chaffjs 0.19.0 → 0.20.0

package/dist/detectors/next-word.js

@@ -0,0 +1,125 @@
+import { densityFindings } from "./ai-phrasing.js";
+/**
+ * The word kinds an article can stand before. A conjunction, a preposition or a verb after "a" is the letter a ("option A
+ * or B", "a is the length").
+ */
+const CONTENT = new Set(["NOUN", "PROPN", "ADJ", "ADV", "NUM"]);
+const LETTER_START = /^\p{L}/u;
+const HAS_LOWER = /\p{Ll}/u;
+/** A lower-case letter before capitals (mRNA, pH is not): the first letter is said by its name. */
+const LETTER_BEFORE_CAPITALS = /^\p{Ll}\p{Lu}{2}/u;
+/** Two tokens with nothing but whitespace between them in the source. Code and markup are hidden from the tokens. */
+const adjacent = (source, left, right) => source.slice(left.span.end, right.span.start).trim() === "";
+const ofGroup = (lexicon, group) => lexicon.filter((entry) => entry.group === group).map((entry) => entry.pattern);
+export const articleSoundsOf = (lexicon) => ({
+    vowelLetters: new Set(ofGroup(lexicon, "vowel-letter")),
+    vowelLetterNames: new Set(ofGroup(lexicon, "letter-name-vowel")),
+    eitherLetterNames: new Set(ofGroup(lexicon, "letter-name-either")),
+    consonant: ofGroup(lexicon, "consonant"),
+    vowel: ofGroup(lexicon, "vowel"),
+    word: new Set(ofGroup(lexicon, "word")),
+    either: ofGroup(lexicon, "either"),
+});
+const startsWithAny = (word, starts) => starts.some((start) => word.startsWith(start));
+/** Whether a word in lower or mixed case starts with a vowel sound; undefined when it is said both ways. */
+const wordSound = (word, sounds) => {
+    const lower = word.toLowerCase();
+    if (startsWithAny(lower, sounds.either))
+        return undefined;
+    if (startsWithAny(lower, sounds.vowel))
+        return true;
+    if (startsWithAny(lower, sounds.consonant))
+        return false;
+    return sounds.vowelLetters.has(lower.charAt(0));
+};
+/** Capitals longer than this that hold a vowel may be said as a word (SALT, GIF), so only the lexicon decides them. */
+const SPELLED_OUT_LENGTH = 3;
+/** Whether capitals not in the lexicon are surely said letter by letter: short (MBA), or with no vowel to say (HTTP). */
+const isSpelledOut = (word, sounds) => word.length <= SPELLED_OUT_LENGTH || ![...word.toLowerCase()].some((letter) => sounds.vowelLetters.has(letter));
+/**
+ * Whether a word starts with a vowel sound. Capitals are read as the lexicon says (a NASA probe), else letter by letter when
+ * surely spelled out (an MBA). undefined when the word may be said both ways (a SQL, an SQL; a or an HTTP) or nothing says how.
+ */
+export const startsWithVowelSound = (word, sounds) => {
+    if (LETTER_BEFORE_CAPITALS.test(word))
+        return sounds.vowelLetterNames.has(word.charAt(0).toUpperCase());
+    if (HAS_LOWER.test(word))
+        return wordSound(word, sounds);
+    if (sounds.either.includes(word) || sounds.eitherLetterNames.has(word.charAt(0)))
+        return undefined;
+    if (sounds.word.has(word))
+        return wordSound(word, sounds);
+    return isSpelledOut(word, sounds) ? sounds.vowelLetterNames.has(word.charAt(0)) : undefined;
+};
+const ARTICLES = { a: false, an: true };
+/** The article the next word takes, written in the case of the one written ("A apple" → "An"). */
+const fitting = (written, vowel) => {
+    const article = vowel ? "an" : "a";
+    return written.charAt(0) === written.charAt(0).toUpperCase() ? `${article.charAt(0).toUpperCase()}${article.slice(1)}` : article;
