chaffjs 0.28.0 → 0.30.0

package/dist/detectors/obligation-without-deadline.js

@@ -0,0 +1,148 @@
+// A duty that happens at a point in time (pay, deliver, return, notify; 支払う, 提出する, 通知する) written with no time at
+// all, in a contract that gives most of its other duties a limit: by when must it be done? The drafter decides whether
+// duties carry limits, so only the minority in one document is pointed at, as vague-deadline does. A sentence holding
+// a vague limit (promptly) is vague-deadline's; one stating an amount (pays a fee of $9,000) sets a price, and its
+// timing is another clause's. The words come from the language package's word lists: timed-duty (the act),
+// duty-marker (shall, will), duty-ending (う, するものとする: what may follow a Japanese act at the sentence end),
+// stated-deadline (by the, までに) with concrete-deadline and vague-deadline, deadline-permission and currency-notation. Pure.
+import { quoteAt } from "./structure-tree.js";
+import { escapeRegExp } from "../orthography.js";
+import { grantsRight, statesConcrete, vagueIn } from "./vague-deadline.js";
+const LATIN = /\p{Script=Latin}/u;
+const LETTER = /\p{L}/u;
+const DIGIT = /\p{Nd}/u;
+const SENTENCE_MARKS = new Set(["。", "．", ".", "!", "?", "！", "？", ";", ":"]);
+/** How many words an aside between the duty marker and the act may hold: "shall, at its option, return". */
+const MARKER_REACH_WORDS = 4;
+/** How far before an amount's trailing mark its number may end (88万円, 1,000 円). */
+const AMOUNT_REACH = 3;
+const withoutEnding = (text) => {
+    const chars = [...text.trimEnd()];
+    const last = chars.findLastIndex((char) => !SENTENCE_MARKS.has(char) && char.trim() !== "");
+    return chars.slice(0, last + 1).join("");
+};
+const latinWords = (lower) => [...lower.matchAll(/[\p{L}']+/gu)].map((match) => ({ word: match[0], at: match.index }));
+/** The act a Latin-script duty binds: an act word within reach after a duty marker (shall return, will notify). */
+const latinActAt = (lower, words) => {
+    const tokens = latinWords(lower);
+    const markerWords = words.markers.map((marker) => marker.toLowerCase().split(/\s+/u));
+    const markerEnds = tokens.flatMap((_, index) => markerWords.filter((parts) => parts.every((part, offset) => tokens[index + offset]?.word === part)).map((parts) => index + parts.length));
+    const acts = new Set(words.acts.map((act) => act.toLowerCase()));
+    const negations = new Set(words.negations.map((word) => word.toLowerCase()));
+    return markerEnds
+        .flatMap((end) => {
+        const reach = tokens.slice(end, end + MARKER_REACH_WORDS);
+        const act = reach.findIndex((token) => acts.has(token.word));
+        const found = reach[act];
+        const marker = tokens[end - 1];
+        if (found === undefined || marker === undefined || reach.slice(0, act).some((token) => negations.has(token.word)))
+            return [];
+        // The act is the verb right after the marker, or after an aside set off by a comma ("shall, at its option, return");
+        // a later act word is an object ("shall send each invoice").
+        const aside = lower
+            .slice(marker.at + marker.word.length, found.at)
+            .trimStart()
+            .startsWith(",");
+        return act === 0 || aside ? [found] : [];
+    })
+        .map((token) => ({ act: token.word, at: token.at }))[0];
+};
+/** The act a duty written as the sentence's last predicate binds (支払う, 通知するものとする, 通知します). */
+const endingActAt = (text, words) => {
+    const ending = withoutEnding(text);
+    const act = words.acts.find((stem) => stem !== "" && words.endings.some((tail) => ending.endsWith(`${stem}${tail}`)));
+    if (act === undefined)
+        return undefined;
+    return { act, at: ending.lastIndexOf(act) };
+};
+const actIn = (text, words) => words.acts.some((act) => LATIN.test(act)) ? latinActAt(text.toLowerCase(), words) : endingActAt(text, words);
+/** An amount (¥1,000, 1,000円, $9,000) in the sentence: it sets what is paid, and when is another clause's. */
+export const statesAmount = (text, currencies) => currencies
+    .filter((entry) => entry.pattern !== "")
+    .some((entry) => {
+    const parts = text.split(entry.pattern);
+    return parts.slice(1).some((after, index) => {
+        const before = parts[index] ?? "";
+        return entry.position === "before" ? DIGIT.test(after.trimStart().charAt(0)) : DIGIT.test(before.trimEnd().slice(-AMOUNT_REACH));
+    });
+});
+const isWordStart = (text, at) => !LETTER.test(text.charAt(at - 1));
+/** The sentence with the words that are no limit blanked out, so that a limit's words inside them are not read. */
+const withoutNotLimits = (text, notLimits) => ({
+    start: text.start,
+    text: notLimits.filter((word) => word !== "").reduce((masked, word) => masked.replaceAll(word, " ".repeat(word.length)), text.text),
+});
+const FRAME = "…";
+/** A phrase in the sentence; a Latin one only as whole words (by the, not by their). */
+const phraseIn = (text, phrase) => {
+    const lower = text.toLowerCase();
+    const wanted = phrase.toLowerCase().trim();
+    if (wanted === "")
+        return false;
+    if (!LATIN.test(wanted))
+        return lower.includes(wanted);
+    return [...lower.matchAll(new RegExp(escapeRegExp(wanted), "gu"))].some((match) => !LETTER.test(lower.charAt(match.index - 1)) && !LETTER.test(lower.charAt(match.index + wanted.length)));
+};
+/** A limit in the sentence: a frame (within … days) or a phrase (by the, までに), outside the words that are no limit. */
+const statesLimit = (text, words) => {
+    const masked = withoutNotLimits(text, words.notLimits);
+    const frames = words.limits.filter((limit) => limit.includes(FRAME));
+    return statesConcrete(masked, frames) || words.limits.some((limit) => !limit.includes(FRAME) && phraseIn(masked.text, limit));
+};
+const limited = (text, words) => statesLimit(text, words) || vagueIn(text, words.vague).length > 0;
+/** For each sentence, the index after the last sentence on its line: the time of a duty often follows it (Fees are due within 30 days). */
+const lineEnds = (texts) => {
+    const ends = [];
+    texts.reduceRight((after, _text, index) => {
+        const end = texts[index + 1]?.sameLine === true ? after : index + 1;
+        ends[index] = end;
+        return end;
+    }, texts.length);
+    return ends;
+};
+/**
+ * Each duty to act at a point in time whose sentence, and the rest of whose line, states no limit, holds no vague one, grants no right and states
+ * no amount, when more sentences of the document state a limit than hold such a duty: these duties are then the
+ * document's minority.
+ */
+export const untimedDuties = (texts, words) => {
+    const ends = lineEnds(texts);
+    const found = texts.flatMap((text, index) => {
+        const line = texts.slice(index, ends[index]);
+        if (line.some((later) => limited(later, words) || words.deferrals.some((deferral) => phraseIn(later.text, deferral))))
+            return [];
+        if (grantsRight(text, words.permissions) || statesAmount(text.text, words.currencies))
+            return [];
+        const act = actIn(text.text, words);
+        if (act === undefined || (LATIN.test(act.act) && !isWordStart(text.text, act.at)))
+            return [];
+        return [{ offset: text.start + act.at, act: text.text.slice(act.at, act.at + act.act.length) }];
+    });
+    const limitedSentences = texts.filter((text) => statesLimit(text, words)).length;
+    return limitedSentences > found.length ? found : [];
+};
+const patternsOf = (doc, lexicon) => (doc.lexicons[lexicon] ?? []).map((entry) => entry.pattern);
+export const obligationWithoutDeadline = (doc) => untimedDuties(doc.sentences.map((sentence, index) => ({
+    start: sentence.span.start,
+    text: sentence.text,
+    sameLine: index > 0 && !doc.source.slice(doc.sentences[index - 1]?.span.end ?? 0, sentence.span.start).includes("\n"),
+})), {
+    acts: patternsOf(doc, "timed-duty"),
+    markers: patternsOf(doc, "duty-marker"),
+    negations: patternsOf(doc, "duty-negation"),
+    endings: patternsOf(doc, "duty-ending"),
+    limits: [...patternsOf(doc, "concrete-deadline"), ...patternsOf(doc, "stated-deadline")],
+    notLimits: patternsOf(doc, "stated-deadline-not"),
+    vague: patternsOf(doc, "vague-deadline"),
+    deferrals: patternsOf(doc, "duty-terms-elsewhere"),
+    permissions: doc.lexicons["deadline-permission"] ?? [],
+    currencies: doc.lexicons["currency-notation"] ?? [],
+}).map((duty) => ({
+    rule: "obligation-without-deadline",
+    severity: "info",
+    line: 0,
+    column: 0,
+    quote: quoteAt(doc.source, duty.offset),
+    values: { word: duty.act, offset: duty.offset },
+}));
+//# sourceMappingURL=obligation-without-deadline.js.map

