chaffjs 0.19.0 → 0.21.0

package/dist/detectors/defined-term-form.d.ts

@@ -0,0 +1,25 @@
+import { type DefinedTerm } from "../structure/definition-use.ts";
+import type { Detector, ProseDocument, Sentence } from "../plugin.ts";
+/** A use of the term set in quotes again after its definition: 「本サービス」 where 本サービス was defined. */
+export declare const quotedUses: (source: string, term: string, uses: readonly number[], definedAt: number) => number[];
+/**
+ * The uses of a capitalised term written in lower case ("services" where "Services" is defined), when they are the fewer
+ * and at most limit: a document that mostly writes it in lower case uses the word in its ordinary sense on purpose.
+ */
+export declare const lowerCaseUses: (source: string, term: string, uses: readonly number[], definedAt: number, limit: number) => number[];
+export declare const definedTermForm: Detector;
+/** The long name written right before an inline definition: the run of name words that ends at its opening bracket. */
+export declare const nameBefore: (sentence: Sentence, bracket: number) => string | undefined;
+export type RepeatedName = {
+    readonly offset: number;
+    readonly name: string;
+    readonly term: string;
+    readonly line: number;
+};
+/**
+ * Each use of a long name after the definition gave it a short name, outside a signature line. A name that is itself a
+ * defined term (during each Subscription Term ("Permitted Use")) is a clause's last words, not the thing being named.
+ */
+export declare const repeatedNames: (doc: ProseDocument, terms: readonly DefinedTerm[]) => RepeatedName[];
+export declare const definedNameRepeated: Detector;
+//# sourceMappingURL=defined-term-form.d.ts.map

package/dist/detectors/defined-term-form.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"defined-term-form.d.ts","sourceRoot":"","sources":["../../src/detectors/defined-term-form.ts"],"names":[],"mappings":"AAGA,OAAO,EAAuC,KAAK,WAAW,EAAE,MAAM,gCAAgC,CAAC;AAGvG,OAAO,KAAK,EAAE,QAAQ,EAAW,aAAa,EAAE,QAAQ,EAAe,MAAM,cAAc,CAAC;AAgC5F,mGAAmG;AACnG,eAAO,MAAM,UAAU,GAAI,QAAQ,MAAM,EAAE,MAAM,MAAM,EAAE,MAAM,SAAS,MAAM,EAAE,EAAE,WAAW,MAAM,KAAG,MAAM,EASxG,CAAC;AAIL;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,QAAQ,MAAM,EAAE,MAAM,MAAM,EAAE,MAAM,SAAS,MAAM,EAAE,EAAE,WAAW,MAAM,EAAE,OAAO,MAAM,KAAG,MAAM,EAM7H,CAAC;AA+BF,eAAO,MAAM,eAAe,EAAE,QAe7B,CAAC;AAaF,uHAAuH;AACvH,eAAO,MAAM,UAAU,GAAI,UAAU,QAAQ,EAAE,SAAS,MAAM,KAAG,MAAM,GAAG,SAWzE,CAAC;AAuBF,MAAM,MAAM,YAAY,GAAG;IAAE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAgB5H;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,KAAK,aAAa,EAAE,OAAO,SAAS,WAAW,EAAE,KAAG,YAAY,EAgB7F,CAAC;AAEF,eAAO,MAAM,mBAAmB,EAAE,QAK3B,CAAC"}

