EACV,OAAO,GAAG,CAAC;IAEX,KAAK,IAAI,MAAM,GAAG,OAAO,EAAE,MAAM,GAAG,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;QAChD,mEAAmE;QACnE,sBAAsB;QACtB,MAAM,IAAI,GAAG,OAAO,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChE,IAAI,IAAI,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YACjE,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,KAAK,CAAC,OAAoB,EAAE,IAAwB;IAClE,OAAO,IAAI,KAAK,SAAS;QACvB,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,YAAY,IAAI,EAAE,CAAC,CAAC;AAC5D,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAgB;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,EAAE,GAAG,MAAM,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;QACxC,0EAA0E;QAC1E,OAAO,CAAC,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAC1C,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,CACpC,CAAC;IACJ,CAAC;IACD,qDAAqD;IACrD,OAAO,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CACjB,KAA0B,EAC1B,OAAoB;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,OAAO,CAAC;QACjB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,KAA6B,EAC7B,MAAgC,EAChC,SAA2B;IAE3B,MAAM,OAAO,GAAG,IAAI,GAAG,CACrB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAChD,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAAsB,EACtB,KAA0B;IAE1B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAe,CAAC;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAe,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,UAAU,CAAC,KAAK,EAAE;gBAChC,UAAU;gBACV,UAAU,EAAE,OAAO,CAAC,UAAU;gBAC9B,IAAI;aACL,CAAC,CAAC;YACH,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;gBACzB,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtB,CAAC;iBAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IAED,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,KAAK,GAAY;QACrB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACjC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAC7C;KACF,CAAC;IACF,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC;AAChD,CAAC"}
+{"version":3,"file":"rules.js","sourceRoot":"","sources":["../../src/decode/rules.ts"],"names":[],"mappings":"AAGA,OAAO,EACL,MAAM,EACN,YAAY,EACZ,KAAK,EACL,YAAY,EACZ,cAAc,GACf,MAAM,mBAAmB,CAAC;AAsC3B;;;;;;;GAOG;AACH,SAAS,aAAa,CAAC,OAAgB;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC;IACzC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,EAAE,GAAG,MAAM,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;QACxC,0EAA0E;QAC1E,OAAO,CAAC,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAC1C,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,CACpC,CAAC;IACJ,CAAC;IACD,qDAAqD;IACrD,OAAO,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CACjB,KAA0B,EAC1B,OAAoB;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,OAAO,KAAK,MAAM,EAAE,CAAC;YACvB,OAAO,OAAO,CAAC;QACjB,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,YAAY,CACnB,KAA6B,EAC7B,MAAgC,EAChC,SAA2B;IAE3B,MAAM,OAAO,GAAG,IAAI,GAAG,CACrB,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAChD,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;IACH,CAAC;AACH,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,cAAc,CAC5B,OAAgB,EAChB,UAAsB,EACtB,KAA0B;IAE1B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAe,CAAC;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAe,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,OAAO,GAAG,UAAU,CAAC,KAAK,EAAE;gBAChC,UAAU;gBACV,UAAU,EAAE,OAAO,CAAC,UAAU;gBAC9B,IAAI;aACL,CAAC,CAAC;YACH,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;gBACzB,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACtB,CAAC;iBAAM,IAAI,OAAO,KAAK,OAAO,EAAE,CAAC;gBAC/B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,4EAA4E;IAC5E,kEAAkE;IAClE,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK,EAAE,CAAC;QAClC,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,CAAC,CAAC;IACzC,CAAC;IAED,IAAI,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,MAAM,KAAK,GAAY;QACrB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CACjC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAC7C;KACF,CAAC;IACF,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC;AAChD,CAAC"}