package/dist/detectors/obligation-without-deadline.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"obligation-without-deadline.js","sourceRoot":"","sources":["../../src/detectors/obligation-without-deadline.ts"],"names":[],"mappings":"AAAA,iHAAiH;AACjH,uHAAuH;AACvH,sHAAsH;AACtH,mHAAmH;AACnH,2GAA2G;AAC3G,2GAA2G;AAC3G,4HAA4H;AAC5H,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAC9C,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,WAAW,EAAE,cAAc,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AA0B3E,MAAM,KAAK,GAAG,mBAAmB,CAAC;AAClC,MAAM,MAAM,GAAG,QAAQ,CAAC;AACxB,MAAM,KAAK,GAAG,SAAS,CAAC;AACxB,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAC9E,4GAA4G;AAC5G,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAC7B,mFAAmF;AACnF,MAAM,YAAY,GAAG,CAAC,CAAC;AAEvB,MAAM,aAAa,GAAG,CAAC,IAAY,EAAU,EAAE;IAC7C,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC;IAClC,MAAM,IAAI,GAAG,KAAK,CAAC,aAAa,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;IAC5F,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC3C,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,KAAa,EAAoD,EAAE,CACrF,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;AAE3F,mHAAmH;AACnH,MAAM,UAAU,GAAG,CAAC,KAAa,EAAE,KAAgB,EAA6D,EAAE;IAChH,MAAM,MAAM,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;IACjC,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IACtF,MAAM,UAAU,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAC7C,WAAW,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,GAAG,MAAM,CAAC,EAAE,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CACzI,CAAC;IACF,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACjE,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IAC7E,OAAO,UAAU;SACd,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QACf,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,kBAAkB,CAAC,CAAC;QAC1D,MAAM,GAAG,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;QAC7D,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC;QACzB,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;QAC/B,IAAI,KAAK,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAAE,OAAO,EAAE,CAAC;QAC7H,qHAAqH;QACrH,6DAA6D;QAC7D,MAAM,KAAK,GAAG,KAAK;aAChB,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,CAAC;aAC/C,SAAS,EAAE;aACX,UAAU,CAAC,GAAG,CAAC,CAAC;QACnB,OAAO,GAAG,KAAK,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC3C,CAAC,CAAC;SACD,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,KAAK,CAAC,IAAI,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC5D,CAAC,CAAC;AAEF,6FAA6F;AAC7F,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,KAAgB,EAA6D,EAAE;IAChH,MAAM,MAAM,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACnC,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,IAAI,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,IAAI,GAAG,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IACtH,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACxC,OAAO,EAAE,GAAG,EAAE,EAAE,EAAE,MAAM,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;AAC9C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,CAAC,IAAY,EAAE,KAAgB,EAA6D,EAAE,CAC1G,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,WAAW,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;AAE/G,8GAA8G;AAC9G,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,UAAmC,EAAW,EAAE,CACzF,UAAU;KACP,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,KAAK,EAAE,CAAC;KACvC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;IACd,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACxC,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC1C,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC;IACnI,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEP,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,EAAU,EAAW,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;AAE7F,mHAAmH;AACnH,MAAM,gBAAgB,GAAG,CAAC,IAAU,EAAE,SAA4B,EAAQ,EAAE,CAAC,CAAC;IAC5E,KAAK,EAAE,IAAI,CAAC,KAAK;IACjB,IAAI,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC;CACpI,CAAC,CAAC;AAEH,MAAM,KAAK,GAAG,GAAG,CAAC;AAElB,wFAAwF;AACxF,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,MAAc,EAAW,EAAE;IACzD,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IACjC,MAAM,MAAM,GAAG,MAAM,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;IAC3C,IAAI,MAAM,KAAK,EAAE;QAAE,OAAO,KAAK,CAAC;IAChC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC;QAAE,OAAO,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACvD,OAAO,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CACrE,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAClH,CAAC;AACJ,CAAC,CAAC;AAEF,uHAAuH;AACvH,MAAM,WAAW,GAAG,CAAC,IAAU,EAAE,KAAgB,EAAW,EAAE;IAC5D,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;IACvD,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;IACrE,OAAO,cAAc,CAAC,MAAM,EAAE,MAAM,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;AAChI,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,IAAU,EAAE,KAAgB,EAAW,EAAE,CAAC,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;AAE7H,2IAA2I;AAC3I,MAAM,QAAQ,GAAG,CAAC,KAAsB,EAAY,EAAE;IACpD,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,KAAK,CAAC,WAAW,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE;QACxC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;QACpE,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC;QAClB,OAAO,GAAG,CAAC;IACb,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IACjB,OAAO,IAAI,CAAC;AACd,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,KAAsB,EAAE,KAAgB,EAAiB,EAAE;IACvF,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC7B,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC1C,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QAC7C,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC;YAAE,OAAO,EAAE,CAAC;QACjI,IAAI,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC,WAAW,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,UAAU,CAAC;YAAE,OAAO,EAAE,CAAC;QACjG,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QACpC,IAAI,GAAG,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;YAAE,OAAO,EAAE,CAAC;QAC7F,OAAO,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,EAAE,EAAE,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,EAAE,GAAG,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAClG,CAAC,CAAC,CAAC;IACH,MAAM,gBAAgB,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC;IACjF,OAAO,gBAAgB,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;AACtD,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,OAAe,EAAY,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAElI,MAAM,CAAC,MAAM,yBAAyB,GAAa,CAAC,GAAG,EAAa,EAAE,CACpE,aAAa,CACX,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;IACtC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK;IAC1B,IAAI,EAAE,QAAQ,CAAC,IAAI;IACnB,QAAQ,EAAE,KAAK,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,IAAI,CAAC,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC;CACtH,CAAC,CAAC,EACH;IACE,IAAI,EAAE,UAAU,CAAC,GAAG,EAAE,YAAY,CAAC;IACnC,OAAO,EAAE,UAAU,CAAC,GAAG,EAAE,aAAa,CAAC;IACvC,SAAS,EAAE,UAAU,CAAC,GAAG,EAAE,eAAe,CAAC;IAC3C,OAAO,EAAE,UAAU,CAAC,GAAG,EAAE,aAAa,CAAC;IACvC,MAAM,EAAE,CAAC,GAAG,UAAU,CAAC,GAAG,EAAE,mBAAmB,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,iBAAiB,CAAC,CAAC;IACxF,SAAS,EAAE,UAAU,CAAC,GAAG,EAAE,qBAAqB,CAAC;IACjD,KAAK,EAAE,UAAU,CAAC,GAAG,EAAE,gBAAgB,CAAC;IACxC,SAAS,EAAE,UAAU,CAAC,GAAG,EAAE,sBAAsB,CAAC;IAClD,WAAW,EAAE,GAAG,CAAC,QAAQ,CAAC,qBAAqB,CAAC,IAAI,EAAE;IACtD,UAAU,EAAE,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,IAAI,EAAE;CACpD,CACF,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACf,IAAI,EAAE,6BAA6B;IACnC,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC;IACvC,MAAM,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE;CAChD,CAAC,CAAC,CAAC"}

package/dist/detectors/open-end-names.d.ts

@@ -0,0 +1,23 @@
+import type { LexiconEntry, ProseDocument, Sentence } from "../plugin.ts";
+/** A quoted name and where the open-end word stands in it. */
+type QuotedName = {
+    readonly name: string;
+    readonly at: readonly number[];
+};
+/**
+ * The names a document gives that contain an open-end word: the ones it quotes, the ones it defines in an aside right
+ * after them (whole, as written there), and the heads of its definition items (which may run on into the definition).
+ */
+export type OpenEndNames = {
+    readonly quoted: readonly QuotedName[];
+    readonly defined: ReadonlySet<string>;
+    readonly heads: readonly string[];
+};
+/** The text inside the aside that opens at offset, or undefined when none opens there or it does not close soon. */
+export declare const asideAt: (source: string, offset: number) => string | undefined;
+/** The names a document gives that contain an open-end word. statements: the language's words of a definition (をいう, means). */
+export declare const openEndNames: (doc: ProseDocument, words: readonly LexiconEntry[], statements: readonly string[]) => OpenEndNames;
+/** Where the word first stands outside a name the document gives (its document offset), or undefined when it never does. */
+export declare const openEndAt: (doc: ProseDocument, sentence: Sentence, word: LexiconEntry, names: OpenEndNames) => number | undefined;
+export {};
+//# sourceMappingURL=open-end-names.d.ts.map