package/dist/detectors/defined-term-form.js

@@ -0,0 +1,170 @@
+// A defined term used later in another form, and a long name used again after its short name was defined. Pure. The
+// definitions come from the document's structure tree (以下「甲」という, ("Seller")); which words may stand in a long
+// name is read from the tokens' parts of speech.
+import { definedTerms, usesOf } from "../structure/definition-use.js";
+import { escapeRegExp } from "../orthography.js";
+import { quoteAt } from "./structure-tree.js";
+const bodyOf = (doc) => doc.sentences.map((sentence) => ({ start: sentence.span.start, text: sentence.text }));
+const finding = (doc, rule, offset, variant, values) => ({
+    rule,
+    severity: "info",
+    line: 0,
+    column: 0,
+    quote: quoteAt(doc.source, offset),
+    values: { ...values, offset },
+    variant,
+});
+const OPENING_QUOTES = new Set(["「", "『", '"', "“"]);
+const CLOSING_QUOTES = new Set(["」", "』", '"', "”"]);
+/** A quoted term right before these is defined again (「X」とは, "X" means), which duplicate-definition reads. */
+const DEFINING_AFTER = /^(?:とは|という|\s+(?:means|shall mean|refers to|has the meaning))/u;
+/** The ending a use may carry past the term: a plural or a possessive (Services, Service's). */
+const WORD_ENDING = /^(?:e?s|['’]s?)/u;
+/** The first definition of each term, with the spans of all its definitions. */
+const firstDefinitions = (terms) => {
+    const groups = new Map();
+    terms.forEach((term) => {
+        const group = groups.get(term.term);
+        if (group === undefined)
+            groups.set(term.term, { term, spans: [term.span] });
+        else
+            group.spans.push(term.span);
+    });
+    return [...groups.values()];
+};
+/** A use of the term set in quotes again after its definition: 「本サービス」 where 本サービス was defined. */
+export const quotedUses = (source, term, uses, definedAt) => uses.filter((offset) => {
+    const after = offset + term.length + (WORD_ENDING.exec(source.slice(offset + term.length, offset + term.length + 2))?.[0].length ?? 0);
+    return (offset > definedAt &&
+        OPENING_QUOTES.has(source.charAt(offset - 1)) &&
+        CLOSING_QUOTES.has(source.charAt(after)) &&
+        !DEFINING_AFTER.test(source.slice(after + 1, after + 20)));
+});
+const UPPER_START = /^\p{Lu}/u;
+/**
+ * The uses of a capitalised term written in lower case ("services" where "Services" is defined), when they are the fewer
+ * and at most limit: a document that mostly writes it in lower case uses the word in its ordinary sense on purpose.
+ */
+export const lowerCaseUses = (source, term, uses, definedAt, limit) => {
+    if (!UPPER_START.test(term) || term.toLowerCase() === term)
+        return [];
+    const later = uses.filter((offset) => offset > definedAt);
+    const lower = later.filter((offset) => source.charAt(offset) === term.charAt(0).toLowerCase() && !isSentenceStart(source, offset));
+    const capital = later.filter((offset) => source.charAt(offset) === term.charAt(0));
+    return lower.length > 0 && lower.length < capital.length && lower.length <= limit ? lower : [];
+};
+/** How far back a sentence's start is looked for: past a few spaces to the mark before them. */
+const LOOK_BACK = 40;
+const SENTENCE_END = new Set([".", "!", "?", ":"]);
+/** A sentence's first word is capitalised whatever it is, so only a lower-case use inside a sentence counts. */
+const isSentenceStart = (source, offset) => {
+    const before = source.slice(Math.max(0, offset - LOOK_BACK), offset);
+    const text = before.trimEnd();
+    return (offset < LOOK_BACK && text === "") || before.slice(text.length).includes("\n") || SENTENCE_END.has(text.at(-1) ?? "");
+};
+/** Where the sentence holding the definition ends: a use inside it ("Customer" means the party named "Customer") belongs to the definition. */
+const definingSentenceEnd = (doc, definition) => doc.sentences.find((sentence) => sentence.span.start <= definition.start && definition.start < sentence.span.end)?.span.end ?? definition.end;
+const VERBAL = new Set(["VERB", "AUX"]);
+/** A word used as a verb ("you input"), not the defined thing ("the Input"): a different word that shares the spelling. */
+const isVerbAt = (doc, offset) => {
+    const tokens = doc.sentences.find((candidate) => candidate.span.start <= offset && offset < candidate.span.end)?.tokens ?? [];
+    const at = tokens.findIndex((candidate) => candidate.span.start <= offset && offset < candidate.span.end);
+    const token = tokens[at];
+    if (token === undefined)
+        return false;
+    // After a subject pronoun the word is its verb (you input), whatever the tagger says; after a possessive (your input) it is a noun.
+    const before = tokens[at - 1];
+    const afterSubject = before?.pos === "PRON" && before.features?.["Poss"] !== "Yes";
+    return VERBAL.has(token.pos) || afterSubject;
+};
+export const definedTermForm = (doc, options) => {
+    if (doc.structure === undefined)
+        return [];
+    const texts = bodyOf(doc);
+    return firstDefinitions(definedTerms(doc.structure)).flatMap(({ term, spans }) => {
+        const uses = usesOf(term.term, texts, spans, "loose");
+        const definingEnd = definingSentenceEnd(doc, term.span);
+        const quoted = quotedUses(doc.source, term.term, uses, definingEnd).map((offset) => finding(doc, "defined-term-form", offset, "quoted", { term: term.term, line: term.line }));
+        const nounUses = uses.filter((offset) => !isVerbAt(doc, offset));
+        const lower = lowerCaseUses(doc.source, term.term, nounUses, definingEnd, options.limit).map((offset) => finding(doc, "defined-term-form", offset, "case", { term: term.term, line: term.line }));
+        return [...quoted, ...lower];
+    });
+};
+/** The words a long name is made of: nouns and names (株式会社 みなと 製作所, Acme Corporation), and the marks inside one. */
+const NAME_PARTS = new Set(["NOUN", "PROPN", "NUM"]);
+const NAME_MARKS = new Set(["・", "&", "＆", ".", ","]);
+/** A long name shorter than this (法律, Customer) is an ordinary word, used in its own sense again. */
+const MIN_NAME_CHARS = 4;
+/** In English a long name has two words or more (Acme Corporation); one capitalised word is the term's own kind of word. */
+const MIN_NAME_WORDS = 2;
+const LATIN = /\p{Script=Latin}/u;
+const isNamePart = (token) => NAME_PARTS.has(token.pos) || NAME_MARKS.has(token.surface);
+/** The long name written right before an inline definition: the run of name words that ends at its opening bracket. */
+export const nameBefore = (sentence, bracket) => {
+    const tokens = (sentence.tokens ?? []).filter((token) => token.span.end <= bracket);
+    const firstOfRun = tokens.findLastIndex((token) => !isNamePart(token)) + 1;
+    if (isModifier(tokens[firstOfRun - 1]))
+        return undefined;
+    const run = tokens.slice(firstOfRun).filter((token, index, all) => index > 0 || !NAME_MARKS.has(token.surface) || all.length === 1);
+    const first = run[0];
+    const end = run.at(-1);
+    if (first === undefined || end === undefined || bracket - end.span.end > 1)
+        return undefined;
+    const name = withoutTrailingMarks(sentence.text.slice(first.span.start - sentence.span.start, end.span.end - sentence.span.start));
+    const words = name.split(/\s+/u).length;
+    return [...name].length < MIN_NAME_CHARS || (LATIN.test(name) && words < MIN_NAME_WORDS) ? undefined : name;
+};
+/**
+ * A word that narrows the name before it: an article, an adjective, a verb, or の (妊娠中の女性従業員, the Beta Preview, 本利用規約).
+ * The words then name a kind of thing the definition picks out, not one party, and are used in their own sense again.
+ */
+const MODIFIERS = new Set(["DET", "ADJ", "VERB", "AUX"]);
+const isModifier = (token) => token !== undefined && (MODIFIERS.has(token.pos) || token.surface === "の");
+const TRAILING_MARKS = new Set([",", ".", " "]);
+const withoutTrailingMarks = (text) => {
+    const chars = [...text];
+    return chars.slice(0, chars.findLastIndex((char) => !TRAILING_MARKS.has(char)) + 1).join("");
+};
+/** Where the inline definition's bracket opens: （以下「甲」という。） or ("Seller"). */
+const bracketBefore = (source, definition) => {
+    const before = source.slice(Math.max(0, definition.start - 2), definition.start + 1);
+    const at = Math.max(before.lastIndexOf("（"), before.lastIndexOf("("));
+    return at === -1 ? -1 : Math.max(0, definition.start - 2) + at;
+};
+/** A line that holds the short name too (甲 株式会社みなと) or little else than the long name: a signature block names both. */
+const isNameLine = (source, offset, name, term) => {
+    const start = source.lastIndexOf("\n", offset - 1) + 1;
+    const end = source.indexOf("\n", offset);
+    const line = source.slice(start, end === -1 ? source.length : end).trim();
+    const SLACK = 12;
+    return line.includes(term) || [...line].length <= [...name].length + SLACK;
+};
+/** A name inside a longer word (東京大学 in 東京大学大学院, Acme in Acmeware) is part of another name. */
+const WORD_CHAR = /[\p{Script=Han}\p{Script=Katakana}\p{Script=Latin}\p{N}ー]/u;
+const standsAlone = (source, offset, name) => !WORD_CHAR.test(source.charAt(offset - 1)) && !WORD_CHAR.test(source.charAt(offset + name.length));
+/**
+ * Each use of a long name after the definition gave it a short name, outside a signature line. A name that is itself a
+ * defined term (during each Subscription Term ("Permitted Use")) is a clause's last words, not the thing being named.
+ */
+export const repeatedNames = (doc, terms) => {
+    const defined = new Set(terms.map((term) => term.term));
+    return terms
+        .filter((term) => term.inline)
+        .flatMap((term) => {
+        const bracket = bracketBefore(doc.source, term.span);
+        const sentence = doc.sentences.find((candidate) => candidate.span.start <= bracket && bracket < candidate.span.end);
+        const name = sentence === undefined || bracket === -1 ? undefined : nameBefore(sentence, bracket);
+        if (name === undefined || name === term.term || term.term.includes(name) || defined.has(name))
+            return [];
+        const pattern = new RegExp(escapeRegExp(name), "gu");
+        return doc.sentences
+            .filter((later) => later.span.start > term.span.end)
+            .flatMap((later) => [...later.text.matchAll(pattern)].map((match) => later.span.start + match.index))
+            .filter((offset) => standsAlone(doc.source, offset, name) && !isNameLine(doc.source, offset, name, term.term))
+            .map((offset) => ({ offset, name, term: term.term, line: term.line }));
+    });
+};
+export const definedNameRepeated = (doc) => doc.structure === undefined
+    ? []
+    : repeatedNames(doc, definedTerms(doc.structure)).map((repeated) => finding(doc, "defined-name-repeated", repeated.offset, "name", { name: repeated.name, term: repeated.term, line: repeated.line }));
+//# sourceMappingURL=defined-term-form.js.map