+};
+/** An article that does not fit the sound of the word after it. */
+const WORD = /\p{L}/u;
+/**
+ * Whether a token is the article "a" or "an": lower case, or capitalised as the first word of the sentence. A capital A
+ * inside a sentence is a letter ("grade A eggs", "vitamin A intake").
+ */
+const isArticle = (token, at, tokens) => {
+    const first = tokens.findIndex((candidate) => WORD.test(candidate.surface));
+    return token.surface === token.surface.toLowerCase() || at === first;
+};
+const slipAt = (source, tokens, at, sounds) => {
+    const article = tokens[at];
+    const next = tokens[at + 1];
+    const takesAn = ARTICLES[article?.surface.toLowerCase() ?? ""];
+    if (article === undefined || next === undefined || takesAn === undefined || !isArticle(article, at, tokens))
+        return undefined;
+    if (!CONTENT.has(next.pos) || !LETTER_START.test(next.surface))
+        return undefined;
+    if (!adjacent(source, article, next))
+        return undefined;
+    const vowel = startsWithVowelSound(next.surface, sounds);
+    return vowel === undefined || vowel === takesAn ? undefined : { article, next, fits: fitting(article.surface, vowel) };
+};
+const slipFinding = (sentence, slip) => ({
+    rule: "",
+    severity: "warning",
+    line: 0,
+    column: 0,
+    quote: sentence.text.trim(),
+    values: { matched: `${slip.article.surface} ${slip.next.surface}`, article: slip.fits, offset: slip.article.span.start },
+});
+/** "a" before a vowel sound or "an" before a consonant sound (a apple, an user), by the sound knowledge in the lexicon. */
+export const articleSound = (doc, options) => {
+    const sounds = articleSoundsOf(options.lexicon ?? []);
+    const findings = doc.sentences.flatMap((sentence) => {
+        const tokens = sentence.tokens ?? [];
+        return tokens.flatMap((_token, at) => {
+            const slip = slipAt(doc.source, tokens, at, sounds);
+            return slip === undefined ? [] : [slipFinding(sentence, slip)];
+        });
+    });
+    return findings.length < options.limit ? [] : findings;
+};
+const ADJECTIVE = "ADJ";
+/** The lexicon's intensifier right before an adjective ("very important"), not before an excepted word ("the very first"). */
+const intensifiedIn = (source, sentence, words) => {
+    const tokens = sentence.tokens ?? [];
+    return tokens.flatMap((token, at) => {
+        const next = tokens[at + 1];
+        if (next === undefined || !words.intensifiers.has(token.surface.toLowerCase()) || next.pos !== ADJECTIVE)
+            return [];
+        if (!adjacent(source, token, next))
+            return [];
+        return words.exceptions.has(next.surface.toLowerCase()) ? [] : [{ sentence, matched: `${token.surface} ${next.surface}`, offset: token.span.start }];
+    });
+};
+const EXCEPTIONS = "very-exception";
+/** "very" + adjective, dense for the document's length: each one leans on "very" instead of a stronger word. */
+export const intensifiedAdjective = (doc, options) => {
+    const words = {
+        intensifiers: new Set((options.lexicon ?? []).map((entry) => entry.pattern)),
+        exceptions: new Set((doc.lexicons[EXCEPTIONS] ?? []).map((entry) => entry.pattern)),
+    };
+    return densityFindings(doc, doc.sentences.flatMap((sentence) => intensifiedIn(doc.source, sentence, words)), options.limit);
+};
+//# sourceMappingURL=next-word.js.map