package/dist/detectors/open-end-names.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"open-end-names.d.ts","sourceRoot":"","sources":["../../src/detectors/open-end-names.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,aAAa,EAAE,QAAQ,EAAS,MAAM,cAAc,CAAC;AAoCjF,8DAA8D;AAC9D,KAAK,UAAU,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,EAAE,EAAE,SAAS,MAAM,EAAE,CAAA;CAAE,CAAC;AAE5E;;;GAGG;AACH,MAAM,MAAM,YAAY,GAAG;IAAE,QAAQ,CAAC,MAAM,EAAE,SAAS,UAAU,EAAE,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAA;CAAE,CAAC;AAgChJ,oHAAoH;AACpH,eAAO,MAAM,OAAO,GAAI,QAAQ,MAAM,EAAE,QAAQ,MAAM,KAAG,MAAM,GAAG,SASjE,CAAC;AAyBF,+HAA+H;AAC/H,eAAO,MAAM,YAAY,GAAI,KAAK,aAAa,EAAE,OAAO,SAAS,YAAY,EAAE,EAAE,YAAY,SAAS,MAAM,EAAE,KAAG,YAKhH,CAAC;AAqBF,4HAA4H;AAC5H,eAAO,MAAM,SAAS,GAAI,KAAK,aAAa,EAAE,UAAU,QAAQ,EAAE,MAAM,YAAY,EAAE,OAAO,YAAY,KAAG,MAAM,GAAG,SACjB,CAAC"}

package/dist/detectors/open-end-names.js

@@ -0,0 +1,95 @@
+import { escapeRegExp } from "../orthography.js";
+import { entryRanges } from "./lexicon-match.js";
+// A word that leaves a list open (等, etc.) is also part of names a document gives itself: 「業務執行取締役等」,
+// 銀行等（…をいう。）, the head of a definition item (四の二 親会社等 …をいう。), and longer nouns built on them
+// (監査等委員会設置会社). There the document closes the list, so the word leaves nothing open. Pure: the document and the
+// language's words come in.
+/** The parts of speech a name is made of. */
+const NAME_POS = new Set(["NOUN", "PROPN"]);
+/** A quoted name: what 「」『』“”"" hold, short enough to be a term rather than a sentence. */
+const QUOTED = /[「『“"]([^「」『』“”"\n]{1,40})[」』”"]/gu;
+/** The brackets of an aside that can define the name in front of it: 銀行等（銀行…をいう。）. */
+const ASIDE_CLOSE = { "（": "）", "(": ")" };
+/** How far an aside is read for its closing bracket. A definition's aside is a clause, not a page. */
+const MAX_ASIDE_CHARS = 600;
+/** The label before the head of a definition item, with the space after it (四の二 , (a) ). */
+const ITEM_LABEL = /^\s*\S{1,8}\s+$/u;
+/** A word right after the head that makes it the subject of a sentence (親会社等は), not the head of an item. */
+const FUNCTION_POS = new Set(["ADP", "AUX", "PART", "SCONJ", "CCONJ", "PUNCT"]);
+const attached = (left, right) => left !== undefined && right !== undefined && left.span.end === right.span.start;
+const isNameToken = (token) => token !== undefined && NAME_POS.has(token.pos);
+/** The index of the first noun of the run written right before index, or index itself when none is. */
+const runStart = (tokens, index) => isNameToken(tokens[index - 1]) && attached(tokens[index - 1], tokens[index]) ? runStart(tokens, index - 1) : index;
+/** The index of the last noun of the run written right after index, or index itself when none is. */
+const runEnd = (tokens, index) => isNameToken(tokens[index + 1]) && attached(tokens[index], tokens[index + 1]) ? runEnd(tokens, index + 1) : index;
+const occurrencesIn = (sentence, entry) => {
+    const tokens = sentence.tokens ?? [];
+    return entryRanges(sentence, entry).flatMap((range) => {
+        const firstWord = tokens[range.start];
+        const lastWord = tokens[range.end - 1];
+        if (firstWord === undefined || lastWord === undefined)
+            return [];
+        const first = runStart(tokens, range.start);
+        const last = runEnd(tokens, range.end - 1);
+        return [{ sentence, tokens, range, first, last, start: firstWord.span.start, end: lastWord.span.end }];
+    });
+};
+const startOf = (occurrence, index) => occurrence.tokens[index]?.span.start ?? occurrence.start;
+const endOf = (occurrence, index) => occurrence.tokens[index]?.span.end ?? occurrence.end;
+/** The noun the word is part of, as written (当該監査等委員会). */
+const nounOf = (source, occurrence) => source.slice(startOf(occurrence, occurrence.first), endOf(occurrence, occurrence.last));
+/** The text inside the aside that opens at offset, or undefined when none opens there or it does not close soon. */
+export const asideAt = (source, offset) => {
+    const open = source.charAt(offset);
+    const close = ASIDE_CLOSE[open];
+    if (close === undefined)
+        return undefined;
+    const chars = Array.from(source.slice(offset + 1, offset + 1 + MAX_ASIDE_CHARS));
+    const depths = [];
+    chars.forEach((char) => depths.push((depths.at(-1) ?? 1) + (char === open ? 1 : 0) - (char === close ? 1 : 0)));
+    const end = depths.indexOf(0);
+    return end === -1 ? undefined : chars.slice(0, end).join("");
+};
+const says = (text, statements) => statements.some((word) => text.includes(word));
+/** Whether the noun is defined in the aside right after it (銀行等（銀行…をいう。）). */
+const definedInAside = (source, occurrence, statements) => {
+    const aside = asideAt(source, endOf(occurrence, occurrence.last));
+    return aside !== undefined && says(aside, statements);
+};
+/** Whether the noun heads a definition item: a label before it, no particle right after the word, and a definition in the rest (四の二 親会社等 …をいう。). */
+const headsItem = (source, occurrence, statements) => {
+    const next = occurrence.tokens[occurrence.range.end];
+    if (next !== undefined && next.surface.trim() !== "" && FUNCTION_POS.has(next.pos))
+        return false;
+    const before = source.slice(occurrence.sentence.span.start, startOf(occurrence, occurrence.first));
+    return ITEM_LABEL.test(before) && says(source.slice(endOf(occurrence, occurrence.last), occurrence.sentence.span.end), statements);
+};
+const quotedNames = (source, words) => [...source.matchAll(QUOTED)].flatMap((match) => {
+    const name = match[1] ?? "";
+    const at = words.flatMap((word) => [...name.matchAll(new RegExp(escapeRegExp(word.pattern), "gu"))].map((found) => found.index));
+    return at.length === 0 ? [] : [{ name, at }];
+});
+/** The names a document gives that contain an open-end word. statements: the language's words of a definition (をいう, means). */
+export const openEndNames = (doc, words, statements) => {
+    const occurrences = doc.sentences.flatMap((sentence) => words.flatMap((word) => occurrencesIn(sentence, word)));
+    const defined = occurrences.filter((occurrence) => definedInAside(doc.source, occurrence, statements)).map((occurrence) => nounOf(doc.source, occurrence));
+    const heads = occurrences.filter((occurrence) => headsItem(doc.source, occurrence, statements)).map((occurrence) => nounOf(doc.source, occurrence));
+    return { quoted: quotedNames(doc.source, words), defined: new Set(defined), heads: [...new Set(heads)] };
+};
+/** Whether a quoted name holds this occurrence: the text around it reads the whole name (当該責任追及等 holds 「責任追及等」). */
+const inQuotedName = (source, occurrence, quoted) => quoted.some(({ name, at }) => at.some((index) => source.startsWith(name, occurrence.start - index)));
+/**
+ * The parts of the noun, cut at word boundaries, that hold the word and a noun before it: 当該監査等委員, 監査等委員, 監査等 … for 当該監査等委員.
+ * A part that is a defined name narrows it (当該金銭等, 特定役員等), and the definition closes its list.
+ */
+const partsHoldingWord = (source, occurrence) => {
+    const { tokens, range, first, last } = occurrence;
+    const starts = tokens.slice(first, range.start).map((token) => token.span.start);
+    const ends = tokens.slice(range.end - 1, last + 1).map((token) => token.span.end);
+    return starts.flatMap((start) => ends.map((end) => source.slice(start, end)));
+};
+const isNamed = (source, occurrence, names) => inQuotedName(source, occurrence, names.quoted) ||
+    partsHoldingWord(source, occurrence).some((part) => names.defined.has(part) || names.heads.some((head) => head.startsWith(part)));
+/** Where the word first stands outside a name the document gives (its document offset), or undefined when it never does. */
+export const openEndAt = (doc, sentence, word, names) => occurrencesIn(sentence, word).find((occurrence) => !isNamed(doc.source, occurrence, names))?.start;
+//# sourceMappingURL=open-end-names.js.map