package/dist/detectors/defined-term-form.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"defined-term-form.js","sourceRoot":"","sources":["../../src/detectors/defined-term-form.ts"],"names":[],"mappings":"AAAA,oHAAoH;AACpH,8GAA8G;AAC9G,iDAAiD;AACjD,OAAO,EAAE,YAAY,EAAE,MAAM,EAAmC,MAAM,gCAAgC,CAAC;AACvG,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAG9C,MAAM,MAAM,GAAG,CAAC,GAAkB,EAAc,EAAE,CAAC,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,EAAE,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;AAE1I,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,IAAY,EAAE,MAAc,EAAE,OAAe,EAAE,MAAiD,EAAW,EAAE,CAAC,CAAC;IAClJ,IAAI;IACJ,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC;IAClC,MAAM,EAAE,EAAE,GAAG,MAAM,EAAE,MAAM,EAAE;IAC7B,OAAO;CACR,CAAC,CAAC;AAEH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AACrD,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AACrD,8GAA8G;AAC9G,MAAM,cAAc,GAAG,gEAAgE,CAAC;AACxF,gGAAgG;AAChG,MAAM,WAAW,GAAG,kBAAkB,CAAC;AAEvC,gFAAgF;AAChF,MAAM,gBAAgB,GAAG,CAAC,KAA6B,EAAqE,EAAE;IAC5H,MAAM,MAAM,GAAG,IAAI,GAAG,EAAgD,CAAC;IACvE,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QACrB,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpC,IAAI,KAAK,KAAK,SAAS;YAAE,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;;YACxE,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnC,CAAC,CAAC,CAAC;IACH,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;AAC9B,CAAC,CAAC;AAEF,mGAAmG;AACnG,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,IAAY,EAAE,IAAuB,EAAE,SAAiB,EAAY,EAAE,CAC/G,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;IACrB,MAAM,KAAK,GAAG,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IACvI,OAAO,CACL,MAAM,GAAG,SAAS;QAClB,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC7C,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACxC,CAAC,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,EAAE,CAAC,CAAC,CAC1D,CAAC;AACJ,CAAC,CAAC,CAAC;AAEL,MAAM,WAAW,GAAG,UAAU,CAAC;AAE/B;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,IAAY,EAAE,IAAuB,EAAE,SAAiB,EAAE,KAAa,EAAY,EAAE;IACjI,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,KAAK,IAAI;QAAE,OAAO,EAAE,CAAC;IACtE,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,GAAG,SAAS,CAAC,CAAC;IAC1D,MAAM,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,IAAI,CAAC,eAAe,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC;IACnI,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IACnF,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;AACjG,CAAC,CAAC;AAEF,gGAAgG;AAChG,MAAM,SAAS,GAAG,EAAE,CAAC;AACrB,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEnD,gHAAgH;AAChH,MAAM,eAAe,GAAG,CAAC,MAAc,EAAE,MAAc,EAAW,EAAE;IAClE,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,GAAG,SAAS,CAAC,EAAE,MAAM,CAAC,CAAC;IACrE,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC;IAC9B,OAAO,CAAC,MAAM,GAAG,SAAS,IAAI,IAAI,KAAK,EAAE,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAChI,CAAC,CAAC;AAEF,+IAA+I;AAC/I,MAAM,mBAAmB,GAAG,CAAC,GAAkB,EAAE,UAAgB,EAAU,EAAE,CAC3E,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,IAAI,UAAU,CAAC,KAAK,IAAI,UAAU,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,GAAG,IAAI,UAAU,CAAC,GAAG,CAAC;AAEhJ,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;AAExC,2HAA2H;AAC3H,MAAM,QAAQ,GAAG,CAAC,GAAkB,EAAE,MAAc,EAAW,EAAE;IAC/D,MAAM,MAAM,GAAG,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,MAAM,IAAI,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,IAAI,EAAE,CAAC;IAC9H,MAAM,EAAE,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,MAAM,IAAI,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC1G,MAAM,KAAK,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;IACzB,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACtC,oIAAoI;IACpI,MAAM,MAAM,GAAG,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC9B,MAAM,YAAY,GAAG,MAAM,EAAE,GAAG,KAAK,MAAM,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC,MAAM,CAAC,KAAK,KAAK,CAAC;IACnF,OAAO,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,YAAY,CAAC;AAC/C,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,eAAe,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACnE,IAAI,GAAG,CAAC,SAAS,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IAC3C,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,OAAO,gBAAgB,CAAC,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE;QAC/E,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC;QACtD,MAAM,WAAW,GAAG,mBAAmB,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QACxD,MAAM,MAAM,GAAG,UAAU,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,WAAW,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CACjF,OAAO,CAAC,GAAG,EAAE,mBAAmB,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAC1F,CAAC;QACF,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,CAAC;QACjE,MAAM,KAAK,GAAG,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,EAAE,QAAQ,EAAE,WAAW,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CACtG,OAAO,CAAC,GAAG,EAAE,mBAAmB,EAAE,MAAM,EAAE,MAAM,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CACxF,CAAC;QACF,OAAO,CAAC,GAAG,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC;IAC/B,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,oHAAoH;AACpH,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;AACrD,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AACtD,qGAAqG;AACrG,MAAM,cAAc,GAAG,CAAC,CAAC;AACzB,4HAA4H;AAC5H,MAAM,cAAc,GAAG,CAAC,CAAC;AACzB,MAAM,KAAK,GAAG,mBAAmB,CAAC;AAElC,MAAM,UAAU,GAAG,CAAC,KAAY,EAAW,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEzG,uHAAuH;AACvH,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAE,OAAe,EAAsB,EAAE;IACpF,MAAM,MAAM,GAAG,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,IAAI,OAAO,CAAC,CAAC;IACpF,MAAM,UAAU,GAAG,MAAM,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC;IAC3E,IAAI,UAAU,CAAC,MAAM,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IACzD,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,KAAK,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC;IACpI,MAAM,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;IACrB,MAAM,GAAG,GAAG,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACvB,IAAI,KAAK,KAAK,SAAS,IAAI,GAAG,KAAK,SAAS,IAAI,OAAO,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC;QAAE,OAAO,SAAS,CAAC;IAC7F,MAAM,IAAI,GAAG,oBAAoB,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IACnI,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC;IACxC,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,GAAG,cAAc,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,KAAK,GAAG,cAAc,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AAC9G,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;AACzD,MAAM,UAAU,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,KAAK,SAAS,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,GAAG,CAAC,CAAC;AAErI,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAEhD,MAAM,oBAAoB,GAAG,CAAC,IAAY,EAAU,EAAE;IACpD,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC;IACxB,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC/F,CAAC,CAAC;AAEF,8EAA8E;AAC9E,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,UAAgB,EAAU,EAAE;IACjE,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IACrF,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC;IACtE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;AACjE,CAAC,CAAC;AAIF,wHAAwH;AACxH,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,MAAc,EAAE,IAAY,EAAE,IAAY,EAAW,EAAE;IACzF,MAAM,KAAK,GAAG,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IACvD,MAAM,GAAG,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACzC,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC1E,MAAM,KAAK,GAAG,EAAE,CAAC;IACjB,OAAO,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,GAAG,KAAK,CAAC;AAC7E,CAAC,CAAC;AAEF,+FAA+F;AAC/F,MAAM,SAAS,GAAG,4DAA4D,CAAC;AAC/E,MAAM,WAAW,GAAG,CAAC,MAAc,EAAE,MAAc,EAAE,IAAY,EAAW,EAAE,CAC5E,CAAC,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,GAAkB,EAAE,KAA6B,EAAkB,EAAE;IACjG,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACxD,OAAO,KAAK;SACT,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC;SAC7B,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAChB,MAAM,OAAO,GAAG,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QACrD,MAAM,QAAQ,GAAG,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,OAAO,IAAI,OAAO,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACpH,MAAM,IAAI,GAAG,QAAQ,KAAK,SAAS,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAClG,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC;YAAE,OAAO,EAAE,CAAC;QACzG,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,IAAI,CAAC,CAAC;QACrD,OAAO,GAAG,CAAC,SAAS;aACjB,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC;aACnD,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;aACpG,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;aAC7G,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAC3E,CAAC,CAAC,CAAC;AACP,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,mBAAmB,GAAa,CAAC,GAAG,EAAa,EAAE,CAC9D,GAAG,CAAC,SAAS,KAAK,SAAS;IACzB,CAAC,CAAC,EAAE;IACJ,CAAC,CAAC,aAAa,CAAC,GAAG,EAAE,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC/D,OAAO,CAAC,GAAG,EAAE,uBAAuB,EAAE,QAAQ,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,QAAQ,CAAC,IAAI,EAAE,CAAC,CAClI,CAAC"}

package/dist/detectors/derived-numbers.d.ts

@@ -0,0 +1,6 @@
+import type { Detector } from "../plugin.ts";
+/** 始まり + 期間 ≠ 終わり。 */
+export declare const durationMismatch: Detector;
+/** 起点の年から数えた年数が、文書の日付と合わない。 */
+export declare const elapsedYearsMismatch: Detector;
+//# sourceMappingURL=derived-numbers.d.ts.map

