chaffjs 0.20.0 → 0.22.0

package/dist/detectors/clause-shape.js

@@ -0,0 +1,149 @@
+import { isWithinAny, quotedIn } from "../quoted-span.js";
+// The shape of a clause read from its words: how it ends, a paired word left alone, an adverb that needs a
+// negation, a particle used where another one belongs. Every word comes from a lexicon; this file knows parts of speech.
+const findingAt = (sentence, offset, values) => ({
+    rule: "",
+    severity: "info",
+    line: 0,
+    column: 0,
+    quote: sentence.text.trim(),
+    values: { ...values, offset },
+});
+const patternsOf = (lexicon, group) => new Set((lexicon ?? []).filter((entry) => group === undefined || entry.group === group).map((entry) => entry.pattern));
+const isSpace = (token) => token !== undefined && token.surface.trim() === "";
+/** The tokens of a sentence outside anything quoted in 「」『』, with their index in the sentence. */
+const ownTokens = (sentence) => {
+    const quoted = quotedIn(sentence);
+    return (sentence.tokens ?? []).flatMap((token, at) => (isWithinAny(quoted, token.span) ? [] : [{ token, at }]));
+};
+const wordOf = (token) => token.lemma ?? token.surface;
+const isListed = (words, token) => token !== undefined && (words.has(token.surface) || words.has(wordOf(token)));
+// --- ending-run: sentences in a row that end with the same words ---------------------------------------------
+const TRAILING_POS = new Set(["PUNCT", "SYM"]);
+const TAIL_POS = new Set(["AUX", "PART"]);
+const VERBAL_POS = new Set(["VERB"]);
+/**
+ * How a sentence ends: its closing auxiliaries, and the verb before them as written (しました, あります, 思います).
+ * A noun before them is left out, so 必要です and 重要です end the same way. No auxiliary, no ending to compare.
+ */
+export const endingOf = (sentence) => {
+    const tokens = (sentence.tokens ?? []).filter((token) => !isSpace(token));
+    const last = tokens.findLastIndex((token) => !TRAILING_POS.has(token.pos));
+    const head = tokens.slice(0, last + 1).findLastIndex((token) => !TAIL_POS.has(token.pos));
+    if (head === last)
+        return undefined;
+    const headToken = tokens[head];
+    const tail = tokens.slice(head + 1, last + 1).map((token) => token.surface);
+    return [headToken !== undefined && VERBAL_POS.has(headToken.pos) ? headToken.surface : "", ...tail].join("");
+};
+/** Runs of consecutive sentences with the same ending. A sentence with no ending to compare closes a run. */
+export const endingRuns = (sentences) => sentences
+    .map((sentence) => ({ sentence, ending: endingOf(sentence) }))
+    .reduce((runs, ended) => {
+    const last = runs.at(-1) ?? [];
+    const continues = ended.ending !== undefined && last[0]?.ending === ended.ending;
+    return continues ? [...runs.slice(0, -1), [...last, ended]] : [...runs, [ended]];
+}, [])
+    .filter((run) => run[0]?.ending !== undefined);
+const runFinding = (run, limit) => {
+    const first = run[0];
+    return {
+        rule: "",
+        severity: "info",
+        line: 0,
+        column: 0,
+        quote: run.map(({ sentence }) => sentence.text.trim()).join(""),
+        values: { word: first?.ending ?? "", count: run.length, limit, offset: first?.sentence.span.start ?? 0 },
+    };
+};
+/** More than limit sentences in a row within one paragraph that end with the same words. */
+export const endingRun = (doc, options) => doc.paragraphs.flatMap((paragraph) => endingRuns(paragraph.sentences)
+    .filter((run) => run.length > options.limit)
+    .map((run) => runFinding(run, options.limit)));
+// --- unpaired-parallel: a word that lists by pairs (〜たり) used once, followed by another clause -----------------
+const CONTENT_POS = new Set(["VERB", "ADJ", "NOUN", "PROPN"]);
+const nextWord = (tokens, at) => tokens.slice(at + 1).find((token) => !isSpace(token));
+/**
+ * The marker is left alone when the sentence holds it once and what follows starts a new clause: a comma,
+ * or a content word other than the lexicon's closer (〜たり、映画を見る / 〜たり見る). 〜たりする and 〜たりもする are whole.
+ */
+const leftAlone = (tokens, at, closers, commas) => {
+    const next = nextWord(tokens, at);
+    if (next === undefined)
+        return false;
+    if (commas.has(next.surface))
+        return true;
+    return CONTENT_POS.has(next.pos) && !isListed(closers, next);
+};
+const CONJUNCTION = "CCONJ";
+const markersIn = (sentence, markers) => ownTokens(sentence).filter(({ token }) => token.pos === CONJUNCTION && markers.has(token.surface));
+/** A sentence that uses the lexicon's paired marker once and goes on without its partner. */
+export const unpairedParallel = (doc, options) => {
+    const markers = patternsOf(options.lexicon, "marker");
+    const closers = patternsOf(options.lexicon, "closer");
+    const commas = patternsOf(options.lexicon, "comma");
+    return doc.sentences.flatMap((sentence) => {
+        const found = markersIn(sentence, markers);
+        const only = found.length === 1 ? found[0] : undefined;
+        if (only === undefined || !leftAlone(sentence.tokens ?? [], only.at, closers, commas))
+            return [];
+        return [findingAt(sentence, only.token.span.start, { word: only.token.surface })];
+    });
+};
+const negationsOf = (lexicon) => ({
+    words: new Set((lexicon ?? []).filter((entry) => entry.group === undefined).map((entry) => entry.pattern)),
+    prefixes: [...patternsOf(lexicon, "prefix")],
+});
+/** A negating word (ない, 違う) or a word opening with a negating prefix (不十分, 無理). */
+const isNegation = (token, negations) => isListed(negations.words, token) || negations.prefixes.some((prefix) => token.surface.length > prefix.length && token.surface.startsWith(prefix));
+/** A negation of the writer's own after the adverb; one inside a quotation does not answer it. */
+const hasNegation = (own, from, negations) => own.some(({ token, at }) => at > from && isNegation(token, negations));
+/** Each of the lexicon's adverbs (全然, 決して) with no word of the negation list after it in its sentence. */
+export const adverbPolarity = (doc, options) => {
+    const adverbs = patternsOf(options.lexicon);
+    const negations = negationsOf(doc.lexicons["negation-word"]);
+    return doc.sentences.flatMap((sentence) => {
+        const own = ownTokens(sentence);
+        return own
+            .filter(({ token, at }) => adverbs.has(token.surface) && !hasNegation(own, at, negations))
+            .map(({ token }) => findingAt(sentence, token.span.start, { word: token.surface }));
+    });
+};
+// --- origin-particle: a comparison particle (より) used for "from" -------------------------------------------------
+/** Adverbs (順次) and punctuation (本日より、発送) stand between the particle and its head. */
+const SKIPPED_POS = new Set(["ADV", "PUNCT"]);
+const PARTICLE = "ADP";
+const NOMINAL_POS = new Set(["NOUN", "PROPN"]);
+/** The first word after the particle, past adverbs (順次) and a bound honorific prefix (ご連絡). */
+const headAfter = (tokens, at) => tokens.findIndex((token, index) => index > at && !isSpace(token) && !SKIPPED_POS.has(token.pos) && token.features?.["Bound"] !== "Yes");
+/**
+ * The head names a start or a sending (開始, 参る, 連絡). A noun head must be the action itself: followed by an object
+ * marker or a light verb (受付を開始, 開始します), not a noun being compared (昨年より販売が増えた).
+ */
+const isOriginHead = (tokens, head, heads, objectMarkers) => {
+    const token = tokens[head];
+    if (token === undefined || !isListed(heads, token))
+        return false;
+    if (!NOMINAL_POS.has(token.pos))
+        return true;
+    const next = nextWord(tokens, head);
+    return next !== undefined && (objectMarkers.has(next.surface) || next.features?.["VerbType"] === "Light");
+};
+const originsIn = (sentence, words) => {
+    const tokens = sentence.tokens ?? [];
+    return ownTokens(sentence)
+        .filter(({ token, at }) => token.pos === PARTICLE && words.particles.has(token.surface) && isOriginHead(tokens, headAfter(tokens, at), words.heads, words.objectMarkers))
+        .map(({ token }) => token);
+};
+const particleWords = (doc, options) => ({
+    particles: patternsOf(options.lexicon),
+    heads: patternsOf(doc.lexicons["origin-head"], "head"),
+    objectMarkers: patternsOf(doc.lexicons["origin-head"], "object"),
+});
+/** The lexicon's particle where it marks a starting point (4月1日より開始), which the lexicon's rewrite says instead. */
+export const originParticle = (doc, options) => {
+    const words = particleWords(doc, options);
+    const rewrites = new Map((options.lexicon ?? []).map((entry) => [entry.pattern, entry.rewrite ?? ""]));
+    return doc.sentences.flatMap((sentence) => originsIn(sentence, words).map((token) => findingAt(sentence, token.span.start, { word: token.surface, suggestion: rewrites.get(token.surface) ?? "" })));
+};
+//# sourceMappingURL=clause-shape.js.map