package/dist/detectors/open-end-names.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"open-end-names.js","sourceRoot":"","sources":["../../src/detectors/open-end-names.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,WAAW,EAAmB,MAAM,oBAAoB,CAAC;AAElE,sGAAsG;AACtG,6FAA6F;AAC7F,gHAAgH;AAChH,4BAA4B;AAE5B,6CAA6C;AAC7C,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAC5C,2FAA2F;AAC3F,MAAM,MAAM,GAAG,oCAAoC,CAAC;AACpD,sFAAsF;AACtF,MAAM,WAAW,GAAqC,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC;AAC7E,sGAAsG;AACtG,MAAM,eAAe,GAAG,GAAG,CAAC;AAC5B,4FAA4F;AAC5F,MAAM,UAAU,GAAG,kBAAkB,CAAC;AACtC,4GAA4G;AAC5G,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC;AAyBhF,MAAM,QAAQ,GAAG,CAAC,IAAuB,EAAE,KAAwB,EAAW,EAAE,CAC9E,IAAI,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;AAClF,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,KAAK,SAAS,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;AAE1G,uGAAuG;AACvG,MAAM,QAAQ,GAAG,CAAC,MAAwB,EAAE,KAAa,EAAU,EAAE,CACnE,WAAW,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;AAErH,qGAAqG;AACrG,MAAM,MAAM,GAAG,CAAC,MAAwB,EAAE,KAAa,EAAU,EAAE,CACjE,WAAW,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;AAEnH,MAAM,aAAa,GAAG,CAAC,QAAkB,EAAE,KAAmB,EAAgB,EAAE;IAC9E,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,WAAW,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;QACpD,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;QACvC,IAAI,SAAS,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACjE,MAAM,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QAC5C,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;QAC3C,OAAO,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,SAAS,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;IACzG,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,UAAsB,EAAE,KAAa,EAAU,EAAE,CAAC,UAAU,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,IAAI,CAAC,KAAK,IAAI,UAAU,CAAC,KAAK,CAAC;AAC5H,MAAM,KAAK,GAAG,CAAC,UAAsB,EAAE,KAAa,EAAU,EAAE,CAAC,UAAU,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,IAAI,CAAC,GAAG,IAAI,UAAU,CAAC,GAAG,CAAC;AAEtH,2DAA2D;AAC3D,MAAM,MAAM,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAU,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,UAAU,EAAE,UAAU,CAAC,KAAK,CAAC,EAAE,KAAK,CAAC,UAAU,EAAE,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;AAE3J,oHAAoH;AACpH,MAAM,CAAC,MAAM,OAAO,GAAG,CAAC,MAAc,EAAE,MAAc,EAAsB,EAAE;IAC5E,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;IACnC,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;IAChC,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,CAAC,GAAG,eAAe,CAAC,CAAC,CAAC;IACjF,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAChH,MAAM,GAAG,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;IAC9B,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC/D,CAAC,CAAC;AAEF,MAAM,IAAI,GAAG,CAAC,IAAY,EAAE,UAA6B,EAAW,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;AAEtH,8EAA8E;AAC9E,MAAM,cAAc,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAE,UAA6B,EAAW,EAAE;IACxG,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,KAAK,CAAC,UAAU,EAAE,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;IAClE,OAAO,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC;AACxD,CAAC,CAAC;AAEF,oJAAoJ;AACpJ,MAAM,SAAS,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAE,UAA6B,EAAW,EAAE;IACnG,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,CAAC,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACrD,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACjG,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,OAAO,CAAC,UAAU,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC;IACnG,OAAO,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,UAAU,EAAE,UAAU,CAAC,IAAI,CAAC,EAAE,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,UAAU,CAAC,CAAC;AACrI,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,MAAc,EAAE,KAA8B,EAAgB,EAAE,CACnF,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IAC7C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC5B,MAAM,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;IACjI,OAAO,EAAE,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC;AAC/C,CAAC,CAAC,CAAC;AAEL,+HAA+H;AAC/H,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,GAAkB,EAAE,KAA8B,EAAE,UAA6B,EAAgB,EAAE;IAC9H,MAAM,WAAW,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC;IAChH,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,cAAc,CAAC,GAAG,CAAC,MAAM,EAAE,UAAU,EAAE,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC,CAAC;IAC3J,MAAM,KAAK,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,MAAM,EAAE,UAAU,EAAE,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC,CAAC;IACpJ,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,OAAO,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,EAAE,KAAK,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;AAC3G,CAAC,CAAC;AAEF,oHAAoH;AACpH,MAAM,YAAY,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAE,MAA6B,EAAW,EAAE,CACtG,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,UAAU,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;AAEvG;;;GAGG;AACH,MAAM,gBAAgB,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAY,EAAE;IAC5E,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,UAAU,CAAC;IAClD,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACjF,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAClF,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;AAChF,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,MAAc,EAAE,UAAsB,EAAE,KAAmB,EAAW,EAAE,CACvF,YAAY,CAAC,MAAM,EAAE,UAAU,EAAE,KAAK,CAAC,MAAM,CAAC;IAC9C,gBAAgB,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAEpI,4HAA4H;AAC5H,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,GAAkB,EAAE,QAAkB,EAAE,IAAkB,EAAE,KAAmB,EAAsB,EAAE,CAC/H,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC"}

package/dist/detectors/ordinal-run-gaps.d.ts

@@ -0,0 +1,8 @@
+import type { ProseDocument } from "../plugin.ts";
+import type { StructureIssue } from "../structure/issues.ts";
+/**
+ * The skips and repeats in the document's 第N回 / Week N runs. A block whose heading or the lines right above it say it
+ * lists only some (抜粋, selected) is not read: its numbers skip by choice.
+ */
+export declare const ordinalRunGaps: (doc: ProseDocument) => StructureIssue[];
+//# sourceMappingURL=ordinal-run-gaps.d.ts.map

package/dist/detectors/ordinal-run-gaps.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ordinal-run-gaps.d.ts","sourceRoot":"","sources":["../../src/detectors/ordinal-run-gaps.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,aAAa,EAAW,MAAM,cAAc,CAAC;AAG3D,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AAkF7D;;;GAGG;AACH,eAAO,MAAM,cAAc,GAAI,KAAK,aAAa,KAAG,cAAc,EASjE,CAAC"}

package/dist/detectors/ordinal-run-gaps.js

@@ -0,0 +1,80 @@
+import { proseAndTablesOf } from "../table-text.js";
+import { runsOf } from "../structure/runs.js";
+import { framesAt, ordinalFramesOf, ordinalRunBreaks, ordinalRuns } from "../structure/ordinal-runs.js";
+const EMPHASIS = String.raw `(?:\*\*|__|\*|_)?`;
+const TABLE_LEAD = new RegExp(String.raw `^[ \t]*\|?[ \t]*${EMPHASIS}`, "u");
+const LIST_LEAD = new RegExp(String.raw `^[ \t]*(?:[-*+]|\d{1,9}[.)])[ \t]+(?:\[[ xX]\][ \t]+)?${EMPHASIS}`, "u");
+const LINE_LEAD = new RegExp(String.raw `^[ \t\u3000]*${EMPHASIS}`, "u");
+const QUOTED_LINE = /^[ \t]*>/u;
+const linesOf = (text) => {
+    const starts = [0, ...[...text.matchAll(/\n/gu)].map((match) => match.index + 1)];
+    return starts.map((start, index) => ({ start, text: text.slice(start, (starts[index + 1] ?? text.length + 1) - 1) }));
+};
+const lineAt = (text, offset) => {
+    const end = text.indexOf("\n", offset);
+    return text.slice(offset, end === -1 ? text.length : end);
+};
+const leadOf = (text, offset, marks, frames) => {
+    const line = lineAt(text, offset);
+    const skipped = marks.exec(line)?.[0].length ?? 0;
+    return { offset: offset + skipped, read: framesAt(line.slice(skipped), frames) };
+};
+const tableBlocks = (text, frames) => runsOf(text)
+    .filter((run) => run[0]?.kind === "table" && run.length > 0)
+    .map((run) => ({ leads: run.map((line) => leadOf(text, line.start, TABLE_LEAD, frames)), start: run[0]?.start ?? 0 }));
+const isQuotedAt = (source, offset) => QUOTED_LINE.test(lineAt(source, source.lastIndexOf("\n", offset - 1) + 1));
+const listBlocks = (doc, text, frames) => doc.lists
+    .filter((list) => !isQuotedAt(doc.source, list.span.start))
+    .map((list) => ({ leads: list.itemSpans.map((item) => leadOf(text, item.start, LIST_LEAD, frames)), start: list.span.start }));
+const sectionIndexAt = (sections, offset) => sections.findLastIndex((section) => section.span.start <= offset);
+/** The lines outside tables, lists and quotations, one block per section. */
+const lineBlocks = (doc, text, frames) => {
+    const inList = (offset) => doc.lists.some((list) => offset >= list.span.start && offset < list.span.end);
+    const bySection = new Map();
+    linesOf(text)
+        .filter((line) => !line.text.includes("|"))
+        .map((line) => leadOf(text, line.start, LINE_LEAD, frames))
+        .filter((lead) => lead.read.length > 0 && !inList(lead.offset) && !isQuotedAt(doc.source, lead.offset))
+        .forEach((lead) => {
+        const key = sectionIndexAt(doc.sections, lead.offset);
+        const leads = bySection.get(key) ?? [];
+        leads.push(lead);
+        bySection.set(key, leads);
+    });
+    return [...bySection.values()].map((leads) => ({ leads, start: leads[0]?.offset ?? 0 }));
+};
+/** How many lines above a block are read for a word saying it lists only some items (a caption, a table's header row). */
+const LINES_ABOVE = 3;
+const LATIN_WORD = /^[A-Za-z][A-Za-z ]*$/u;
+const markerTest = (marker) => {
+    if (!LATIN_WORD.test(marker))
+        return (text) => text.includes(marker);
+    const word = new RegExp(String.raw `(?<![A-Za-z])${marker}(?![A-Za-z])`, "iu");
+    return (text) => word.test(text);
+};
+/** The heading over the block and the few lines right above it, as written. */
+const contextAbove = (doc, start) => {
+    const section = doc.sections[sectionIndexAt(doc.sections, start)];
+    const from = section?.span.start ?? 0;
+    const above = linesOf(doc.source.slice(from, start))
+        .map((line) => line.text)
+        .filter((line) => line.trim() !== "")
+        .slice(-LINES_ABOVE);
+    return [section?.heading ?? "", ...above].join("\n");
+};
+/**
+ * The skips and repeats in the document's 第N回 / Week N runs. A block whose heading or the lines right above it say it
+ * lists only some (抜粋, selected) is not read: its numbers skip by choice.
+ */
+export const ordinalRunGaps = (doc) => {
+    const frames = ordinalFramesOf(doc.lexicons["ordinal-frame"] ?? []);
+    if (frames.length === 0)
+        return [];
+    const partial = (doc.lexicons["partial-listing-marker"] ?? []).map((entry) => markerTest(entry.pattern));
+    const text = proseAndTablesOf(doc);
+    return [...tableBlocks(text, frames), ...listBlocks(doc, text, frames), ...lineBlocks(doc, text, frames)]
+        .filter((block) => block.leads.some((lead) => lead.read.length > 0))
+        .filter((block) => !partial.some((names) => names(contextAbove(doc, block.start))))
+        .flatMap((block) => ordinalRuns(block.leads).flatMap(ordinalRunBreaks));
+};
+//# sourceMappingURL=ordinal-run-gaps.js.map