package/dist/detectors/derived-numbers.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"derived-numbers.d.ts","sourceRoot":"","sources":["../../src/detectors/derived-numbers.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA+C,MAAM,cAAc,CAAC;AAkE1F,sBAAsB;AACtB,eAAO,MAAM,gBAAgB,EAAE,QAiB9B,CAAC;AAwEF,+BAA+B;AAC/B,eAAO,MAAM,oBAAoB,EAAE,QAiClC,CAAC"}

package/dist/detectors/derived-numbers.js

@@ -0,0 +1,155 @@
+import { inDocumentOrder } from "../structure/issues.js";
+import { dateStampIndexes, stampCandidates } from "../date-stamp.js";
+import { overlapsAny, spanIndex } from "../compare/spans.js";
+import { yearOf } from "../derived/date-arithmetic.js";
+import { durationMismatches } from "../derived/durations.js";
+import { elapsedMismatches } from "../derived/elapsed.js";
+import { quoteAt } from "./structure-tree.js";
+/** 期間の単位の語彙表。木の数量の単位がどれかに入れば、その単位の期間。 */
+const DURATION_LEXICONS = [
+    ["duration-day", "day"],
+    ["duration-week", "week"],
+    ["duration-month", "month"],
+    ["duration-year", "year"],
+];
+const patternsOf = (doc, id) => (doc.lexicons[id] ?? []).map((entry) => entry.pattern);
+const positioned = (doc, id, position) => (doc.lexicons[id] ?? []).filter((entry) => (entry.position ?? "before") === position).map((entry) => entry.pattern.toLowerCase());
+/** 数のすぐ後ろ（空白一つまで）に書いた単位まで。木の数量は単位の前で終わることがある（30 years の 30）。 */
+const endWithUnit = (source, end, unit) => {
+    const gap = source.charAt(end) === " " ? 1 : 0;
+    return unit !== "" && source.startsWith(unit, end + gap) ? end + gap + unit.length : end;
+};
+const quantitiesOf = (tree, source) => inDocumentOrder(tree).flatMap((node) => {
+    if (node.kind !== "quantity")
+        return [];
+    const unit = String(node.attrs["unit"] ?? "");
+    return [{ start: node.span.start, end: endWithUnit(source, node.span.end, unit), amount: Number(node.attrs["value"]), unit: unit.normalize("NFKC") }];
+});
+const datesOf = (tree) => inDocumentOrder(tree).flatMap((node) => (node.kind === "date" ? [{ ...node.span, value: String(node.attrs["value"]) }] : []));
+const NEAR = 12;
+const textBefore = (doc, span) => doc.source
+    .slice(Math.max(0, span.start - NEAR), span.start)
+    .toLowerCase()
+    .trimEnd();
+const textAfter = (doc, span) => doc.source
+    .slice(span.end, span.end + NEAR)
+    .toLowerCase()
+    .trimStart();
+/** 「約3か月」「3か月程度」"about 3 months": 目安の期間は足し算に使わない。 */
+const isApproximate = (doc, span) => positioned(doc, "approximate-marker", "before").some((word) => textBefore(doc, span).endsWith(word)) ||
+    positioned(doc, "approximate-marker", "after").some((word) => textAfter(doc, span).startsWith(word));
+const durationsOf = (doc, quantities) => quantities.flatMap((quantity) => {
+    const unit = DURATION_LEXICONS.find(([id]) => patternsOf(doc, id).some((pattern) => pattern.normalize("NFKC") === quantity.unit))?.[1];
+    return unit === undefined || isApproximate(doc, quantity) ? [] : [{ start: quantity.start, end: quantity.end, amount: quantity.amount, unit }];
+});
+const sentencesOf = (doc) => doc.sentences.map((sentence) => sentence.span);
+/** 始まり + 期間 ≠ 終わり。 */
+export const durationMismatch = (doc) => {
+    if (doc.structure === undefined)
+        return [];
+    const durations = durationsOf(doc, quantitiesOf(doc.structure, doc.source));
+    return durationMismatches(sentencesOf(doc), datesOf(doc.structure), durations).map((mismatch) => ({
+        rule: "duration-mismatch",
+        severity: "warning",
+        line: 0,
+        column: 0,
+        quote: quoteAt(doc.source, mismatch.end.start),
+        values: {
+            start: doc.source.slice(mismatch.start.start, mismatch.start.end),
+            duration: doc.source.slice(mismatch.duration.start, mismatch.duration.end),
+            end: doc.source.slice(mismatch.end.start, mismatch.end.end),
+            expected: mismatch.expected,
+            offset: mismatch.end.start,
+        },
+    }));
+};
+/** 四桁の年。木が日付と読まなかった英語の年（founded in 2015、born 1980）もここで読む。 */
+const BARE_YEAR = /(?<![\p{N}.,])[12]\d{3}(?!\p{N})/gu;
+const yearsOf = (doc, tree, quantities) => {
+    const dated = datesOf(tree).flatMap((date) => {
+        const year = yearOf(date.value);
+        return year === undefined ? [] : [{ start: date.start, end: date.end, year }];
+    });
+    const taken = spanIndex([...datesOf(tree), ...quantities]);
+    const bare = [...doc.source.matchAll(BARE_YEAR)].flatMap((match) => {
+        const span = { start: match.index, end: match.index + match[0].length };
+        return overlapsAny(taken, span) ? [] : [{ ...span, year: Number(match[0]) }];
+    });
+    return [...dated, ...bare];
+};
+const LATIN = /^[a-z ]+$/iu;
+const WORD_CHAR = /[\p{L}\p{N}]/u;
+/** 語彙表の語が書かれた所。英字の語は語の切れ目でだけ。 */
+const wordsOf = (doc, id) => patternsOf(doc, id).flatMap((pattern) => {
+    const lower = doc.source.toLowerCase();
+    const needle = pattern.toLowerCase();
+    const found = [];
+    for (let at = lower.indexOf(needle); at !== -1; at = lower.indexOf(needle, at + needle.length)) {
+        const bounded = !LATIN.test(pattern) || (!WORD_CHAR.test(lower.charAt(at - 1)) && !WORD_CHAR.test(lower.charAt(at + needle.length)));
+        if (bounded)
+            found.push({ start: at, end: at + needle.length, pattern });
+    }
+    return found;
+});
+/** 年齢: 単位が年齢の語（歳）か、年齢の印（aged、years old）の付いた数量。 */
+const isAge = (doc, quantity) => patternsOf(doc, "age-unit").some((unit) => unit.normalize("NFKC") === quantity.unit) ||
+    positioned(doc, "age-marker", "before").some((word) => textBefore(doc, quantity).endsWith(word)) ||
+    positioned(doc, "age-marker", "after").some((word) => textAfter(doc, quantity).startsWith(word));
+const elapsedOf = (doc, quantities) => quantities.filter((quantity) => !isAge(doc, quantity) && patternsOf(doc, "elapsed-unit").some((unit) => unit.normalize("NFKC") === quantity.unit));
+/** 「aged 45」の 45: 英語の木は単位の無い数を読まないので、年齢の印のすぐ後ろの数を読む。 */
+const AGE_NUMBER = /^\s*(\d{1,3})(?!\d)/u;
+const markedAges = (doc) => positioned(doc, "age-marker", "before").flatMap((marker) => wordsOf(doc, "age-marker")
+    .filter((word) => word.pattern.toLowerCase() === marker)
+    .flatMap((word) => {
+    const number = AGE_NUMBER.exec(doc.source.slice(word.end));
+    if (number === null)
+        return [];
+    const start = word.end + number[0].length - (number[1] ?? "").length;
+    return [{ start, end: word.end + number[0].length, amount: Number(number[1]) }];
+}));
+/** 文書の日付の年: 日付だけの段落か「更新日：」の段落（date-stamp）の、最初の日付。 */
+const referenceYear = (doc, tree) => {
+    const dates = datesOf(tree);
+    const candidates = stampCandidates(doc.source, doc.paragraphs.map((paragraph) => paragraph.span), dates);
+    const stamps = dateStampIndexes(candidates, patternsOf(doc, "date-stamp-label"));
+    const paragraph = doc.paragraphs.find((_, index) => stamps.has(index));
+    const date = paragraph === undefined ? undefined : dates.find((candidate) => candidate.start >= paragraph.span.start && candidate.end <= paragraph.span.end);
+    return date === undefined ? undefined : yearOf(date.value);
+};
+/** 起点の年から数えた年数が、文書の日付と合わない。 */
+export const elapsedYearsMismatch = (doc) => {
+    const tree = doc.structure;
+    const reference = tree === undefined ? undefined : referenceYear(doc, tree);
+    if (tree === undefined || reference === undefined)
+        return [];
+    const quantities = quantitiesOf(tree, doc.source);
+    const quantityAges = quantities.filter((quantity) => isAge(doc, quantity));
+    const taken = spanIndex(quantityAges);
+    const ages = [...quantityAges, ...markedAges(doc).filter((age) => !overlapsAny(taken, age))].toSorted((left, right) => left.start - right.start);
+    const input = {
+        source: doc.source,
+        sentences: sentencesOf(doc),
+        years: yearsOf(doc, tree, quantities),
+        foundings: wordsOf(doc, "elapsed-origin"),
+        births: wordsOf(doc, "birth-origin"),
+        elapsed: elapsedOf(doc, quantities),
+        ages,
+        links: patternsOf(doc, "elapsed-link"),
+        reference,
+    };
+    return elapsedMismatches(input).map((mismatch) => ({
+        rule: "elapsed-years-mismatch",
+        severity: "warning",
+        line: 0,
+        column: 0,
+        quote: quoteAt(doc.source, mismatch.elapsed.start),
+        values: {
+            written: doc.source.slice(mismatch.elapsed.start, mismatch.elapsed.end),
+            origin: doc.source.slice(mismatch.origin.start, mismatch.origin.end),
+            reference,
+            expected: mismatch.expected,
+            offset: mismatch.elapsed.start,
+        },
+    }));
+};
+//# sourceMappingURL=derived-numbers.js.map