package/dist/detectors/clause-shape.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"clause-shape.js","sourceRoot":"","sources":["../../src/detectors/clause-shape.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,mBAAmB,CAAC;AAE1D,2GAA2G;AAC3G,yHAAyH;AAEzH,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,MAAc,EAAE,MAAiD,EAAW,EAAE,CAAC,CAAC;IACrH,IAAI,EAAE,EAAE;IACR,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IAC3B,MAAM,EAAE,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE;CAC9B,CAAC,CAAC;AAEH,MAAM,UAAU,GAAG,CAAC,OAA4B,EAAE,KAAc,EAAuB,EAAE,CACvF,IAAI,GAAG,CAAC,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;AAEzH,MAAM,OAAO,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC;AAE1G,kGAAkG;AAClG,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAoD,EAAE;IACzF,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC;IAClC,OAAO,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;AAClH,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,KAAY,EAAU,EAAE,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,OAAO,CAAC;AAEtE,MAAM,QAAQ,GAAG,CAAC,KAA0B,EAAE,KAAwB,EAAW,EAAE,CACjF,KAAK,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;AAEhF,gHAAgH;AAEhH,MAAM,YAAY,GAAwB,IAAI,GAAG,CAAC,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;AACpE,MAAM,QAAQ,GAAwB,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC;AAC/D,MAAM,UAAU,GAAwB,IAAI,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;AAE1D;;;GAGG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,QAAkB,EAAsB,EAAE;IACjE,MAAM,MAAM,GAAG,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;IAC1E,MAAM,IAAI,GAAG,MAAM,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3E,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IAC1F,IAAI,IAAI,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IACpC,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC/B,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC5E,OAAO,CAAC,SAAS,KAAK,SAAS,IAAI,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC/G,CAAC,CAAC;AAIF,6GAA6G;AAC7G,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,SAA8B,EAAa,EAAE,CACtE,SAAS;KACN,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;KAC7D,MAAM,CAAY,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;IACjC,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/B,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,KAAK,SAAS,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,MAAM,KAAK,KAAK,CAAC,MAAM,CAAC;IACjF,OAAO,SAAS,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;AACnF,CAAC,EAAE,EAAE,CAAC;KACL,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,MAAM,KAAK,SAAS,CAAC,CAAC;AAEnD,MAAM,UAAU,GAAG,CAAC,GAAqB,EAAE,KAAa,EAAW,EAAE;IACnE,MAAM,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;IACrB,OAAO;QACL,IAAI,EAAE,EAAE;QACR,QAAQ,EAAE,MAAM;QAChB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;QAC/D,MAAM,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,IAAI,CAAC,EAAE;KACzG,CAAC;AACJ,CAAC,CAAC;AAEF,4FAA4F;AAC5F,MAAM,CAAC,MAAM,SAAS,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE,CAC7D,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,SAAS,EAAE,EAAE,CACnC,UAAU,CAAC,SAAS,CAAC,SAAS,CAAC;KAC5B,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;KAC3C,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAChD,CAAC;AAEJ,kHAAkH;AAElH,MAAM,WAAW,GAAwB,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAEnF,MAAM,QAAQ,GAAG,CAAC,MAAwB,EAAE,EAAU,EAAqB,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;AAEpI;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,MAAwB,EAAE,EAAU,EAAE,OAA4B,EAAE,MAA2B,EAAW,EAAE;IAC7H,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAClC,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACrC,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,OAAO,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;AAC/D,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,OAAO,CAAC;AAE5B,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,OAA4B,EAAoD,EAAE,CACvH,SAAS,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,WAAW,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;AAErG,6FAA6F;AAC7F,MAAM,CAAC,MAAM,gBAAgB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACpE,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;IACtD,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;IACtD,MAAM,MAAM,GAAG,UAAU,CAAC,OAAO,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;IACpD,OAAO,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;QACxC,MAAM,KAAK,GAAG,SAAS,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QACvD,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,EAAE,IAAI,CAAC,EAAE,EAAE,OAAO,EAAE,MAAM,CAAC;YAAE,OAAO,EAAE,CAAC;QACjG,OAAO,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IACpF,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAMF,MAAM,WAAW,GAAG,CAAC,OAA4B,EAAa,EAAE,CAAC,CAAC;IAChE,KAAK,EAAE,IAAI,GAAG,CAAC,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC1G,QAAQ,EAAE,CAAC,GAAG,UAAU,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;CAC7C,CAAC,CAAC;AAEH,mFAAmF;AACnF,MAAM,UAAU,GAAG,CAAC,KAAY,EAAE,SAAoB,EAAW,EAAE,CACjE,QAAQ,CAAC,SAAS,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC;AAIpJ,kGAAkG;AAClG,MAAM,WAAW,GAAG,CAAC,GAAsB,EAAE,IAAY,EAAE,SAAoB,EAAW,EAAE,CAC1F,GAAG,CAAC,IAAI,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,IAAI,IAAI,UAAU,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC,CAAC;AAEzE,0GAA0G;AAC1G,MAAM,CAAC,MAAM,cAAc,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAClE,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAC5C,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,CAAC,QAAQ,CAAC,eAAe,CAAC,CAAC,CAAC;IAC7D,OAAO,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;QACxC,MAAM,GAAG,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAC;QAChC,OAAO,GAAG;aACP,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,EAAE,EAAE,EAAE,SAAS,CAAC,CAAC;aACzF,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,IAAI,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IACxF,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,oHAAoH;AAEpH,sFAAsF;AACtF,MAAM,WAAW,GAAwB,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;AACnE,MAAM,QAAQ,GAAG,KAAK,CAAC;AACvB,MAAM,WAAW,GAAwB,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAEpE,+FAA+F;AAC/F,MAAM,SAAS,GAAG,CAAC,MAAwB,EAAE,EAAU,EAAU,EAAE,CACjE,MAAM,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAC;AAE1I;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,MAAwB,EAAE,IAAY,EAAE,KAA0B,EAAE,aAAkC,EAAW,EAAE;IACvI,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3B,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,QAAQ,CAAC,KAAK,EAAE,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IACjE,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC7C,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IACpC,OAAO,IAAI,KAAK,SAAS,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,OAAO,CAAC,CAAC;AAC5G,CAAC,CAAC;AAIF,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,KAAoB,EAAW,EAAE;IACtE,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,SAAS,CAAC,QAAQ,CAAC;SACvB,MAAM,CACL,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,EAAE,CAChB,KAAK,CAAC,GAAG,KAAK,QAAQ,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,YAAY,CAAC,MAAM,EAAE,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,EAAE,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,aAAa,CAAC,CAChJ;SACA,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;AAC/B,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,GAAkB,EAAE,OAAwB,EAAiB,EAAE,CAAC,CAAC;IACtF,SAAS,EAAE,UAAU,CAAC,OAAO,CAAC,OAAO,CAAC;IACtC,KAAK,EAAE,UAAU,CAAC,GAAG,CAAC,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;IACtD,aAAa,EAAE,UAAU,CAAC,GAAG,CAAC,QAAQ,CAAC,aAAa,CAAC,EAAE,QAAQ,CAAC;CACjE,CAAC,CAAC;AAEH,mHAAmH;AACnH,MAAM,CAAC,MAAM,cAAc,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAClE,MAAM,KAAK,GAAG,aAAa,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IAC1C,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IACvG,OAAO,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CACxC,SAAS,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,IAAI,EAAE,KAAK,CAAC,OAAO,EAAE,UAAU,EAAE,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CACzJ,CAAC;AACJ,CAAC,CAAC"}