package/dist/detectors/next-word.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"next-word.js","sourceRoot":"","sources":["../../src/detectors/next-word.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAC;AAEnD;;;GAGG;AACH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC;AAEhE,MAAM,YAAY,GAAG,SAAS,CAAC;AAC/B,MAAM,SAAS,GAAG,SAAS,CAAC;AAC5B,mGAAmG;AACnG,MAAM,sBAAsB,GAAG,mBAAmB,CAAC;AAEnD,qHAAqH;AACrH,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAE,IAAW,EAAE,KAAY,EAAW,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC;AAarI,MAAM,OAAO,GAAG,CAAC,OAAgB,EAAE,KAAa,EAAY,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAE9I,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,OAAgB,EAAiB,EAAE,CAAC,CAAC;IACnE,YAAY,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC;IACvD,gBAAgB,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,mBAAmB,CAAC,CAAC;IAChE,iBAAiB,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,oBAAoB,CAAC,CAAC;IAClE,SAAS,EAAE,OAAO,CAAC,OAAO,EAAE,WAAW,CAAC;IACxC,KAAK,EAAE,OAAO,CAAC,OAAO,EAAE,OAAO,CAAC;IAChC,IAAI,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IACvC,MAAM,EAAE,OAAO,CAAC,OAAO,EAAE,QAAQ,CAAC;CACnC,CAAC,CAAC;AAEH,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC;AAE3H,4GAA4G;AAC5G,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,MAAqB,EAAuB,EAAE;IAC7E,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IACjC,IAAI,aAAa,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,CAAC;QAAE,OAAO,SAAS,CAAC;IAC1D,IAAI,aAAa,CAAC,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACpD,IAAI,aAAa,CAAC,KAAK,EAAE,MAAM,CAAC,SAAS,CAAC;QAAE,OAAO,KAAK,CAAC;IACzD,OAAO,MAAM,CAAC,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;AAClD,CAAC,CAAC;AAEF,uHAAuH;AACvH,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAE7B,yHAAyH;AACzH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,MAAqB,EAAW,EAAE,CACpE,IAAI,CAAC,MAAM,IAAI,kBAAkB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;AAElH;;;GAGG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,IAAY,EAAE,MAAqB,EAAuB,EAAE;IAC/F,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,MAAM,CAAC,gBAAgB,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IACxG,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACzD,IAAI,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,iBAAiB,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IACnG,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IAC1D,OAAO,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,gBAAgB,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9F,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAsC,EAAE,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC;AAE3E,kGAAkG;AAClG,MAAM,OAAO,GAAG,CAAC,OAAe,EAAE,KAAc,EAAU,EAAE;IAC1D,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC;IACnC,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC;AACnI,CAAC,CAAC;AAIF,mEAAmE;AACnE,MAAM,IAAI,GAAG,QAAQ,CAAC;AAEtB;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,KAAY,EAAE,EAAU,EAAE,MAAwB,EAAW,EAAE;IAChF,MAAM,KAAK,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC;IAC5E,OAAO,KAAK,CAAC,OAAO,KAAK,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,IAAI,EAAE,KAAK,KAAK,CAAC;AACvE,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,MAAc,EAAE,MAAwB,EAAE,EAAU,EAAE,MAAqB,EAAoB,EAAE;IAC/G,MAAM,OAAO,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;IAC3B,MAAM,IAAI,GAAG,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IAC5B,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC,WAAW,EAAE,IAAI,EAAE,CAAC,CAAC;IAC/D,IAAI,OAAO,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS,IAAI,OAAO,KAAK,SAAS,IAAI,CAAC,SAAS,CAAC,OAAO,EAAE,EAAE,EAAE,MAAM,CAAC;QAAE,OAAO,SAAS,CAAC;IAC9H,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,SAAS,CAAC;IACjF,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IACvD,MAAM,KAAK,GAAG,oBAAoB,CAAC,IAAI,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IACzD,OAAO,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC;AACzH,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,QAAkB,EAAE,IAAU,EAAW,EAAE,CAAC,CAAC;IAChE,IAAI,EAAE,EAAE;IACR,QAAQ,EAAE,SAAS;IACnB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IAC3B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,OAAO,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,OAAO,EAAE,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,EAAE;CACzH,CAAC,CAAC;AAEH,2HAA2H;AAC3H,MAAM,CAAC,MAAM,YAAY,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAChE,MAAM,MAAM,GAAG,eAAe,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;IACtD,MAAM,QAAQ,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;QAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;QACrC,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE,EAAE,EAAE;YACnC,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,CAAC,CAAC;YACpD,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC;QACjE,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IACH,OAAO,QAAQ,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC;AACzD,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,KAAK,CAAC;AAMxB,8HAA8H;AAC9H,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,QAAkB,EAAE,KAAuB,EAAiB,EAAE;IACnG,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,EAAE,EAAE;QAClC,MAAM,IAAI,GAAG,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;QAC5B,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,IAAI,IAAI,CAAC,GAAG,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACpH,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC;YAAE,OAAO,EAAE,CAAC;QAC9C,OAAO,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,EAAE,GAAG,KAAK,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,EAAE,EAAE,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;IACvJ,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,gBAAgB,CAAC;AAEpC,gHAAgH;AAChH,MAAM,CAAC,MAAM,oBAAoB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACxE,MAAM,KAAK,GAAG;QACZ,YAAY,EAAE,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QAC5E,UAAU,EAAE,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;KACpF,CAAC;IACF,OAAO,eAAe,CACpB,GAAG,EACH,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC,EAC/E,OAAO,CAAC,KAAK,CACd,CAAC;AACJ,CAAC,CAAC"}