package/dist/detectors/derived-numbers.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"derived-numbers.js","sourceRoot":"","sources":["../../src/detectors/derived-numbers.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACzD,OAAO,EAAE,gBAAgB,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAC;AACrE,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,qBAAqB,CAAC;AAC7D,OAAO,EAAE,MAAM,EAAqB,MAAM,+BAA+B,CAAC;AAC1E,OAAO,EAAE,kBAAkB,EAAkC,MAAM,yBAAyB,CAAC;AAC7F,OAAO,EAAE,iBAAiB,EAA4C,MAAM,uBAAuB,CAAC;AACpG,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAE9C,yCAAyC;AACzC,MAAM,iBAAiB,GAAiD;IACtE,CAAC,cAAc,EAAE,KAAK,CAAC;IACvB,CAAC,eAAe,EAAE,MAAM,CAAC;IACzB,CAAC,gBAAgB,EAAE,OAAO,CAAC;IAC3B,CAAC,eAAe,EAAE,MAAM,CAAC;CAC1B,CAAC;AAIF,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,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,EAAU,EAAE,QAA4B,EAAY,EAAE,CAC5F,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,QAAQ,IAAI,QAAQ,CAAC,KAAK,QAAQ,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC;AAEpI,gEAAgE;AAChE,MAAM,WAAW,GAAG,CAAC,MAAc,EAAE,GAAW,EAAE,IAAY,EAAU,EAAE;IACxE,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC/C,OAAO,IAAI,KAAK,EAAE,IAAI,MAAM,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC;AAC3F,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,CAAC,IAAmB,EAAE,MAAc,EAAc,EAAE,CACvE,eAAe,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAc,EAAE;IACjD,IAAI,IAAI,CAAC,IAAI,KAAK,UAAU;QAAE,OAAO,EAAE,CAAC;IACxC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;IAC9C,OAAO,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,WAAW,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;AACxJ,CAAC,CAAC,CAAC;AAEL,MAAM,OAAO,GAAG,CAAC,IAAmB,EAAgB,EAAE,CACpD,eAAe,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AAEhI,MAAM,IAAI,GAAG,EAAE,CAAC;AAEhB,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAE,IAAU,EAAU,EAAE,CAC5D,GAAG,CAAC,MAAM;KACP,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC;KACjD,WAAW,EAAE;KACb,OAAO,EAAE,CAAC;AACf,MAAM,SAAS,GAAG,CAAC,GAAkB,EAAE,IAAU,EAAU,EAAE,CAC3D,GAAG,CAAC,MAAM;KACP,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,GAAG,IAAI,CAAC;KAChC,WAAW,EAAE;KACb,SAAS,EAAE,CAAC;AAEjB,qDAAqD;AACrD,MAAM,aAAa,GAAG,CAAC,GAAkB,EAAE,IAAU,EAAW,EAAE,CAChE,UAAU,CAAC,GAAG,EAAE,oBAAoB,EAAE,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IACpG,UAAU,CAAC,GAAG,EAAE,oBAAoB,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;AAEvG,MAAM,WAAW,GAAG,CAAC,GAAkB,EAAE,UAA+B,EAAc,EAAE,CACtF,UAAU,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;IAC9B,MAAM,IAAI,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACvI,OAAO,IAAI,KAAK,SAAS,IAAI,aAAa,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,GAAG,EAAE,MAAM,EAAE,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;AACjJ,CAAC,CAAC,CAAC;AAEL,MAAM,WAAW,GAAG,CAAC,GAAkB,EAAU,EAAE,CAAC,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;AAEnG,sBAAsB;AACtB,MAAM,CAAC,MAAM,gBAAgB,GAAa,CAAC,GAAG,EAAa,EAAE;IAC3D,IAAI,GAAG,CAAC,SAAS,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IAC3C,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,YAAY,CAAC,GAAG,CAAC,SAAS,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;IAC5E,OAAO,kBAAkB,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QAChG,IAAI,EAAE,mBAAmB;QACzB,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC;QAC9C,MAAM,EAAE;YACN,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC;YACjE,QAAQ,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC;YAC1E,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,EAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC;YAC3D,QAAQ,EAAE,QAAQ,CAAC,QAAQ;YAC3B,MAAM,EAAE,QAAQ,CAAC,GAAG,CAAC,KAAK;SAC3B;KACF,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,6DAA6D;AAC7D,MAAM,SAAS,GAAG,oCAAoC,CAAC;AAEvD,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,IAAmB,EAAE,UAA+B,EAAU,EAAE;IACnG,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAU,EAAE;QACnD,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAChC,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;IAChF,CAAC,CAAC,CAAC;IACH,MAAM,KAAK,GAAG,SAAS,CAAC,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,UAAU,CAAC,CAAC,CAAC;IAC3D,MAAM,IAAI,GAAG,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAU,EAAE;QACzE,MAAM,IAAI,GAAG,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;QACxE,OAAO,WAAW,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,EAAE,IAAI,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAC/E,CAAC,CAAC,CAAC;IACH,OAAO,CAAC,GAAG,KAAK,EAAE,GAAG,IAAI,CAAC,CAAC;AAC7B,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,aAAa,CAAC;AAC5B,MAAM,SAAS,GAAG,eAAe,CAAC;AAElC,iCAAiC;AACjC,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,EAAU,EAAgB,EAAE,CAC/D,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE;IACtC,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;IACvC,MAAM,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;IACrC,MAAM,KAAK,GAAiB,EAAE,CAAC;IAC/B,KAAK,IAAI,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;QAC/F,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACrI,IAAI,OAAO;YAAE,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC;IAC3E,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC,CAAC;AAEL,kDAAkD;AAClD,MAAM,KAAK,GAAG,CAAC,GAAkB,EAAE,QAAkB,EAAW,EAAE,CAChE,UAAU,CAAC,GAAG,EAAE,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,QAAQ,CAAC,IAAI,CAAC;IACpF,UAAU,CAAC,GAAG,EAAE,YAAY,EAAE,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IAChG,UAAU,CAAC,GAAG,EAAE,YAAY,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;AAEnG,MAAM,SAAS,GAAG,CAAC,GAAkB,EAAE,UAA+B,EAAa,EAAE,CACnF,UAAU,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,QAAQ,CAAC,IAAI,UAAU,CAAC,GAAG,EAAE,cAAc,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;AAErJ,wDAAwD;AACxD,MAAM,UAAU,GAAG,sBAAsB,CAAC;AAE1C,MAAM,UAAU,GAAG,CAAC,GAAkB,EAAa,EAAE,CACnD,UAAU,CAAC,GAAG,EAAE,YAAY,EAAE,QAAQ,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,CACzD,OAAO,CAAC,GAAG,EAAE,YAAY,CAAC;KACvB,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,KAAK,MAAM,CAAC;KACvD,OAAO,CAAC,CAAC,IAAI,EAAa,EAAE;IAC3B,MAAM,MAAM,GAAG,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3D,IAAI,MAAM,KAAK,IAAI;QAAE,OAAO,EAAE,CAAC;IAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IACrE,OAAO,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAClF,CAAC,CAAC,CACL,CAAC;AAEJ,qDAAqD;AACrD,MAAM,aAAa,GAAG,CAAC,GAAkB,EAAE,IAAmB,EAAsB,EAAE;IACpF,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAC5B,MAAM,UAAU,GAAG,eAAe,CAChC,GAAG,CAAC,MAAM,EACV,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,EACjD,KAAK,CACN,CAAC;IACF,MAAM,MAAM,GAAG,gBAAgB,CAAC,UAAU,EAAE,UAAU,CAAC,GAAG,EAAE,kBAAkB,CAAC,CAAC,CAAC;IACjF,MAAM,SAAS,GAAG,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;IACvE,MAAM,IAAI,GAAG,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,KAAK,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,SAAS,CAAC,GAAG,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7J,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAC7D,CAAC,CAAC;AAEF,+BAA+B;AAC/B,MAAM,CAAC,MAAM,oBAAoB,GAAa,CAAC,GAAG,EAAa,EAAE;IAC/D,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC;IAC3B,MAAM,SAAS,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,aAAa,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;IAC5E,IAAI,IAAI,KAAK,SAAS,IAAI,SAAS,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IAC7D,MAAM,UAAU,GAAG,YAAY,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAClD,MAAM,YAAY,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC,CAAC;IAC3E,MAAM,KAAK,GAAG,SAAS,CAAC,YAAY,CAAC,CAAC;IACtC,MAAM,IAAI,GAAG,CAAC,GAAG,YAAY,EAAE,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;IACjJ,MAAM,KAAK,GAAG;QACZ,MAAM,EAAE,GAAG,CAAC,MAAM;QAClB,SAAS,EAAE,WAAW,CAAC,GAAG,CAAC;QAC3B,KAAK,EAAE,OAAO,CAAC,GAAG,EAAE,IAAI,EAAE,UAAU,CAAC;QACrC,SAAS,EAAE,OAAO,CAAC,GAAG,EAAE,gBAAgB,CAAC;QACzC,MAAM,EAAE,OAAO,CAAC,GAAG,EAAE,cAAc,CAAC;QACpC,OAAO,EAAE,SAAS,CAAC,GAAG,EAAE,UAAU,CAAC;QACnC,IAAI;QACJ,KAAK,EAAE,UAAU,CAAC,GAAG,EAAE,cAAc,CAAC;QACtC,SAAS;KACV,CAAC;IACF,OAAO,iBAAiB,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACjD,IAAI,EAAE,wBAAwB;QAC9B,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC;QAClD,MAAM,EAAE;YACN,OAAO,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC;YACvE,MAAM,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,QAAQ,CAAC,MAAM,CAAC,GAAG,CAAC;YACpE,SAAS;YACT,QAAQ,EAAE,QAAQ,CAAC,QAAQ;YAC3B,MAAM,EAAE,QAAQ,CAAC,OAAO,CAAC,KAAK;SAC/B;KACF,CAAC,CAAC,CAAC;AACN,CAAC,CAAC"}