package/dist/detectors/code-fences.d.ts

@@ -0,0 +1,34 @@
+export type CodeFence = {
+    /** Where the opening fence line starts. */
+    readonly start: number;
+    /** The first word of the info string, lower-cased; "" when the fence names none. */
+    readonly language: string;
+    /** The lines between the fences. */
+    readonly lines: readonly string[];
+    /** Where the block ends: the end of the closing fence line, or of the document. */
+    readonly end: number;
+};
+export type CodeSpan = {
+    readonly start: number;
+    readonly text: string;
+};
+type Opening = {
+    readonly mark: string;
+    readonly length: number;
+};
+/** The fence a code block's first line opens, with its info string; undefined for an indented block, which has none. */
+export declare const fenceOf: (line: string) => {
+    readonly fence: Opening;
+    readonly info: string;
+} | undefined;
+/** The language an info string names: its first word, without braces or a leading dot ({r}, {.python}). */
+export declare const languageOf: (info: string) => string;
+/**
+ * Every fenced code block, in order, as the Markdown parser reads it: a fence in a list item is one, an indented code
+ * block (four spaces, no fence) is not, and an unclosed fence runs to the end. A fence inside a quotation is left out.
+ */
+export declare const codeFences: (source: string) => readonly CodeFence[];
+/** Inline code spans in a stretch of text: a run of backticks closed by a run of the same length. */
+export declare const codeSpans: (text: string, offset?: number) => CodeSpan[];
+export {};
+//# sourceMappingURL=code-fences.d.ts.map

package/dist/detectors/code-fences.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"code-fences.d.ts","sourceRoot":"","sources":["../../src/detectors/code-fences.ts"],"names":[],"mappings":"AAMA,MAAM,MAAM,SAAS,GAAG;IACtB,2CAA2C;IAC3C,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,oFAAoF;IACpF,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,oCAAoC;IACpC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAC;IAClC,mFAAmF;IACnF,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;CACtB,CAAC;AAEF,MAAM,MAAM,QAAQ,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAIzE,KAAK,OAAO,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAElE,wHAAwH;AACxH,eAAO,MAAM,OAAO,GAAI,MAAM,MAAM,KAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAAG,SAS3F,CAAC;AAEF,2GAA2G;AAC3G,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,MAMzC,CAAC;AAkBF;;;GAGG;AACH,eAAO,MAAM,UAAU,GAAI,QAAQ,MAAM,KAAG,SAAS,SAAS,EAe7D,CAAC;AAEF,qGAAqG;AACrG,eAAO,MAAM,SAAS,GAAI,MAAM,MAAM,EAAE,eAAU,KAAG,QAAQ,EAiB5D,CAAC"}