package/dist/decode/run-lattice.d.ts

@@ -0,0 +1,52 @@
+import { type Lattice } from "./lattice.js";
+import type { ResolvedHints } from "./hints.js";
+import type { RunLattice } from "./lattice-slice.js";
+export { runLattice } from "./lattice-slice.js";
+import type { Dictionary } from "../dictionary/dictionary.js";
+import type { ReadingUnit } from "./locking.js";
+import { type EdgeRule } from "./rules.js";
+import type { DecodedWord } from "./word.js";
+/**
+ * Gather the reading units into the words the boundaries make of them.
+ *
+ * Where the two decodes disagree the readings win, which is what "advisory"
+ * means for the spacing decode: a boundary falling inside a reading unit is
+ * dropped rather than allowed to break the unit up. 玩儿 read as `wánr` is one
+ * syllable over two characters, and no spacing can be put between them.
+ */
+export declare function groupUnits<Unit extends ReadingUnit>(units: readonly Unit[], boundaries: readonly number[]): readonly (readonly Unit[])[];
+/**
+ * Assemble one word from the units the boundaries kept together.
+ */
+export declare function wordFrom(dictionary: Dictionary, lattice: Lattice, group: readonly ReadingUnit[]): DecodedWord;
+/**
+ * The lattice a run decodes over: stage 1 built, then stage 4 applied.
+ *
+ * The rules run between building and decoding rather than over the output,
+ * which is the whole point of them — they change what the decode gets to choose
+ * between, so everything downstream sees one lattice rather than a decode and
+ * a list of corrections to it.
+ */
+export declare function ruledLattice(dictionary: Dictionary, run: string, rules: readonly EdgeRule[], hints: ResolvedHints | undefined): Lattice;
+/**
+ * Cut a decode of the whole lattice down to the run at the end of it.
+ *
+ * Everything before `at` was context rather than run — 汉字 a caller put in
+ * front so that the decode could see them — and is dropped once it has done its
+ * work.
+ */
+export declare function runGroups<Unit extends ReadingUnit>({ lattice, at }: RunLattice, units: readonly Unit[]): readonly (readonly Unit[])[];
+/**
+ * Decode a Han run with the lattice: the recommended path.
+ *
+ * Stages 1 to 3 of ALGORITHM.md end to end — build every candidate reading,
+ * lock the positions that cannot vary, score the rest — where the greedy
+ * baseline commits to the longest match at each position and never revisits it.
+ *
+ * `before` is 汉字 standing in front of the run that the decode should see and
+ * not report: the 汉字 a number written in digits would have been written with.
+ * A run is decoded on its own without it, and a run reached that way has no idea
+ * what preceded it — 个人 alone is the word `gèrén`, while the 两 of 2个人 makes
+ * 两个 the word and leaves 人 to itself.
+ */
+//# sourceMappingURL=run-lattice.d.ts.map

package/dist/decode/run-lattice.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"run-lattice.d.ts","sourceRoot":"","sources":["../../src/decode/run-lattice.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,KAAK,OAAO,EAAE,MAAM,cAAc,CAAC;AAC1D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAErD,OAAO,EAAE,UAAU,EAAE,MAAM,oBAAoB,CAAC;AAShD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,6BAA6B,CAAC;AAC9D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EAAkB,KAAK,QAAQ,EAAE,MAAM,YAAY,CAAC;AAC3D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAE7C;;;;;;;GAOG;AACH,wBAAgB,UAAU,CAAC,IAAI,SAAS,WAAW,EACjD,KAAK,EAAE,SAAS,IAAI,EAAE,EACtB,UAAU,EAAE,SAAS,MAAM,EAAE,GAC5B,SAAS,CAAC,SAAS,IAAI,EAAE,CAAC,EAAE,CAkB9B;AAED;;GAEG;AACH,wBAAgB,QAAQ,CACtB,UAAU,EAAE,UAAU,EACtB,OAAO,EAAE,OAAO,EAChB,KAAK,EAAE,SAAS,WAAW,EAAE,GAC5B,WAAW,CAiBb;AAED;;;;;;;GAOG;AACH,wBAAgB,YAAY,CAC1B,UAAU,EAAE,UAAU,EACtB,GAAG,EAAE,MAAM,EACX,KAAK,EAAE,SAAS,QAAQ,EAAE,EAC1B,KAAK,EAAE,aAAa,GAAG,SAAS,GAC/B,OAAO,CAMT;AAED;;;;;;GAMG;AACH,wBAAgB,SAAS,CAAC,IAAI,SAAS,WAAW,EAChD,EAAE,OAAO,EAAE,EAAE,EAAE,EAAE,UAAU,EAC3B,KAAK,EAAE,SAAS,IAAI,EAAE,GACrB,SAAS,CAAC,SAAS,IAAI,EAAE,CAAC,EAAE,CAK9B;AAED;;;;;;;;;;;;GAYG"}