package/dist/detectors/document-shape.d.ts

@@ -0,0 +1,18 @@
+import type { Detector, LengthUnit, Paragraph, ProseDocument, Section } from "../plugin.ts";
+/** The document's title: its first top-level heading, when it has words. */
+export declare const titleSectionOf: (doc: ProseDocument) => Section | undefined;
+/** The paragraph right under the title, with nothing but blank lines between them. A list, a table or an image first is not a lead. */
+export declare const leadParagraphOf: (doc: ProseDocument, title: Section) => Paragraph | undefined;
+/** How much of the title the lead repeats: the share of the title's character trigrams found in the lead, in percent. */
+export declare const echoPercent: (title: string, lead: string) => number;
+/**
+ * The lead restates the title: it repeats most of the title and adds fewer than `limit` content words of its own.
+ * Both must hold: a lead that adds little but is about something else (a date, a byline) is not a restatement.
+ */
+export declare const noLead: Detector;
+/** A heading's length in the document's unit: words, or characters with each Latin word counted as two. */
+export declare const headingLength: (heading: string, unit: LengthUnit) => number;
+export declare const isParagraphHeading: (heading: string) => boolean;
+/** Each title or heading longer than the limit. The first top-level heading is the title; every other is a heading. */
+export declare const titleLength: Detector;
+//# sourceMappingURL=document-shape.d.ts.map

package/dist/detectors/document-shape.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"document-shape.d.ts","sourceRoot":"","sources":["../../src/detectors/document-shape.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAE,QAAQ,EAAW,UAAU,EAAiB,SAAS,EAAE,aAAa,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AAUpH,4EAA4E;AAC5E,eAAO,MAAM,cAAc,GAAI,KAAK,aAAa,KAAG,OAAO,GAAG,SAAiG,CAAC;AAEhK,uIAAuI;AACvI,eAAO,MAAM,eAAe,GAAI,KAAK,aAAa,EAAE,OAAO,OAAO,KAAG,SAAS,GAAG,SAIhF,CAAC;AAEF,yHAAyH;AACzH,eAAO,MAAM,WAAW,GAAI,OAAO,MAAM,EAAE,MAAM,MAAM,KAAG,MAKzD,CAAC;AAaF;;;GAGG;AACH,eAAO,MAAM,MAAM,EAAE,QAcpB,CAAC;AASF,2GAA2G;AAC3G,eAAO,MAAM,aAAa,GAAI,SAAS,MAAM,EAAE,MAAM,UAAU,KAAG,MAMjE,CAAC;AAQF,eAAO,MAAM,kBAAkB,GAAI,SAAS,MAAM,KAAG,OAAgD,CAAC;AAMtG,uHAAuH;AACvH,eAAO,MAAM,WAAW,EAAE,QAazB,CAAC"}