package/dist/detectors/code-fences.js

@@ -0,0 +1,100 @@
+// Fenced code blocks (``` or ~~~) and inline code spans of a Markdown source. Pure. A fence's language is the first word
+// of its info string (```bash, ```{r}); a fence inside a blockquote is someone else's and is not read.
+import { parse } from "../markdown-read.js";
+import { spanOf } from "../markdown-node.js";
+import { eachPreOrder } from "../tree-walk.js";
+const MIN_FENCE = 3;
+/** The fence a code block's first line opens, with its info string; undefined for an indented block, which has none. */
+export const fenceOf = (line) => {
+    const mark = line.charAt(0);
+    if (mark !== "`" && mark !== "~")
+        return undefined;
+    const length = [...line].findIndex((char) => char !== mark);
+    const run = length === -1 ? line.length : length;
+    if (run < MIN_FENCE)
+        return undefined;
+    const info = line.slice(run).trim();
+    // A backtick fence's info string has no backtick: ```js``` on one line is an inline span.
+    return mark === "`" && info.includes("`") ? undefined : { fence: { mark, length: run }, info };
+};
+/** The language an info string names: its first word, without braces or a leading dot ({r}, {.python}). */
+export const languageOf = (info) => {
+    const word = info.split(/\s/u)[0] ?? "";
+    const opened = word.startsWith("{") ? word.slice(1) : word;
+    const bare = opened.startsWith(".") ? opened.slice(1) : opened;
+    const cut = [...bare].findIndex((char) => char === "{" || char === "}" || char === ",");
+    return (cut === -1 ? bare : bare.slice(0, cut)).toLowerCase();
+};
+/** How many backticks run from `at`. */
+const runAt = (text, at) => {
+    const state = { end: at };
+    while (text.charAt(state.end) === "`")
+        state.end += 1;
+    return state.end - at;
+};
+/** The fences of the last source read: the three rules that read fences parse a document once. */
+const lastRead = { source: undefined, fences: [] };
+const isQuoted = (source, offset) => source
+    .slice(source.lastIndexOf("\n", offset - 1) + 1, offset)
+    .trimStart()
+    .startsWith(">");
+/**
+ * Every fenced code block, in order, as the Markdown parser reads it: a fence in a list item is one, an indented code
+ * block (four spaces, no fence) is not, and an unclosed fence runs to the end. A fence inside a quotation is left out.
+ */
+export const codeFences = (source) => {
+    if (lastRead.source === source)
+        return lastRead.fences;
+    const fences = [];
+    eachPreOrder(parse(source), (node) => {
+        const span = node.type === "code" ? spanOf(node) : undefined;
+        if (span === undefined || isQuoted(source, span.start))
+            return;
+        const lineEnd = source.indexOf("\n", span.start);
+        // A fenced block starts at its fence; an indented block starts at its indentation, which is no fence.
+        const opening = fenceOf(source.slice(span.start, lineEnd === -1 ? source.length : lineEnd).trimEnd());
+        if (opening === undefined)
+            return;
+        fences.push({ start: span.start, language: languageOf(opening.info), lines: (node.value ?? "").split("\n"), end: span.end });
+    });
+    lastRead.source = source;
+    lastRead.fences = fences;
+    return fences;
+};
+/** Inline code spans in a stretch of text: a run of backticks closed by a run of the same length. */
+export const codeSpans = (text, offset = 0) => {
+    const spans = [];
+    const state = { at: 0 };
+    while (state.at < text.length) {
+        const open = text.indexOf("`", state.at);
+        if (open === -1)
+            break;
+        const length = runAt(text, open);
+        const fence = "`".repeat(length);
+        const close = findRun(text, fence, open + length);
+        if (close === -1) {
+            state.at = open + length;
+        }
+        else {
+            spans.push({ start: offset + open, text: text.slice(open + length, close).trim() });
+            state.at = close + length;
+        }
+    }
+    return spans;
+};
+/** The next run of exactly this many backticks from `from`, or -1. */
+const findRun = (text, run, from) => {
+    const state = { at: from };
+    while (state.at < text.length) {
+        const found = text.indexOf(run, state.at);
+        if (found === -1)
+            return -1;
+        if (text.charAt(found + run.length) !== "`" && text.charAt(found - 1) !== "`")
+            return found;
+        state.at = found + run.length;
+        while (text.charAt(state.at) === "`")
+            state.at += 1;
+    }
+    return -1;
+};
+//# sourceMappingURL=code-fences.js.map

package/dist/detectors/code-fences.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"code-fences.js","sourceRoot":"","sources":["../../src/detectors/code-fences.ts"],"names":[],"mappings":"AAAA,yHAAyH;AACzH,uGAAuG;AACvG,OAAO,EAAE,KAAK,EAAE,MAAM,qBAAqB,CAAC;AAC5C,OAAO,EAAE,MAAM,EAAE,MAAM,qBAAqB,CAAC;AAC7C,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAe/C,MAAM,SAAS,GAAG,CAAC,CAAC;AAIpB,wHAAwH;AACxH,MAAM,CAAC,MAAM,OAAO,GAAG,CAAC,IAAY,EAAkE,EAAE;IACtG,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;IAC5B,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG;QAAE,OAAO,SAAS,CAAC;IACnD,MAAM,MAAM,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;IAC5D,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;IACjD,IAAI,GAAG,GAAG,SAAS;QAAE,OAAO,SAAS,CAAC;IACtC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACpC,0FAA0F;IAC1F,OAAO,IAAI,KAAK,GAAG,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,CAAC;AACjG,CAAC,CAAC;AAEF,2GAA2G;AAC3G,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAU,EAAE;IACjD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IAC3D,MAAM,IAAI,GAAG,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/D,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC;IACxF,OAAO,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;AAChE,CAAC,CAAC;AAEF,wCAAwC;AACxC,MAAM,KAAK,GAAG,CAAC,IAAY,EAAE,EAAU,EAAU,EAAE;IACjD,MAAM,KAAK,GAAG,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC;IAC1B,OAAO,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,GAAG;QAAE,KAAK,CAAC,GAAG,IAAI,CAAC,CAAC;IACtD,OAAO,KAAK,CAAC,GAAG,GAAG,EAAE,CAAC;AACxB,CAAC,CAAC;AAEF,kGAAkG;AAClG,MAAM,QAAQ,GAAiE,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;AAEjH,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAE,MAAc,EAAW,EAAE,CAC3D,MAAM;KACH,KAAK,CAAC,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC;KACvD,SAAS,EAAE;KACX,UAAU,CAAC,GAAG,CAAC,CAAC;AAErB;;;GAGG;AACH,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,MAAc,EAAwB,EAAE;IACjE,IAAI,QAAQ,CAAC,MAAM,KAAK,MAAM;QAAE,OAAO,QAAQ,CAAC,MAAM,CAAC;IACvD,MAAM,MAAM,GAAgB,EAAE,CAAC;IAC/B,YAAY,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,EAAE;QACnC,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAC7D,IAAI,IAAI,KAAK,SAAS,IAAI,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO;QAC/D,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;QACjD,sGAAsG;QACtG,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;QACtG,IAAI,OAAO,KAAK,SAAS;YAAE,OAAO;QAClC,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;IAC/H,CAAC,CAAC,CAAC;IACH,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC;IACzB,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC;IACzB,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF,qGAAqG;AACrG,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,MAAM,GAAG,CAAC,EAAc,EAAE;IAChE,MAAM,KAAK,GAAe,EAAE,CAAC;IAC7B,MAAM,KAAK,GAAG,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC;IACxB,OAAO,KAAK,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC9B,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC;QACzC,IAAI,IAAI,KAAK,CAAC,CAAC;YAAE,MAAM;QACvB,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QACjC,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,GAAG,MAAM,CAAC,CAAC;QAClD,IAAI,KAAK,KAAK,CAAC,CAAC,EAAE,CAAC;YACjB,KAAK,CAAC,EAAE,GAAG,IAAI,GAAG,MAAM,CAAC;QAC3B,CAAC;aAAM,CAAC;YACN,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,GAAG,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,MAAM,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACpF,KAAK,CAAC,EAAE,GAAG,KAAK,GAAG,MAAM,CAAC;QAC5B,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF,sEAAsE;AACtE,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,GAAW,EAAE,IAAY,EAAU,EAAE;IAClE,MAAM,KAAK,GAAG,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC;IAC3B,OAAO,KAAK,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC9B,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,CAAC,CAAC;QAC1C,IAAI,KAAK,KAAK,CAAC,CAAC;YAAE,OAAO,CAAC,CAAC,CAAC;QAC5B,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,KAAK,GAAG,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,GAAG;YAAE,OAAO,KAAK,CAAC;QAC5F,KAAK,CAAC,EAAE,GAAG,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC;QAC9B,OAAO,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,GAAG;YAAE,KAAK,CAAC,EAAE,IAAI,CAAC,CAAC;IACtD,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC,CAAC"}