package/dist/detectors/ordinal-run-gaps.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ordinal-run-gaps.js","sourceRoot":"","sources":["../../src/detectors/ordinal-run-gaps.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAC;AACpD,OAAO,EAAE,MAAM,EAAE,MAAM,sBAAsB,CAAC;AAE9C,OAAO,EAAE,QAAQ,EAAE,eAAe,EAAE,gBAAgB,EAAE,WAAW,EAAgC,MAAM,8BAA8B,CAAC;AAOtI,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,mBAAmB,CAAC;AAC/C,MAAM,UAAU,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,mBAAmB,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAC;AAC5E,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,yDAAyD,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAC;AACjH,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,gBAAgB,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAC;AACxE,MAAM,WAAW,GAAG,WAAW,CAAC;AAEhC,MAAM,OAAO,GAAG,CAAC,IAAY,EAAU,EAAE;IACvC,MAAM,MAAM,GAAG,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;IAClF,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AACxH,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAc,EAAU,EAAE;IACtD,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACvC,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;AAC5D,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAc,EAAE,KAAa,EAAE,MAA+B,EAAQ,EAAE;IACpG,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IAClC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC;IAClD,OAAO,EAAE,MAAM,EAAE,MAAM,GAAG,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC,EAAE,CAAC;AACnF,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,MAA+B,EAAW,EAAE,CAC7E,MAAM,CAAC,IAAI,CAAC;KACT,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,IAAI,KAAK,OAAO,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC;KAC3D,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,UAAU,EAAE,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;AAE3H,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,MAAc,EAAW,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AAE3I,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,IAAY,EAAE,MAA+B,EAAW,EAAE,CAChG,GAAG,CAAC,KAAK;KACN,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;KAC1D,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;AAEnI,MAAM,cAAc,GAAG,CAAC,QAA4B,EAAE,MAAc,EAAU,EAAE,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC,CAAC;AAEnJ,6EAA6E;AAC7E,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,IAAY,EAAE,MAA+B,EAAW,EAAE;IAChG,MAAM,MAAM,GAAG,CAAC,MAAc,EAAW,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC1H,MAAM,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC5C,OAAO,CAAC,IAAI,CAAC;SACV,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;SAC1C,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC;SAC1D,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;SACtG,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAChB,MAAM,GAAG,GAAG,cAAc,CAAC,GAAG,CAAC,QAAQ,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;QACtD,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC;QACvC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACjB,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;IAC5B,CAAC,CAAC,CAAC;IACL,OAAO,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;AAC3F,CAAC,CAAC;AAEF,0HAA0H;AAC1H,MAAM,WAAW,GAAG,CAAC,CAAC;AACtB,MAAM,UAAU,GAAG,uBAAuB,CAAC;AAE3C,MAAM,UAAU,GAAG,CAAC,MAAc,EAA+B,EAAE;IACjE,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC;QAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACrE,MAAM,IAAI,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,gBAAgB,MAAM,cAAc,EAAE,IAAI,CAAC,CAAC;IAC9E,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACnC,CAAC,CAAC;AAEF,+EAA+E;AAC/E,MAAM,YAAY,GAAG,CAAC,GAAkB,EAAE,KAAa,EAAU,EAAE;IACjE,MAAM,OAAO,GAAG,GAAG,CAAC,QAAQ,CAAC,cAAc,CAAC,GAAG,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,CAAC;IAClE,MAAM,IAAI,GAAG,OAAO,EAAE,IAAI,CAAC,KAAK,IAAI,CAAC,CAAC;IACtC,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;SACjD,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SACxB,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC;SACpC,KAAK,CAAC,CAAC,WAAW,CAAC,CAAC;IACvB,OAAO,CAAC,OAAO,EAAE,OAAO,IAAI,EAAE,EAAE,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACvD,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,GAAkB,EAAoB,EAAE;IACrE,MAAM,MAAM,GAAG,eAAe,CAAC,GAAG,CAAC,QAAQ,CAAC,eAAe,CAAC,IAAI,EAAE,CAAC,CAAC;IACpE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACnC,MAAM,OAAO,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,wBAAwB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;IACzG,MAAM,IAAI,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IACnC,OAAO,CAAC,GAAG,WAAW,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC;SACtG,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;SACnE,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,YAAY,CAAC,GAAG,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;SAClF,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,WAAW,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC,CAAC;AAC5E,CAAC,CAAC"}

package/dist/detectors/orthographic-variant.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"orthographic-variant.d.ts","sourceRoot":"","sources":["../../src/detectors/orthographic-variant.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA2C,MAAM,cAAc,CAAC;AAqLtF;;;;GAIG;AACH,eAAO,MAAM,mBAAmB,EAAE,QAUjC,CAAC"}
+{"version":3,"file":"orthographic-variant.d.ts","sourceRoot":"","sources":["../../src/detectors/orthographic-variant.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA2C,MAAM,cAAc,CAAC;AA8MtF;;;;GAIG;AACH,eAAO,MAAM,mBAAmB,EAAE,QAUjC,CAAC"}