package/dist/detectors/document-shape.js

@@ -0,0 +1,94 @@
+// The shape of a document's top: whether the lead says anything the title does not, and whether a title or heading
+// runs too long to scan. Pure; both read only the sections, paragraphs and tokens chaff has already built.
+import { newContentMorphemes } from "./content-morphemes.js";
+import { trigrams } from "./heading-overlap.js";
+import { findingAt, writtenHeadings } from "./markup-finding.js";
+/** A title shorter than this many character trigrams matches any lead by chance. */
+const MIN_TITLE_GRAMS = 4;
+/** The share of the title's trigrams the lead must repeat before its new words are counted: a lead on another subject is not an echo. */
+const MIN_ECHO_PERCENT = 50;
+const PERCENT = 100;
+/** The document's title: its first top-level heading, when it has words. */
+export const titleSectionOf = (doc) => doc.sections.find((section) => section.depth === 1 && section.heading.trim() !== "");
+/** The paragraph right under the title, with nothing but blank lines between them. A list, a table or an image first is not a lead. */
+export const leadParagraphOf = (doc, title) => {
+    const first = doc.paragraphs.find((paragraph) => paragraph.span.start >= title.span.start && paragraph.span.start < title.span.end);
+    if (first === undefined || first.sentences.length === 0)
+        return undefined;
+    return doc.source.slice(title.span.start, first.span.start).trim() === "" ? first : undefined;
+};
+/** How much of the title the lead repeats: the share of the title's character trigrams found in the lead, in percent. */
+export const echoPercent = (title, lead) => {
+    const titleGrams = trigrams(title);
+    if (titleGrams.size < MIN_TITLE_GRAMS)
+        return 0;
+    const leadGrams = trigrams(lead);
+    return Math.round(([...titleGrams].filter((gram) => leadGrams.has(gram)).length / titleGrams.size) * PERCENT);
+};
+const leadOf = (doc) => {
+    const title = titleSectionOf(doc);
+    const lead = title === undefined ? undefined : leadParagraphOf(doc, title);
+    if (title === undefined || lead === undefined || title.headingTokens === undefined)
+        return undefined;
+    const leadTokens = lead.sentences.flatMap((sentence) => sentence.tokens ?? []);
+    const leadText = lead.sentences.map((sentence) => sentence.text).join(" ");
+    return { title, lead, fresh: newContentMorphemes(title.headingTokens, leadTokens), echo: echoPercent(title.heading, leadText) };
+};
+/**
+ * The lead restates the title: it repeats most of the title and adds fewer than `limit` content words of its own.
+ * Both must hold: a lead that adds little but is about something else (a date, a byline) is not a restatement.
+ */
+export const noLead = (doc, options) => {
+    const found = leadOf(doc);
+    if (found === undefined || found.echo < MIN_ECHO_PERCENT || found.fresh >= options.limit)
+        return [];
+    const first = found.lead.sentences[0];
+    return [
+        {
+            rule: "no-lead",
+            severity: "info",
+            line: 0,
+            column: 0,
+            quote: first?.text.trim() ?? "",
+            values: { title: found.title.heading, count: found.fresh, echo: found.echo, limit: options.limit, offset: first?.span.start ?? found.lead.span.start },
+        },
+    ];
+};
+/** A run of Latin letters and digits inside Japanese: a word, read at about the width of two characters. */
+const LATIN_WORD = /[A-Za-z0-9][A-Za-z0-9._\-/@+#]*/gu;
+const LATIN_WORD_CHARS = 2;
+/** Inline Markdown left in a heading's text (`code`, **bold**, _emphasis_), which the reader does not see. */
+const INLINE_MARKS = /[`*_]/gu;
+/** A heading's length in the document's unit: words, or characters with each Latin word counted as two. */
+export const headingLength = (heading, unit) => {
+    const plain = heading.replace(INLINE_MARKS, "");
+    if (unit === "word")
+        return plain.split(/\s+/u).filter((word) => /[\p{L}\p{N}]/u.test(word)).length;
+    const latinWords = [...plain.matchAll(LATIN_WORD)];
+    const rest = plain.replace(LATIN_WORD, "").replace(/\s+/gu, "");
+    return [...rest].length + latinWords.length * LATIN_WORD_CHARS;
+};
+/**
+ * A heading that holds a sentence end followed by more text: a description pasted into the title line, not a name written
+ * too long. A paragraph above a ---- rule is already left out by writtenHeadings.
+ */
+const SENTENCES_INSIDE = /[。．！？](?=\s*\S)|\p{Ll}{2}[.!?]["'”’)]?\s+["'“‘(]?\p{Lu}/u;
+export const isParagraphHeading = (heading) => SENTENCES_INSIDE.test(heading.trim());
+/** The heading's words without a number label at its head (Chapter 2, 1.), from the section the heading opens. */
+const unlabeledOf = (doc, heading) => doc.sections.find((section) => section.span.start === heading.end)?.unlabeledHeading ?? heading.text;
+/** Each title or heading longer than the limit. The first top-level heading is the title; every other is a heading. */
+export const titleLength = (doc, options) => {
+    const headings = writtenHeadings(doc);
+    const title = headings.find((heading) => heading.depth === 1);
+    return headings
+        .map((heading) => ({ heading, words: unlabeledOf(doc, heading) }))
+        .filter(({ words }) => !isParagraphHeading(words))
+        .map(({ heading, words }) => ({ heading, count: headingLength(words, doc.lengthUnit) }))
+        .filter(({ count }) => count > options.limit)
+        .map(({ heading, count }) => ({
+        ...findingAt(doc, heading, { heading: heading.text.trim(), count, limit: options.limit }, heading === title ? "title" : undefined),
+        rule: "title-length",
+        severity: "info",
+    }));
+};
+//# sourceMappingURL=document-shape.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"document-shape.js","sourceRoot":"","sources":["../../src/detectors/document-shape.ts"],"names":[],"mappings":"AAAA,mHAAmH;AACnH,2GAA2G;AAC3G,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,OAAO,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AAChD,OAAO,EAAE,SAAS,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAGjE,oFAAoF;AACpF,MAAM,eAAe,GAAG,CAAC,CAAC;AAE1B,yIAAyI;AACzI,MAAM,gBAAgB,GAAG,EAAE,CAAC;AAE5B,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB,4EAA4E;AAC5E,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,GAAkB,EAAuB,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,IAAI,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;AAEhK,uIAAuI;AACvI,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,GAAkB,EAAE,KAAc,EAAyB,EAAE;IAC3F,MAAM,KAAK,GAAG,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACpI,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC1E,OAAO,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AAChG,CAAC,CAAC;AAEF,yHAAyH;AACzH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,KAAa,EAAE,IAAY,EAAU,EAAE;IACjE,MAAM,UAAU,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IACnC,IAAI,UAAU,CAAC,IAAI,GAAG,eAAe;QAAE,OAAO,CAAC,CAAC;IAChD,MAAM,SAAS,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IACjC,OAAO,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,GAAG,UAAU,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,CAAC;AAChH,CAAC,CAAC;AAIF,MAAM,MAAM,GAAG,CAAC,GAAkB,EAAoB,EAAE;IACtD,MAAM,KAAK,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAClC,MAAM,IAAI,GAAG,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,eAAe,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;IAC3E,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,aAAa,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACrG,MAAM,UAAU,GAAG,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC;IAC/E,MAAM,QAAQ,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC3E,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,mBAAmB,CAAC,KAAK,CAAC,aAAa,EAAE,UAAU,CAAC,EAAE,IAAI,EAAE,WAAW,CAAC,KAAK,CAAC,OAAO,EAAE,QAAQ,CAAC,EAAE,CAAC;AAClI,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,MAAM,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC1D,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,GAAG,gBAAgB,IAAI,KAAK,CAAC,KAAK,IAAI,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IACpG,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;IACtC,OAAO;QACL;YACE,IAAI,EAAE,SAAS;YACf,QAAQ,EAAE,MAAM;YAChB,IAAI,EAAE,CAAC;YACP,MAAM,EAAE,CAAC;YACT,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE;YAC/B,MAAM,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE;SACvJ;KACF,CAAC;AACJ,CAAC,CAAC;AAEF,4GAA4G;AAC5G,MAAM,UAAU,GAAG,mCAAmC,CAAC;AACvD,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAE3B,8GAA8G;AAC9G,MAAM,YAAY,GAAG,SAAS,CAAC;AAE/B,2GAA2G;AAC3G,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,OAAe,EAAE,IAAgB,EAAU,EAAE;IACzE,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC;IAChD,IAAI,IAAI,KAAK,MAAM;QAAE,OAAO,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC;IACpG,MAAM,UAAU,GAAG,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC;IACnD,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;IAChE,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,GAAG,UAAU,CAAC,MAAM,GAAG,gBAAgB,CAAC;AACjE,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,gBAAgB,GAAG,0DAA0D,CAAC;AAEpF,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,OAAe,EAAW,EAAE,CAAC,gBAAgB,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;AAEtG,kHAAkH;AAClH,MAAM,WAAW,GAAG,CAAC,GAAkB,EAAE,OAAsB,EAAU,EAAE,CACzE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,KAAK,OAAO,CAAC,GAAG,CAAC,EAAE,gBAAgB,IAAI,OAAO,CAAC,IAAI,CAAC;AAEvG,uHAAuH;AACvH,MAAM,CAAC,MAAM,WAAW,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC/D,MAAM,QAAQ,GAAG,eAAe,CAAC,GAAG,CAAC,CAAC;IACtC,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC;IAC9D,OAAO,QAAQ;SACZ,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,WAAW,CAAC,GAAG,EAAE,OAAO,CAAC,EAAE,CAAC,CAAC;SACjE,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAC;SACjD,GAAG,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,aAAa,CAAC,KAAK,EAAE,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;SACvF,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;SAC5C,GAAG,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC;QAC5B,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,EAAE,EAAE,OAAO,EAAE,OAAO,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,EAAE,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;QAClI,IAAI,EAAE,cAAc;QACpB,QAAQ,EAAE,MAAM;KACjB,CAAC,CAAC,CAAC;AACR,CAAC,CAAC"}