package/dist/detectors/cross-definitions.d.ts

@@ -0,0 +1,4 @@
+import type { CrossDetector } from "../plugin.ts";
+/** A term defined in other words than in most files of the run that define it: reported in the files of the fewer wording. */
+export declare const crossDocDuplicateDefinition: CrossDetector;
+//# sourceMappingURL=cross-definitions.d.ts.map

package/dist/detectors/cross-definitions.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-definitions.d.ts","sourceRoot":"","sources":["../../src/detectors/cross-definitions.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAiD,MAAM,cAAc,CAAC;AA2CjG,8HAA8H;AAC9H,eAAO,MAAM,2BAA2B,EAAE,aAIzC,CAAC"}

package/dist/detectors/cross-definitions.js

@@ -0,0 +1,43 @@
+import { inDocumentOrder } from "../structure/issues.js";
+import { definitionsOutOfStep } from "../structure/cross-definitions.js";
+import { quoteAt } from "./structure-tree.js";
+// cross-doc-duplicate-definition reads definitions as duplicate-definition does (the tree's definition nodes), and compares
+// what each file says after the term: 「X」とは、…をいう / "X" means ….
+/**
+ * A definition that stands alone: not one held within an article or section (scope local), and not the inline kind
+ * (以下「X」という, ("X")), whose meaning is written before the term rather than after it.
+ */
+const standsAlone = (node) => node.kind === "definition" && node.attrs["scope"] !== "local" && node.attrs["placement"] !== "inline" && node.attrs["term"] !== undefined;
+/** The language's words of a definition (をいう, means): the definition-statement word list. */
+const statementWordsOf = (doc) => (doc.lexicons["definition-statement"] ?? []).map((entry) => entry.pattern);
+/**
+ * What the sentence says after the term, up to its end. Empty when the sentence has none of the words of a definition:
+ * 「『AI が書いた』とは言いません」 names the term without defining it.
+ */
+const bodyAfter = (doc, node, statementWords) => {
+    const sentence = doc.sentences.find((candidate) => candidate.span.start <= node.span.start && node.span.start < candidate.span.end);
+    if (sentence === undefined || node.span.end > sentence.span.end)
+        return "";
+    const text = doc.source.slice(sentence.span.start, sentence.span.end);
+    return statementWords.some((word) => text.includes(word)) ? doc.source.slice(node.span.end, sentence.span.end) : "";
+};
+const definitionsOf = (doc, tree) => {
+    const statementWords = statementWordsOf(doc);
+    return inDocumentOrder(tree)
+        .filter(standsAlone)
+        .map((node) => ({ term: String(node.attrs["term"]), body: bodyAfter(doc, node, statementWords), offset: node.span.start }));
+};
+const findingOf = (byPath, odd) => {
+    const doc = byPath.get(odd.path);
+    if (doc === undefined)
+        return [];
+    const values = { term: odd.definition.term, count: odd.files, example: odd.usual.path, offset: odd.definition.offset };
+    return [{ path: odd.path, finding: { rule: "", severity: "warning", line: 0, column: 0, quote: quoteAt(doc.source, odd.definition.offset), values } }];
+};
+/** A term defined in other words than in most files of the run that define it: reported in the files of the fewer wording. */
+export const crossDocDuplicateDefinition = (docs) => {
+    const files = docs.flatMap((doc) => (doc.structure === undefined ? [] : [{ path: doc.path, definitions: definitionsOf(doc, doc.structure) }]));
+    const byPath = new Map(docs.map((doc) => [doc.path, doc]));
+    return definitionsOutOfStep(files).flatMap((odd) => findingOf(byPath, odd));
+};
+//# sourceMappingURL=cross-definitions.js.map

package/dist/detectors/cross-definitions.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-definitions.js","sourceRoot":"","sources":["../../src/detectors/cross-definitions.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACzD,OAAO,EAAE,oBAAoB,EAAoD,MAAM,mCAAmC,CAAC;AAC3H,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAE9C,4HAA4H;AAC5H,gEAAgE;AAEhE;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,IAAmB,EAAW,EAAE,CACnD,IAAI,CAAC,IAAI,KAAK,YAAY,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,KAAK,QAAQ,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,SAAS,CAAC;AAE5I,6FAA6F;AAC7F,MAAM,gBAAgB,GAAG,CAAC,GAAkB,EAAY,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,sBAAsB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEtI;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,GAAkB,EAAE,IAAmB,EAAE,cAAiC,EAAU,EAAE;IACvG,MAAM,QAAQ,GAAG,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACpI,IAAI,QAAQ,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,GAAG,QAAQ,CAAC,IAAI,CAAC,GAAG;QAAE,OAAO,EAAE,CAAC;IAC3E,MAAM,IAAI,GAAG,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACtE,OAAO,cAAc,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACtH,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,GAAkB,EAAE,IAAmB,EAAuB,EAAE;IACrF,MAAM,cAAc,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IAC7C,OAAO,eAAe,CAAC,IAAI,CAAC;SACzB,MAAM,CAAC,WAAW,CAAC;SACnB,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,EAAE,IAAI,EAAE,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;AAChI,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,CAAC,MAA0C,EAAE,GAAwB,EAAqB,EAAE;IAC5G,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACjC,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACjC,MAAM,MAAM,GAAG,EAAE,IAAI,EAAE,GAAG,CAAC,UAAU,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,OAAO,EAAE,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC;IACvH,OAAO,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC;AACzJ,CAAC,CAAC;AAEF,8HAA8H;AAC9H,MAAM,CAAC,MAAM,2BAA2B,GAAkB,CAAC,IAAI,EAAE,EAAE;IACjE,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,WAAW,EAAE,aAAa,CAAC,GAAG,EAAE,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC/I,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;IAC3D,OAAO,oBAAoB,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,SAAS,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC;AAC9E,CAAC,CAAC"}

package/dist/detectors/cross-facts.d.ts

@@ -0,0 +1,4 @@
+import type { CrossDetector } from "../plugin.ts";
+/** A labelled fact given one value in some files of the run and another in the rest: reported in the files of the fewer value. */
+export declare const crossDocFactConflict: CrossDetector;
+//# sourceMappingURL=cross-facts.d.ts.map