package/dist/detectors/orthographic-variant.js

@@ -1,10 +1,11 @@
 import { groupBy, oddSpellings } from "../spelling-variants.js";
-import { kanjiSkeleton, katakanaKey, lemmaReading } from "../kana-spelling.js";
+import { dropsOkurigana, kanjiSkeleton, katakanaKey, lemmaReading } from "../kana-spelling.js";
 import { isKatakanaWord, stemOf } from "../long-vowel.js";
 import { QUOTATION_MARKS, isWithinAny, quotedSpans } from "../quoted-span.js";
 import { nameSpans, touchesAny } from "../team-names.js";
 import { furiganaSpans } from "../furigana.js";
 import { acronymsIn, isCapitalsNotSpelling } from "../capitals-with-small.js";
+import { isPartOfAddress } from "./address-word.js";
 /** Parts of speech that carry a word of their own. Particles and endings are written one way by the grammar. */
 const CONTENT = new Set(["NOUN", "VERB", "ADJ", "ADV", "PRON", "CCONJ", "DET"]);
 const JAPANESE_WORD = /^[\p{Script=Han}々ぁ-ゖァ-ヺー]+$/u;
@@ -33,7 +34,15 @@ const TE_FORM_CONJUNCTION = /^\p{Script=Han}.*[てで]$/u;
 /** After a て-form (見てみる). The tokenizer reads the て as a joining word, or the whole て-form as a conjunction (追って / みる). */
 const followsTeForm = (previous) => (previous?.pos === "SCONJ" && (previous.surface === "て" || previous.surface === "で")) ||
     (previous?.pos === "CCONJ" && TE_FORM_CONJUNCTION.test(previous.surface));
-const isHelper = ({ token, previous }) => token.features?.["Bound"] === "Yes" || followsTeForm(previous);
+/** は or も after a word ending in で, て or く: the ない after them negates (事実ではない, どれでもない, 高くはない); it is not 無い. */
+const TOPIC_AFTER_LINK = new Set(["は", "も"]);
+const LINK_END = /[でてく]$/u;
+const NEGATION = new Set(["ない", "無い"]);
+/** では and でも read as one word after a space (2 ではなく). */
+const LINK_AND_TOPIC = new Set(["では", "でも"]);
+const followsLinkAndTopic = (previous, before) => LINK_AND_TOPIC.has(previous?.surface ?? "") || (previous?.pos === "ADP" && TOPIC_AFTER_LINK.has(previous.surface) && LINK_END.test(before?.surface ?? ""));
+const negatesAfterTopic = ({ token, previous, before }) => token.pos === "ADJ" && NEGATION.has(token.lemma ?? "") && followsLinkAndTopic(previous, before);
+const isHelper = (word) => word.token.features?.["Bound"] === "Yes" || followsTeForm(word.previous) || negatesAfterTopic(word);
 const isTouching = (left, right) => left !== undefined && left.span.end === right.span.start;
 const HIRAGANA_ONLY = /^[ぁ-ゖー]+$/u;
 /** A kana noun written onto a noun (全員ぶん): a suffix, which may not be the free noun of the same reading (文). */
@@ -59,13 +68,18 @@ const ONE_KATAKANA_HEAD = /^\p{Script=Katakana}(?!\p{Script=Katakana})/u;
 const KATAKANA_END = /\p{Script=Katakana}$/u;
 /** A piece the tokenizer cut out of a katakana run (ミスっ read as ミ and スる): not a word of its own. */
 const isCutFromKatakana = ({ token, previous }) => ONE_KATAKANA_HEAD.test(token.surface) && token.surface.length > 1 && isTouching(previous, token) && KATAKANA_END.test(previous?.surface ?? "");
+/**
+ * で read as the verb 出る right after a word with no particle between (JSON でない): the copula's で, which the tokenizer
+ * misreads after a space or a name. 出る after a particle (結果がでない) or at the head of a sentence (でない音) is the verb.
+ */
+const isCopulaReadAsVerb = ({ token, previous }) => token.pos === "VERB" && token.surface === "で" && previous !== undefined && previous.pos !== "ADP";
 const readingKeyOf = (word, { skip, idioms }) => {
     const { token } = word;
     const lemma = token.lemma;
     if (!CONTENT.has(token.pos) || token.reading === undefined || lemma === undefined || !JAPANESE_WORD.test(lemma) || skip.has(lemma))
         return undefined;
     const reading = lemmaReading(token.surface, lemma, token.reading);
-    if (reading === undefined || [...reading].length < MIN_READING || skip.has(reading) || isCutFromKatakana(word))
+    if (reading === undefined || [...reading].length < MIN_READING || skip.has(reading) || isCutFromKatakana(word) || isCopulaReadAsVerb(word))
         return undefined;
     return `${token.pos}|${useOf(word, idioms)}|${reading}`;
 };
@@ -77,21 +91,27 @@ const readingWords = (sentence, isOpen, skips) => {
         const local = token.span.start - sentence.span.start;
         if (key === undefined || token.lemma === undefined || !isOpen(local, local + token.surface.length))
             return [];
-        return [{ key, spelling: token.lemma, sentence, offset: token.span.start }];
+        const next = tokens[at + 1];
+        const compoundHead = token.pos === "NOUN" && next?.pos === "NOUN" && isTouching(token, next);
+        return [{ key, spelling: token.lemma, sentence, offset: token.span.start, compoundHead }];
     });
 };
 /**
  * Same reading is not yet same word: 書く and 描く are both カク. Spellings meet only when they keep the same kanji (引っ越し, 引越し)
  * or one is all kana. A kana spelling joins the kanji one only when the reading has a single kanji spelling in the document;
- * with two (橋 and 箸 beside はし) it cannot say which, and is left out.
+ * with two (橋 and 箸 beside はし) it cannot say which, and is left out. Official Japanese drops okurigana at the head of a compound
+ * (取扱事業者 beside 個人情報の取扱い), so where the document writes a word both ways, its compound heads are compared apart.
  */
 const byKanji = (words) => [...groupBy(words, (word) => word.key).values()].flatMap((group) => {
     const skeletons = [...new Set(group.map((word) => kanjiSkeleton(word.spelling)).filter((skeleton) => skeleton !== ""))];
+    const spellings = [...new Set(group.map((word) => word.spelling))];
+    const apart = spellings.some((spelling) => dropsOkurigana(spelling, spellings)) ? "|compound" : "";
     return group.flatMap((word) => {
         const own = kanjiSkeleton(word.spelling);
         if (own === "" && skeletons.length > 1)
             return [];
-        return [{ ...word, key: `${word.key}|${own === "" ? (skeletons[0] ?? "") : own}` }];
+        const use = word.compoundHead === true ? apart : "";
+        return [{ ...word, key: `${word.key}|${own === "" ? (skeletons[0] ?? "") : own}${use}` }];
     });
 });
 /** Katakana nouns keyed with small kana and ー evened out, spelled without the final ー that katakana-long-vowel reads. */
@@ -112,9 +132,6 @@ const latinSpelling = (word) => word
     .split("-")
     .map((part) => part.charAt(0).toLowerCase() + part.slice(1))
     .join("-");
-/** A word joined to a dot, an at sign or a slash is part of an address or a file name (github.com, user@example), spelled as it must be. */
-const ADDRESS_NEIGHBOUR = /[.@/\\_]/u;
-const isPartOfAddress = (text, start, end) => ADDRESS_NEIGHBOUR.test(text.charAt(start - 1)) || (ADDRESS_NEIGHBOUR.test(text.charAt(end)) && /\w/u.test(text.charAt(end + 1)));
 /**
  * Latin words keyed without hyphens and case (e-mail and email, GitHub and Github). A word in capitals only (TEAMS, MAY) is
  * emphasis or an acronym, not a spelling of the word, and is left out; so are a name in capitals with an ending (SENDs) and