package/dist/detectors/email-letter.d.ts

@@ -0,0 +1,46 @@
+import type { Detector, Span } from "../plugin.ts";
+/** One line of the writer's own text. breaks: a heading, the header or the signature stood between it and the line before. */
+export type BodyLine = {
+    readonly start: number;
+    readonly text: string;
+    readonly breaks: boolean;
+};
+/**
+ * Whether a line holds one of the words. A Latin word opens the line (Dear, Best regards) and is followed by no letter,
+ * so "Hi" does not match "History"; a Japanese word may stand anywhere in it (いつもお世話になっております).
+ */
+export declare const lineHas: (line: string, words: readonly string[]) => boolean;
+/**
+ * The writer's own lines, in order: not blank once the header and signature are masked (prose), not a quoted reply, not a
+ * heading. A heading, or a line the source has and prose masks (a header, a signature), breaks the text into parts.
+ */
+export declare const bodyLines: (source: string, prose: string, replyQuotes: readonly Span[]) => BodyLine[];
+export type FrameSlip = {
+    readonly variant: "greeting" | "closing";
+    readonly offset: number;
+};
+/** A missing greeting (none in the first lines) and a missing closing (none in the last lines of any part). */
+export declare const frameSlips: (lines: readonly BodyLine[], greetings: readonly string[], closings: readonly string[]) => FrameSlip[];
+export declare const emailFrame: Detector;
+export type SubjectLine = {
+    readonly start: number;
+    readonly subject: string;
+};
+/** "Subject: Lunch" as its name and value; undefined for a line that is no field. */
+export declare const fieldOf: (line: string) => {
+    readonly name: string;
+    readonly value: string;
+} | undefined;
+/** The subject lines (Subject: …, 件名：…), outside a quoted reply, with reply and forward marks (Re:, Fwd:) taken off the value. */
+export declare const subjectLines: (source: string, fields: readonly string[], replyMarks: readonly string[], replyQuotes: readonly Span[]) => SubjectLine[];
+/** "Re: Fwd: Lunch" is "Lunch": the marks a mail program adds are not what the writer wrote. Read piece by piece between colons. */
+export declare const withoutReplyMarks: (subject: string, marks: readonly string[]) => string;
+export declare const emailSubject: Detector;
+/** Where the text first mentions an attachment (添付します, attached), outside the phrases that only look like one (attached to). */
+export declare const firstMention: (lines: readonly BodyLine[], mentions: readonly string[], notMentions: readonly string[]) => number | undefined;
+/** An attachment shown: a line that labels it, a file name with a listed extension, or the archive's note. text is the writer's own (no quoted reply). */
+export declare const showsAttachment: (text: string, labels: readonly string[], extensions: readonly string[], notes: readonly string[]) => boolean;
+/** The source without someone else's words: a quoted reply's spans blanked and its "> " lines dropped. An old message's attachment is not this one's. */
+export declare const ownText: (source: string, replyQuotes: readonly Span[]) => string;
+export declare const attachmentMissing: Detector;
+//# sourceMappingURL=email-letter.d.ts.map

package/dist/detectors/email-letter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"email-letter.d.ts","sourceRoot":"","sources":["../../src/detectors/email-letter.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,QAAQ,EAA0B,IAAI,EAAE,MAAM,cAAc,CAAC;AAO3E,8HAA8H;AAC9H,MAAM,MAAM,QAAQ,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,OAAO,CAAA;CAAE,CAAC;AAOnG;;;GAGG;AACH,eAAO,MAAM,OAAO,GAAI,MAAM,MAAM,EAAE,OAAO,SAAS,MAAM,EAAE,KAAG,OAOhE,CAAC;AAoCF;;;GAGG;AACH,eAAO,MAAM,SAAS,GAAI,QAAQ,MAAM,EAAE,OAAO,MAAM,EAAE,aAAa,SAAS,IAAI,EAAE,KAAG,QAAQ,EAgB/F,CAAC;AAaF,MAAM,MAAM,SAAS,GAAG;IAAE,QAAQ,CAAC,OAAO,EAAE,UAAU,GAAG,SAAS,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAE9F,+GAA+G;AAC/G,eAAO,MAAM,UAAU,GAAI,OAAO,SAAS,QAAQ,EAAE,EAAE,WAAW,SAAS,MAAM,EAAE,EAAE,UAAU,SAAS,MAAM,EAAE,KAAG,SAAS,EAU3H,CAAC;AAEF,eAAO,MAAM,UAAU,EAAE,QAWtB,CAAC;AAEJ,MAAM,MAAM,WAAW,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAAC;AAM/E,qFAAqF;AACrF,eAAO,MAAM,OAAO,GAAI,MAAM,MAAM,KAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAAG,SAc1F,CAAC;AAEF,kIAAkI;AAClI,eAAO,MAAM,YAAY,GAAI,QAAQ,MAAM,EAAE,QAAQ,SAAS,MAAM,EAAE,EAAE,YAAY,SAAS,MAAM,EAAE,EAAE,aAAa,SAAS,IAAI,EAAE,KAAG,WAAW,EAQhJ,CAAC;AAEF,oIAAoI;AACpI,eAAO,MAAM,iBAAiB,GAAI,SAAS,MAAM,EAAE,OAAO,SAAS,MAAM,EAAE,KAAG,MAM7E,CAAC;AAEF,eAAO,MAAM,YAAY,EAAE,QAevB,CAAC;AAEL,gIAAgI;AAChI,eAAO,MAAM,YAAY,GAAI,OAAO,SAAS,QAAQ,EAAE,EAAE,UAAU,SAAS,MAAM,EAAE,EAAE,aAAa,SAAS,MAAM,EAAE,KAAG,MAAM,GAAG,SAQ/H,CAAC;AAoBF,0JAA0J;AAC1J,eAAO,MAAM,eAAe,GAAI,MAAM,MAAM,EAAE,QAAQ,SAAS,MAAM,EAAE,EAAE,YAAY,SAAS,MAAM,EAAE,EAAE,OAAO,SAAS,MAAM,EAAE,KAAG,OAIlI,CAAC;AAEF,yJAAyJ;AACzJ,eAAO,MAAM,OAAO,GAAI,QAAQ,MAAM,EAAE,aAAa,SAAS,IAAI,EAAE,KAAG,MAgBtE,CAAC;AAEF,eAAO,MAAM,iBAAiB,EAAE,QAiB/B,CAAC"}