package/dist/detectors/cross-facts.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-facts.d.ts","sourceRoot":"","sources":["../../src/detectors/cross-facts.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAuD,MAAM,cAAc,CAAC;AAsDvG,kIAAkI;AAClI,eAAO,MAAM,oBAAoB,EAAE,aAIlC,CAAC"}

package/dist/detectors/cross-facts.js

@@ -0,0 +1,53 @@
+import { nameSpans } from "../compare/proper-nouns.js";
+import { overlapsAny, spanIndex } from "../compare/spans.js";
+import { factValues } from "../facts/fact-values.js";
+import { labelledFacts } from "../facts/labelled-facts.js";
+import { tableFacts } from "../facts/table-facts.js";
+import { emailSpans, versionSpans } from "../facts/address-values.js";
+import { scopedFacts } from "../facts/fact-scope.js";
+import { factsOutOfStep, writtenValue } from "../facts/cross-facts.js";
+import { factWordsOf } from "./fact-consistency.js";
+import { quoteAt } from "./structure-tree.js";
+// cross-doc-fact-conflict reads facts as fact-conflict does (fact-consistency.ts), with two more kinds of value that a docs
+// site repeats from page to page: an email address and a version of three parts or more (2.4.1, v3).
+/** The proper nouns, and the addresses and versions, read as values named by what they are. A name inside an address is the address. */
+const namedValues = (doc) => {
+    const extra = [...emailSpans(doc.source), ...versionSpans(doc.source)];
+    const taken = spanIndex(extra);
+    return [...nameSpans(doc).filter((span) => !overlapsAny(taken, span)), ...extra];
+};
+const summaryHeadings = (doc) => (doc.lexicons["summary-heading"] ?? []).map((entry) => entry.pattern);
+/** A colon right before the value, past emphasis marks: "Fee: $12", 「料金：1,200円」, "**Fee:** $12". */
+const COLON_BEFORE = /[:：][\s*_]*$/u;
+/**
+ * A fact written as an entry: "Item: value", or a table cell. A fact in a sentence (「指摘は 3 件です」) is about what that
+ * page is saying, and the same words on another page are often about something else; within one file, fact-conflict reads them.
+ */
+const isEntry = (source, fact) => COLON_BEFORE.test(source.slice(source.lastIndexOf("\n", fact.value.start - 1) + 1, fact.value.start));
+const factsOf = (doc, tree) => {
+    const values = factValues(tree, doc.source, namedValues(doc));
+    const entries = labelledFacts(doc.source, values, factWordsOf(doc)).filter((fact) => isEntry(doc.source, fact));
+    return scopedFacts([...entries, ...tableFacts(doc.source, values)], tree, doc.source, summaryHeadings(doc));
+};
+const findingOf = (byPath, odd) => {
+    const doc = byPath.get(odd.path);
+    const usualDoc = byPath.get(odd.usual.path);
+    if (doc === undefined || usualDoc === undefined)
+        return [];
+    const values = {
+        label: odd.fact.label,
+        value: writtenValue(doc.source, odd.fact.value),
+        other: writtenValue(usualDoc.source, odd.usual.value.value),
+        count: odd.files,
+        example: odd.usual.path,
+        offset: odd.fact.value.start,
+    };
+    return [{ path: odd.path, finding: { rule: "", severity: "warning", line: 0, column: 0, quote: quoteAt(doc.source, odd.fact.value.start), values } }];
+};
+/** A labelled fact given one value in some files of the run and another in the rest: reported in the files of the fewer value. */
+export const crossDocFactConflict = (docs) => {
+    const files = docs.flatMap((doc) => (doc.structure === undefined ? [] : [{ path: doc.path, facts: factsOf(doc, doc.structure) }]));
+    const byPath = new Map(docs.map((doc) => [doc.path, doc]));
+    return factsOutOfStep(files).flatMap((odd) => findingOf(byPath, odd));
+};
+//# sourceMappingURL=cross-facts.js.map

package/dist/detectors/cross-facts.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-facts.js","sourceRoot":"","sources":["../../src/detectors/cross-facts.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAC;AACvD,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,qBAAqB,CAAC;AAC7D,OAAO,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAa,MAAM,4BAA4B,CAAC;AACtE,OAAO,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AACrD,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AACtE,OAAO,EAAE,WAAW,EAAmB,MAAM,wBAAwB,CAAC;AACtE,OAAO,EAAE,cAAc,EAAE,YAAY,EAAsB,MAAM,yBAAyB,CAAC;AAC3F,OAAO,EAAE,WAAW,EAAE,MAAM,uBAAuB,CAAC;AACpD,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAE9C,4HAA4H;AAC5H,qGAAqG;AAErG,wIAAwI;AACxI,MAAM,WAAW,GAAG,CAAC,GAAkB,EAAU,EAAE;IACjD,MAAM,KAAK,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,GAAG,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;IACvE,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;IAC/B,OAAO,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,CAAC;AACnF,CAAC,CAAC;AAEF,MAAM,eAAe,GAAG,CAAC,GAAkB,EAAY,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEhI,oGAAoG;AACpG,MAAM,YAAY,GAAG,eAAe,CAAC;AAErC;;;GAGG;AACH,MAAM,OAAO,GAAG,CAAC,MAAc,EAAE,IAAU,EAAW,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;AAE/J,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,IAAmB,EAAgB,EAAE;IACxE,MAAM,MAAM,GAAG,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC;IAC9D,MAAM,OAAO,GAAG,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC;IAChH,OAAO,WAAW,CAAC,CAAC,GAAG,OAAO,EAAE,GAAG,UAAU,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,MAAM,EAAE,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC;AAC9G,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,CAAC,MAA0C,EAAE,GAAkB,EAAqB,EAAE;IACtG,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACjC,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC5C,IAAI,GAAG,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IAC3D,MAAM,MAAM,GAAG;QACb,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK;QACrB,KAAK,EAAE,YAAY,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC;QAC/C,KAAK,EAAE,YAAY,CAAC,QAAQ,CAAC,MAAM,EAAE,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC;QAC3D,KAAK,EAAE,GAAG,CAAC,KAAK;QAChB,OAAO,EAAE,GAAG,CAAC,KAAK,CAAC,IAAI;QACvB,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK;KAC7B,CAAC;IACF,OAAO,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC;AACxJ,CAAC,CAAC;AAEF,kIAAkI;AAClI,MAAM,CAAC,MAAM,oBAAoB,GAAkB,CAAC,IAAI,EAAE,EAAE;IAC1D,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACnI,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;IAC3D,OAAO,cAAc,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,SAAS,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC;AACxE,CAAC,CAAC"}

package/dist/detectors/cross-registry/cross-doc-duplicate-definition.d.ts

@@ -0,0 +1,3 @@
+import type { CrossDetector } from "../../plugin.ts";
+export declare const detector: CrossDetector;
+//# sourceMappingURL=cross-doc-duplicate-definition.d.ts.map