package/dist/detectors/orthographic-variant.js.map

@@ -1 +1 @@
-{"version":3,"file":"orthographic-variant.js","sourceRoot":"","sources":["../../src/detectors/orthographic-variant.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAoC,MAAM,yBAAyB,CAAC;AAClG,OAAO,EAAE,aAAa,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/E,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,eAAe,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAC9E,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACzD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,UAAU,EAAE,qBAAqB,EAAE,MAAM,2BAA2B,CAAC;AAO9E,gHAAgH;AAChH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;AAChF,MAAM,aAAa,GAAG,8BAA8B,CAAC;AACrD,sFAAsF;AACtF,MAAM,WAAW,GAAG,CAAC,CAAC;AACtB,uEAAuE;AACvE,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAC3B,MAAM,aAAa,GAAG,CAAC,CAAC;AAExB,mJAAmJ;AACnJ,MAAM,QAAQ,GAAG,+BAA+B,CAAC;AACjD,MAAM,OAAO,GAAG,kBAAkB,CAAC;AACnC,6HAA6H;AAC7H,MAAM,KAAK,GAAG,4BAA4B,CAAC;AAE3C,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,EAAU,EAAY,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAExH;;;GAGG;AACH,MAAM,QAAQ,GAAG,CAAC,QAAkB,EAAE,KAAwB,EAAU,EAAE;IACxE,MAAM,MAAM,GAAG,CAAC,GAAG,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,eAAe,CAAC,EAAE,GAAG,aAAa,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;IACjG,MAAM,KAAK,GAAG,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC9C,OAAO,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC;AAChG,CAAC,CAAC;AAKF,qJAAqJ;AACrJ,MAAM,mBAAmB,GAAG,yBAAyB,CAAC;AAEtD,2HAA2H;AAC3H,MAAM,aAAa,GAAG,CAAC,QAA2B,EAAW,EAAE,CAC7D,CAAC,QAAQ,EAAE,GAAG,KAAK,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,OAAO,KAAK,GAAG,CAAC,CAAC;IACrF,CAAC,QAAQ,EAAE,GAAG,KAAK,OAAO,IAAI,mBAAmB,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;AAE5E,MAAM,QAAQ,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CAAC,KAAK,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,IAAI,aAAa,CAAC,QAAQ,CAAC,CAAC;AAE7H,MAAM,UAAU,GAAG,CAAC,IAAuB,EAAE,KAAY,EAAW,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;AAEhI,MAAM,aAAa,GAAG,YAAY,CAAC;AAEnC,gHAAgH;AAChH,MAAM,YAAY,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CAC/D,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,QAAQ,EAAE,GAAG,KAAK,MAAM,IAAI,UAAU,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEvH,oHAAoH;AACpH,MAAM,WAAW,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAa,EAAE,MAA2B,EAAW,EAAE,CACnG,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,QAAQ,EAAE,GAAG,KAAK,KAAK,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,OAAO,GAAG,QAAQ,CAAC,OAAO,GAAG,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;AAE/J;;;;GAIG;AACH,MAAM,KAAK,GAAG,CAAC,IAAe,EAAE,MAA2B,EAAU,EAAE;IACrE,IAAI,WAAW,CAAC,IAAI,EAAE,MAAM,CAAC;QAAE,OAAO,OAAO,CAAC;IAC9C,IAAI,YAAY,CAAC,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC;IACxC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC;QAAE,OAAO,MAAM,CAAC;IACnC,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,WAAW,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;AAC7G,CAAC,CAAC;AAEF,uHAAuH;AACvH,MAAM,iBAAiB,GAAG,8CAA8C,CAAC;AACzE,MAAM,YAAY,GAAG,uBAAuB,CAAC;AAE7C,qGAAqG;AACrG,MAAM,iBAAiB,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CACpE,iBAAiB,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,UAAU,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,QAAQ,EAAE,OAAO,IAAI,EAAE,CAAC,CAAC;AAIjJ,MAAM,YAAY,GAAG,CAAC,IAAe,EAAE,EAAE,IAAI,EAAE,MAAM,EAAS,EAAsB,EAAE;IACpF,MAAM,EAAE,KAAK,EAAE,GAAG,IAAI,CAAC;IACvB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;IAC1B,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACrJ,MAAM,OAAO,GAAG,YAAY,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IAClE,IAAI,OAAO,KAAK,SAAS,IAAI,CAAC,GAAG,OAAO,CAAC,CAAC,MAAM,GAAG,WAAW,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,iBAAiB,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IACjI,OAAO,GAAG,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;AAC1D,CAAC,CAAC;AAEF,0HAA0H;AAC1H,MAAM,YAAY,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAE,KAAY,EAAY,EAAE;IAClF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,EAAY,EAAE;QAC5C,MAAM,GAAG,GAAG,YAAY,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;QAC7F,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;QACrD,IAAI,GAAG,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;YAAE,OAAO,EAAE,CAAC;QAC9G,OAAO,CAAC,EAAE,GAAG,EAAE,QAAQ,EAAE,KAAK,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;IAC9E,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,OAAO,GAAG,CAAC,KAAwB,EAAY,EAAE,CACrD,CAAC,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IACjE,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC;IACxH,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAY,EAAE;QACtC,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACzC,IAAI,GAAG,KAAK,EAAE,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,EAAE,CAAC;QAClD,OAAO,CAAC,EAAE,GAAG,IAAI,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,IAAI,GAAG,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;IACtF,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEL,yHAAyH;AACzH,MAAM,aAAa,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAY,EAAE,CACrE,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAY,EAAE;IAClD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;IACrD,IAAI,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,CAAC,cAAc,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;QAAE,OAAO,EAAE,CAAC;IACtH,MAAM,GAAG,GAAG,WAAW,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,QAAQ,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;AACjJ,CAAC,CAAC,CAAC;AAEL,MAAM,UAAU,GAAG,0CAA0C,CAAC;AAC9D,MAAM,KAAK,GAAG,SAAS,CAAC;AAExB;;;GAGG;AACH,MAAM,aAAa,GAAG,CAAC,IAAY,EAAU,EAAE,CAC7C,IAAI;KACD,KAAK,CAAC,GAAG,CAAC;KACV,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KAC3D,IAAI,CAAC,GAAG,CAAC,CAAC;AAEf,4IAA4I;AAC5I,MAAM,iBAAiB,GAAG,WAAW,CAAC;AAEtC,MAAM,eAAe,GAAG,CAAC,IAAY,EAAE,KAAa,EAAE,GAAW,EAAW,EAAE,CAC5E,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AAEnI;;;;GAIG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAE,IAAyB,EAAY,EAAE;IAC7F,MAAM,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,OAAO,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAY,EAAE;QACzE,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACtB,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACrD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC;QACtC,IAAI,KAAK,CAAC,MAAM,GAAG,aAAa,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,qBAAqB,CAAC,IAAI,EAAE,QAAQ,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;YAAE,OAAO,EAAE,CAAC;QAC1I,IAAI,eAAe,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC;YAAE,OAAO,EAAE,CAAC;QAC7F,OAAO,CAAC,EAAE,GAAG,EAAE,SAAS,KAAK,EAAE,EAAE,QAAQ,EAAE,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;IACtI,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,yFAAyF;AACzF,MAAM,UAAU,GAAG,CAAC,KAAwB,EAAE,EAAE,IAAI,EAAE,KAAK,EAAuB,EAAU,EAAE,CAC5F,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,IAAI,CAAC,GAAG,IAAI,KAAK,CAAC,QAAQ,KAAK,KAAK,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC;AAE5F,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAE,GAAwB,EAAW,EAAE,CAAC,CAAC;IAClF,IAAI,EAAE,EAAE;IACR,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IACpC,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,EAAE,KAAK,EAAE,UAAU,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,EAAE,EAAE,GAAG,CAAC,EAAE,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE;CAC/I,CAAC,CAAC;AAEH;;;;GAIG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACvE,MAAM,KAAK,GAAU,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,CAAC,GAAG,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,EAAE,CAAC;IAC7I,MAAM,MAAM,GAAG,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;IAC5G,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,YAAY,CAAC,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;IACzG,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC;IAC3F,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IACjG,MAAM,GAAG,GAAG,CAAC,OAAO,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;IAC9F,MAAM,GAAG,GAAG,CAAC,GAAG,OAAO,EAAE,GAAG,QAAQ,EAAE,GAAG,KAAK,CAAC,CAAC;IAChD,4FAA4F;IAC5F,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;AACrH,CAAC,CAAC"}
+{"version":3,"file":"orthographic-variant.js","sourceRoot":"","sources":["../../src/detectors/orthographic-variant.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAoC,MAAM,yBAAyB,CAAC;AAClG,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/F,OAAO,EAAE,cAAc,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,eAAe,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAC9E,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACzD,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,UAAU,EAAE,qBAAqB,EAAE,MAAM,2BAA2B,CAAC;AAC9E,OAAO,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AAUpD,gHAAgH;AAChH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;AAChF,MAAM,aAAa,GAAG,8BAA8B,CAAC;AACrD,sFAAsF;AACtF,MAAM,WAAW,GAAG,CAAC,CAAC;AACtB,uEAAuE;AACvE,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAC3B,MAAM,aAAa,GAAG,CAAC,CAAC;AAExB,mJAAmJ;AACnJ,MAAM,QAAQ,GAAG,+BAA+B,CAAC;AACjD,MAAM,OAAO,GAAG,kBAAkB,CAAC;AACnC,6HAA6H;AAC7H,MAAM,KAAK,GAAG,4BAA4B,CAAC;AAE3C,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,EAAU,EAAY,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAExH;;;GAGG;AACH,MAAM,QAAQ,GAAG,CAAC,QAAkB,EAAE,KAAwB,EAAU,EAAE;IACxE,MAAM,MAAM,GAAG,CAAC,GAAG,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,eAAe,CAAC,EAAE,GAAG,aAAa,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;IACjG,MAAM,KAAK,GAAG,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC9C,OAAO,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC;AAChG,CAAC,CAAC;AAKF,qJAAqJ;AACrJ,MAAM,mBAAmB,GAAG,yBAAyB,CAAC;AAEtD,2HAA2H;AAC3H,MAAM,aAAa,GAAG,CAAC,QAA2B,EAAW,EAAE,CAC7D,CAAC,QAAQ,EAAE,GAAG,KAAK,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,KAAK,GAAG,IAAI,QAAQ,CAAC,OAAO,KAAK,GAAG,CAAC,CAAC;IACrF,CAAC,QAAQ,EAAE,GAAG,KAAK,OAAO,IAAI,mBAAmB,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;AAE5E,gHAAgH;AAChH,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAC7C,MAAM,QAAQ,GAAG,SAAS,CAAC;AAC3B,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAEvC,yDAAyD;AACzD,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAE7C,MAAM,mBAAmB,GAAG,CAAC,QAA2B,EAAE,MAAyB,EAAW,EAAE,CAC9F,cAAc,CAAC,GAAG,CAAC,QAAQ,EAAE,OAAO,IAAI,EAAE,CAAC,IAAI,CAAC,QAAQ,EAAE,GAAG,KAAK,KAAK,IAAI,gBAAgB,CAAC,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,MAAM,EAAE,OAAO,IAAI,EAAE,CAAC,CAAC,CAAC;AAE7J,MAAM,iBAAiB,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAa,EAAW,EAAE,CAC5E,KAAK,CAAC,GAAG,KAAK,KAAK,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,EAAE,CAAC,IAAI,mBAAmB,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;AAElG,MAAM,QAAQ,GAAG,CAAC,IAAe,EAAW,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,IAAI,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,iBAAiB,CAAC,IAAI,CAAC,CAAC;AAEnJ,MAAM,UAAU,GAAG,CAAC,IAAuB,EAAE,KAAY,EAAW,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,KAAK,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;AAEhI,MAAM,aAAa,GAAG,YAAY,CAAC;AAEnC,gHAAgH;AAChH,MAAM,YAAY,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CAC/D,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,QAAQ,EAAE,GAAG,KAAK,MAAM,IAAI,UAAU,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEvH,oHAAoH;AACpH,MAAM,WAAW,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAa,EAAE,MAA2B,EAAW,EAAE,CACnG,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,QAAQ,EAAE,GAAG,KAAK,KAAK,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,OAAO,GAAG,QAAQ,CAAC,OAAO,GAAG,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;AAE/J;;;;GAIG;AACH,MAAM,KAAK,GAAG,CAAC,IAAe,EAAE,MAA2B,EAAU,EAAE;IACrE,IAAI,WAAW,CAAC,IAAI,EAAE,MAAM,CAAC;QAAE,OAAO,OAAO,CAAC;IAC9C,IAAI,YAAY,CAAC,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC;IACxC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC;QAAE,OAAO,MAAM,CAAC;IACnC,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,WAAW,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;AAC7G,CAAC,CAAC;AAEF,uHAAuH;AACvH,MAAM,iBAAiB,GAAG,8CAA8C,CAAC;AACzE,MAAM,YAAY,GAAG,uBAAuB,CAAC;AAE7C,qGAAqG;AACrG,MAAM,iBAAiB,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CACpE,iBAAiB,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,UAAU,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,QAAQ,EAAE,OAAO,IAAI,EAAE,CAAC,CAAC;AAEjJ;;;GAGG;AACH,MAAM,kBAAkB,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAa,EAAW,EAAE,CACrE,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,KAAK,CAAC,OAAO,KAAK,GAAG,IAAI,QAAQ,KAAK,SAAS,IAAI,QAAQ,CAAC,GAAG,KAAK,KAAK,CAAC;AAIpG,MAAM,YAAY,GAAG,CAAC,IAAe,EAAE,EAAE,IAAI,EAAE,MAAM,EAAS,EAAsB,EAAE;IACpF,MAAM,EAAE,KAAK,EAAE,GAAG,IAAI,CAAC;IACvB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;IAC1B,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACrJ,MAAM,OAAO,GAAG,YAAY,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IAClE,IAAI,OAAO,KAAK,SAAS,IAAI,CAAC,GAAG,OAAO,CAAC,CAAC,MAAM,GAAG,WAAW,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,iBAAiB,CAAC,IAAI,CAAC,IAAI,kBAAkB,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IAC7J,OAAO,GAAG,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;AAC1D,CAAC,CAAC;AAEF,0HAA0H;AAC1H,MAAM,YAAY,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAE,KAAY,EAAY,EAAE;IAClF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,EAAY,EAAE;QAC5C,MAAM,GAAG,GAAG,YAAY,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;QAC7F,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;QACrD,IAAI,GAAG,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;YAAE,OAAO,EAAE,CAAC;QAC9G,MAAM,IAAI,GAAG,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;QAC5B,MAAM,YAAY,GAAG,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,IAAI,EAAE,GAAG,KAAK,MAAM,IAAI,UAAU,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAC7F,OAAO,CAAC,EAAE,GAAG,EAAE,QAAQ,EAAE,KAAK,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,YAAY,EAAE,CAAC,CAAC;IAC5F,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF;;;;;GAKG;AACH,MAAM,OAAO,GAAG,CAAC,KAAwB,EAAY,EAAE,CACrD,CAAC,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IACjE,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC;IACxH,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;IACnE,MAAM,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,cAAc,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,CAAC;IACnG,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAY,EAAE;QACtC,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACzC,IAAI,GAAG,KAAK,EAAE,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,EAAE,CAAC;QAClD,MAAM,GAAG,GAAG,IAAI,CAAC,YAAY,KAAK,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;QACpD,OAAO,CAAC,EAAE,GAAG,IAAI,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,IAAI,GAAG,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,GAAG,EAAE,EAAE,CAAC,CAAC;IAC5F,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEL,yHAAyH;AACzH,MAAM,aAAa,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAY,EAAE,CACrE,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAY,EAAE;IAClD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;IACrD,IAAI,KAAK,CAAC,GAAG,KAAK,MAAM,IAAI,CAAC,cAAc,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;QAAE,OAAO,EAAE,CAAC;IACtH,MAAM,GAAG,GAAG,WAAW,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,OAAO,CAAC,GAAG,GAAG,CAAC,CAAC,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,QAAQ,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;AACjJ,CAAC,CAAC,CAAC;AAEL,MAAM,UAAU,GAAG,0CAA0C,CAAC;AAC9D,MAAM,KAAK,GAAG,SAAS,CAAC;AAExB;;;GAGG;AACH,MAAM,aAAa,GAAG,CAAC,IAAY,EAAU,EAAE,CAC7C,IAAI;KACD,KAAK,CAAC,GAAG,CAAC;KACV,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KAC3D,IAAI,CAAC,GAAG,CAAC,CAAC;AAEf;;;;GAIG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAE,IAAyB,EAAY,EAAE;IAC7F,MAAM,QAAQ,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC3C,OAAO,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAY,EAAE;QACzE,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACtB,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACrD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC;QACtC,IAAI,KAAK,CAAC,MAAM,GAAG,aAAa,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,qBAAqB,CAAC,IAAI,EAAE,QAAQ,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;YAAE,OAAO,EAAE,CAAC;QAC1I,IAAI,eAAe,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC;YAAE,OAAO,EAAE,CAAC;QAC7F,OAAO,CAAC,EAAE,GAAG,EAAE,SAAS,KAAK,EAAE,EAAE,QAAQ,EAAE,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;IACtI,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,yFAAyF;AACzF,MAAM,UAAU,GAAG,CAAC,KAAwB,EAAE,EAAE,IAAI,EAAE,KAAK,EAAuB,EAAU,EAAE,CAC5F,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,IAAI,CAAC,GAAG,IAAI,KAAK,CAAC,QAAQ,KAAK,KAAK,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC;AAE5F,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAE,GAAwB,EAAW,EAAE,CAAC,CAAC;IAClF,IAAI,EAAE,EAAE;IACR,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IACpC,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,EAAE,KAAK,EAAE,UAAU,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,EAAE,EAAE,GAAG,CAAC,EAAE,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,MAAM,EAAE;CAC/I,CAAC,CAAC;AAEH;;;;GAIG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACvE,MAAM,KAAK,GAAU,EAAE,IAAI,EAAE,IAAI,GAAG,CAAC,CAAC,GAAG,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,EAAE,CAAC;IAC7I,MAAM,MAAM,GAAG,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,QAAQ,EAAE,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;IAC5G,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,YAAY,CAAC,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;IACzG,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC;IAC3F,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IACjG,MAAM,GAAG,GAAG,CAAC,OAAO,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;IAC9F,MAAM,GAAG,GAAG,CAAC,GAAG,OAAO,EAAE,GAAG,QAAQ,EAAE,GAAG,KAAK,CAAC,CAAC;IAChD,4FAA4F;IAC5F,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;AACrH,CAAC,CAAC"}