package/dist/detectors/cross-registry/cross-doc-duplicate-definition.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-doc-duplicate-definition.d.ts","sourceRoot":"","sources":["../../../src/detectors/cross-registry/cross-doc-duplicate-definition.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAC;AAGrD,eAAO,MAAM,QAAQ,EAAE,aAA2C,CAAC"}

package/dist/detectors/cross-registry/cross-doc-duplicate-definition.js

@@ -0,0 +1,3 @@
+import { crossDocDuplicateDefinition } from "../cross-definitions.js";
+export const detector = crossDocDuplicateDefinition;
+//# sourceMappingURL=cross-doc-duplicate-definition.js.map

package/dist/detectors/cross-registry/cross-doc-duplicate-definition.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-doc-duplicate-definition.js","sourceRoot":"","sources":["../../../src/detectors/cross-registry/cross-doc-duplicate-definition.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,2BAA2B,EAAE,MAAM,yBAAyB,CAAC;AAEtE,MAAM,CAAC,MAAM,QAAQ,GAAkB,2BAA2B,CAAC"}

package/dist/detectors/cross-registry/cross-doc-fact-conflict.d.ts

@@ -0,0 +1,3 @@
+import type { CrossDetector } from "../../plugin.ts";
+export declare const detector: CrossDetector;
+//# sourceMappingURL=cross-doc-fact-conflict.d.ts.map

package/dist/detectors/cross-registry/cross-doc-fact-conflict.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-doc-fact-conflict.d.ts","sourceRoot":"","sources":["../../../src/detectors/cross-registry/cross-doc-fact-conflict.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAC;AAGrD,eAAO,MAAM,QAAQ,EAAE,aAAoC,CAAC"}

package/dist/detectors/cross-registry/cross-doc-fact-conflict.js

@@ -0,0 +1,3 @@
+import { crossDocFactConflict } from "../cross-facts.js";
+export const detector = crossDocFactConflict;
+//# sourceMappingURL=cross-doc-fact-conflict.js.map

package/dist/detectors/cross-registry/cross-doc-fact-conflict.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-doc-fact-conflict.js","sourceRoot":"","sources":["../../../src/detectors/cross-registry/cross-doc-fact-conflict.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,oBAAoB,EAAE,MAAM,mBAAmB,CAAC;AAEzD,MAAM,CAAC,MAAM,QAAQ,GAAkB,oBAAoB,CAAC"}

package/dist/detectors/currency-notation.d.ts

@@ -0,0 +1,29 @@
+import type { Detector } from "../plugin.ts";
+export type CurrencyForm = {
+    readonly pattern: string;
+    readonly position: "before" | "after";
+    readonly currency: string;
+};
+/** 金額 1 つ。form は書き方（側と記号）、currency はその通貨。 */
+export type WrittenAmount = {
+    readonly offset: number;
+    readonly written: string;
+    readonly form: string;
+    readonly currency: string;
+};
+/** 少ないほうの書き方の金額と、多いほうの書き方の例、その通貨の金額の数。 */
+export type CurrencyMinority = {
+    readonly odd: WrittenAmount;
+    readonly usual: string;
+    readonly count: number;
+    readonly of: number;
+};
+/** 文字列の中の金額と、その書き方。重なる所に二つの書き方が当たれば、記号の長いほう（US$ と $ なら US$、米ドル と ドル なら 米ドル）。 */
+export declare const amountsIn: (text: string, forms: readonly CurrencyForm[], multipliers: readonly string[]) => WrittenAmount[];
+/**
+ * 通貨ごとに、多いほうと違う書き方の金額。少ないほうが limit パーセントを超えれば、使い分けている文書と読んで言わない。
+ * いちばん多い書き方が同じ数で並べば、どちらにも立たない。
+ */
+export declare const currencyMinorities: (amounts: readonly WrittenAmount[], limit: number) => CurrencyMinority[];
+export declare const currencyNotation: Detector;
+//# sourceMappingURL=currency-notation.d.ts.map

package/dist/detectors/currency-notation.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"currency-notation.d.ts","sourceRoot":"","sources":["../../src/detectors/currency-notation.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA0B,MAAM,cAAc,CAAC;AAQrE,MAAM,MAAM,YAAY,GAAG;IAAE,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,QAAQ,EAAE,QAAQ,GAAG,OAAO,CAAC;IAAC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAA;CAAE,CAAC;AAE1H,6CAA6C;AAC7C,MAAM,MAAM,aAAa,GAAG;IAAE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAA;CAAE,CAAC;AAEpI,0CAA0C;AAC1C,MAAM,MAAM,gBAAgB,GAAG;IAAE,QAAQ,CAAC,GAAG,EAAE,aAAa,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAA;CAAE,CAAC;AAwBpI,gFAAgF;AAChF,eAAO,MAAM,SAAS,GAAI,MAAM,MAAM,EAAE,OAAO,SAAS,YAAY,EAAE,EAAE,aAAa,SAAS,MAAM,EAAE,KAAG,aAAa,EAY5D,CAAC;AAQ3D;;;GAGG;AACH,eAAO,MAAM,kBAAkB,GAAI,SAAS,SAAS,aAAa,EAAE,EAAE,OAAO,MAAM,KAAG,gBAAgB,EAQpC,CAAC;AAOnE,eAAO,MAAM,gBAAgB,EAAE,QAW9B,CAAC"}