package/dist/detectors/orthography.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA0C,MAAM,cAAc,CAAC;AAOrF,6DAA6D;AAC7D,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAgHF;;;GAGG;AACH,eAAO,MAAM,YAAY,EAAE,QAyB1B,CAAC"}
+{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA0C,MAAM,cAAc,CAAC;AAQrF,6DAA6D;AAC7D,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAmHF;;;GAGG;AACH,eAAO,MAAM,YAAY,EAAE,QA0B1B,CAAC"}

package/dist/detectors/orthography.js

@@ -3,6 +3,7 @@ import { escapeRegExp, latinBoundaries, occurrencesOutside } from "../orthograph
 import { minorityReport } from "../spacing-minority.js";
 import { isWithinAny, quotedSpans } from "../quoted-span.js";
 import { digitRunAround, endsWithDivisionLabel, isNumberName, sequenceLabelStarts } from "../number-name.js";
+import { isSpacedLocationPart } from "../location-part.js";
 /** チームが chaff.yaml の prefer に書いた「使わない書き方」。書いていなければ何も言わない。 */
 export const preferredTerm = (doc, options) => {
     const pairs = (options.lexicon ?? []).flatMap((entry) => (entry.instead_of === undefined ? [] : [{ avoid: entry.pattern, use: entry.instead_of }]));
@@ -48,6 +49,7 @@ const calendarStarts = (sentence, units) => new Set(calendarRuns(sentence.text,
 /**
  * 番号・識別子として書かれた数（number-name.ts）、日付・時刻の数、「第1節」の後ろの境目、文頭の項目の番号（「一 JIS」）の後ろの境目は、空け方の好みではないので数えない。
  * 日付・時刻は前の境目（「は 9月」「午後3時」「令和 3 年」）も数えない。日付はまとめて一つの書き方で、数量の空け方の票にはしない。
+ * 「本社 5階」のように名前の後ろに空白で区切った階・部屋の番号の前の空白も、所在の組の区切りなので数えない（location-part.ts）。
  */
 const isCounted = (sentence, boundary, context, calendar) => {
     if (followsItemNumber(sentence.text, boundary, context.itemNumbers))
@@ -61,6 +63,8 @@ const isCounted = (sentence, boundary, context, calendar) => {
         return true;
     if (calendar.has(run.start))
         return false;
+    if (boundary.kind === "before-digit" && isSpacedLocationPart(sentence.text, run, sentence.tokens, sentence.span.start, context.floors))
+        return false;
     return !isNumberName(sentence.text, run, sentence.tokens, sentence.span.start, context);
 };
 const patternList = (doc, id) => (doc.lexicons[id] ?? []).map((entry) => entry.pattern);
@@ -118,6 +122,7 @@ export const latinSpacing = (doc, options) => {
         calendar: calendarUnitsOf(doc),
         divisions: labelsAt(doc, "after"),
         itemNumbers: numberedItems(patternList(doc, "item-number"), doc.prose ?? doc.source),
+        floors: { units: patternsOf(doc, "floor-unit"), levels: patternsOf(doc, "floor-level") },
     };
     const located = doc.sentences.flatMap((sentence) => {
         const quoted = quotedSpans(sentence.text);