package/dist/detectors/currency-notation.js

@@ -0,0 +1,58 @@
+import { escapeRegExp } from "../orthography.js";
+import { formMinority } from "./form-minority.js";
+import { quoteAround } from "./quote-around.js";
+const AMOUNT = "[0-9０-９]+(?:[,，][0-9０-９]{3})*(?:[.．][0-9０-９]+)?";
+const alternation = (words) => words.map(escapeRegExp).join("|");
+/** 記号が英字で始まる・終わるなら、語の中（USDT、dollarsign）を読まないよう英字の切れ目を求める。 */
+const latinEdge = (pattern, side) => {
+    const edge = side === "start" ? pattern.charAt(0) : pattern.charAt(pattern.length - 1);
+    if (!/[A-Za-z]/u.test(edge))
+        return "";
+    return side === "start" ? "(?<![A-Za-z])" : "(?![A-Za-z])";
+};
+const amountPattern = (form, multipliers) => {
+    const mark = escapeRegExp(form.pattern);
+    const multiplier = multipliers.length === 0 ? "" : `(?:\\s?(?:${alternation(multipliers)}))?`;
+    return form.position === "before"
+        ? new RegExp(`${latinEdge(form.pattern, "start")}${mark}\\s?${AMOUNT}${multiplier}(?![A-Za-z0-9０-９])`, "gu")
+        : new RegExp(`(?<![\\w.,，０-９])${AMOUNT}${multiplier}\\s?${mark}${latinEdge(form.pattern, "end")}`, "gu");
+};
+const overlaps = (left, right) => left.offset < right.offset + right.written.length && right.offset < left.offset + left.written.length;
+/** 文字列の中の金額と、その書き方。重なる所に二つの書き方が当たれば、記号の長いほう（US$ と $ なら US$、米ドル と ドル なら 米ドル）。 */
+export const amountsIn = (text, forms, multipliers) => forms
+    .toSorted((left, right) => right.pattern.length - left.pattern.length)
+    .flatMap((form) => [...text.matchAll(amountPattern(form, multipliers))].map((match) => ({
+    offset: match.index,
+    written: match[0],
+    form: `${form.position}:${form.pattern}`,
+    currency: form.currency,
+})))
+    .reduce((kept, amount) => (kept.some((other) => overlaps(other, amount)) ? kept : [...kept, amount]), [])
+    .toSorted((left, right) => left.offset - right.offset);
+const minorityOf = (ofCurrency, limit) => {
+    const minority = formMinority(ofCurrency, limit);
+    if (minority === undefined)
+        return [];
+    return minority.odd.map((amount) => ({ odd: amount, usual: minority.usual.written, count: minority.odd.length, of: ofCurrency.length }));
+};
+/**
+ * 通貨ごとに、多いほうと違う書き方の金額。少ないほうが limit パーセントを超えれば、使い分けている文書と読んで言わない。
+ * いちばん多い書き方が同じ数で並べば、どちらにも立たない。
+ */
+export const currencyMinorities = (amounts, limit) => [...new Set(amounts.map((amount) => amount.currency))]
+    .flatMap((currency) => minorityOf(amounts.filter((amount) => amount.currency === currency), limit))
+    .toSorted((left, right) => left.odd.offset - right.odd.offset);
+const formsOf = (doc) => (doc.lexicons["currency-notation"] ?? []).flatMap((entry) => entry.position === undefined || entry.group === undefined ? [] : [{ pattern: entry.pattern, position: entry.position, currency: entry.group }]);
+export const currencyNotation = (doc, options) => {
+    const text = doc.prose ?? doc.source;
+    const multipliers = (doc.lexicons["amount-multiplier"] ?? []).map((entry) => entry.pattern);
+    return currencyMinorities(amountsIn(text, formsOf(doc), multipliers), options.limit).map(({ odd, usual, count, of }) => ({
+        rule: "",
+        severity: "info",
+        line: 0,
+        column: 0,
+        quote: quoteAround(text, odd.offset, odd.offset + odd.written.length),
+        values: { written: odd.written.trim(), usual: usual.trim(), count, of, offset: odd.offset },
+    }));
+};
+//# sourceMappingURL=currency-notation.js.map

package/dist/detectors/currency-notation.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"currency-notation.js","sourceRoot":"","sources":["../../src/detectors/currency-notation.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAahD,MAAM,MAAM,GAAG,iDAAiD,CAAC;AAEjE,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAU,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAE5F,6DAA6D;AAC7D,MAAM,SAAS,GAAG,CAAC,OAAe,EAAE,IAAqB,EAAU,EAAE;IACnE,MAAM,IAAI,GAAG,IAAI,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACvF,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,EAAE,CAAC;IACvC,OAAO,IAAI,KAAK,OAAO,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,cAAc,CAAC;AAC7D,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,IAAkB,EAAE,WAA8B,EAAU,EAAE;IACnF,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACxC,MAAM,UAAU,GAAG,WAAW,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,WAAW,CAAC,WAAW,CAAC,KAAK,CAAC;IAC9F,OAAO,IAAI,CAAC,QAAQ,KAAK,QAAQ;QAC/B,CAAC,CAAC,IAAI,MAAM,CAAC,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,GAAG,IAAI,OAAO,MAAM,GAAG,UAAU,oBAAoB,EAAE,IAAI,CAAC;QAC5G,CAAC,CAAC,IAAI,MAAM,CAAC,mBAAmB,MAAM,GAAG,UAAU,OAAO,IAAI,GAAG,SAAS,CAAC,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;AAC7G,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,IAAmB,EAAE,KAAoB,EAAW,EAAE,CACtE,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;AAExG,gFAAgF;AAChF,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,KAA8B,EAAE,WAA8B,EAAmB,EAAE,CACzH,KAAK;KACF,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;KACrE,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAChB,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,aAAa,CAAC,IAAI,EAAE,WAAW,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACnE,MAAM,EAAE,KAAK,CAAC,KAAK;IACnB,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;IACjB,IAAI,EAAE,GAAG,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,OAAO,EAAE;IACxC,QAAQ,EAAE,IAAI,CAAC,QAAQ;CACxB,CAAC,CAAC,CACJ;KACA,MAAM,CAAkB,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,EAAE,MAAM,CAAC,CAAC,EAAE,EAAE,CAAC;KACzH,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC;AAE3D,MAAM,UAAU,GAAG,CAAC,UAAoC,EAAE,KAAa,EAAsB,EAAE;IAC7F,MAAM,QAAQ,GAAG,YAAY,CAAC,UAAU,EAAE,KAAK,CAAC,CAAC;IACjD,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACtC,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,EAAE,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;AAC3I,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,OAAiC,EAAE,KAAa,EAAsB,EAAE,CACzG,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;KACnD,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CACpB,UAAU,CACR,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,QAAQ,KAAK,QAAQ,CAAC,EACxD,KAAK,CACN,CACF;KACA,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AAEnE,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAkB,EAAE,CACrD,CAAC,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAC1D,KAAK,CAAC,QAAQ,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,QAAQ,EAAE,KAAK,CAAC,QAAQ,EAAE,QAAQ,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAC/I,CAAC;AAEJ,MAAM,CAAC,MAAM,gBAAgB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACpE,MAAM,IAAI,GAAG,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,MAAM,CAAC;IACrC,MAAM,WAAW,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC5F,OAAO,kBAAkB,CAAC,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,GAAG,CAAC,EAAE,WAAW,CAAC,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC;QACvH,IAAI,EAAE,EAAE;QACR,QAAQ,EAAE,MAAM;QAChB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,WAAW,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC;QACrE,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE;KAC5F,CAAC,CAAC,CAAC;AACN,CAAC,CAAC"}

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

@@ -1 +1 @@
-{"version":3,"file":"custom.d.ts","sourceRoot":"","sources":["../../src/detectors/custom.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA2B,MAAM,cAAc,CAAC;AA6CtE,wGAAwG;AACxG,eAAO,MAAM,WAAW,EAAE,QASzB,CAAC;AAEF,uIAAuI;AACvI,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAEF,kJAAkJ;AAClJ,eAAO,MAAM,YAAY,EAAE,QAa1B,CAAC"}
+{"version":3,"file":"custom.d.ts","sourceRoot":"","sources":["../../src/detectors/custom.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA4C,MAAM,cAAc,CAAC;AAoDvF,wGAAwG;AACxG,eAAO,MAAM,WAAW,EAAE,QAUzB,CAAC;AAEF,uIAAuI;AACvI,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAEF,kJAAkJ;AAClJ,eAAO,MAAM,YAAY,EAAE,QAa1B,CAAC"}