package/dist/detectors/number-style.d.ts

@@ -0,0 +1,31 @@
+import type { Detector, Sentence } from "../plugin.ts";
+export type NumberKind = "count" | "grouping" | "percent" | "count-small" | "count-large";
+/** One number as written: its kind, the style it is written in, the text, and where it starts in the document. */
+export type NumberMark = {
+    readonly kind: NumberKind;
+    readonly style: string;
+    readonly written: string;
+    readonly offset: number;
+};
+/** The language's words for numbers. countWords is in order from one (the position is the value). */
+export type NumberWords = {
+    readonly counters: readonly string[];
+    readonly kanji: readonly string[];
+    readonly percentUnits: readonly string[];
+    readonly countWords: readonly string[];
+    /** Plural nouns of measure (days, miles) that take figures in every style: a number before one is not a count. */
+    readonly figureNouns: readonly string[];
+};
+/** Every number mark in a sentence, less those inside a quotation. */
+export declare const numberMarksIn: (sentence: Sentence, words: NumberWords, language: string, source: string) => NumberMark[];
+/**
+ * The marks not in the document's usual style, when they are few enough to be slips: at most limitPercent of the kind.
+ * The usual style is the most common one; on a tie, the one the document used first.
+ */
+export declare const minorityStyles: (marks: readonly NumberMark[], limitPercent: number) => {
+    readonly odd: NumberMark[];
+    readonly usual: NumberMark | undefined;
+};
+/** For each kind of number, the marks written in the document's less common style. */
+export declare const numberStyleConsistency: Detector;
+//# sourceMappingURL=number-style.d.ts.map

package/dist/detectors/number-style.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"number-style.d.ts","sourceRoot":"","sources":["../../src/detectors/number-style.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,QAAQ,EAA0B,QAAQ,EAAe,MAAM,cAAc,CAAC;AAE5F,MAAM,MAAM,UAAU,GAAG,OAAO,GAAG,UAAU,GAAG,SAAS,GAAG,aAAa,GAAG,aAAa,CAAC;AAE1F,kHAAkH;AAClH,MAAM,MAAM,UAAU,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,UAAU,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAElI,qGAAqG;AACrG,MAAM,MAAM,WAAW,GAAG;IACxB,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAC;IAClC,QAAQ,CAAC,YAAY,EAAE,SAAS,MAAM,EAAE,CAAC;IACzC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,kHAAkH;IAClH,QAAQ,CAAC,WAAW,EAAE,SAAS,MAAM,EAAE,CAAC;CACzC,CAAC;AAiJF,sEAAsE;AACtE,eAAO,MAAM,aAAa,GAAI,UAAU,QAAQ,EAAE,OAAO,WAAW,EAAE,UAAU,MAAM,EAAE,QAAQ,MAAM,KAAG,UAAU,EAUlH,CAAC;AAIF;;;GAGG;AACH,eAAO,MAAM,cAAc,GAAI,OAAO,SAAS,UAAU,EAAE,EAAE,cAAc,MAAM,KAAG;IAAE,QAAQ,CAAC,GAAG,EAAE,UAAU,EAAE,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,UAAU,GAAG,SAAS,CAAA;CAQvJ,CAAC;AAyBF,sFAAsF;AACtF,eAAO,MAAM,sBAAsB,EAAE,QAUpC,CAAC"}

package/dist/detectors/number-style.js

@@ -0,0 +1,182 @@
+// How a document writes its numbers: 3件 or 三件, 1,000 or 1000, 10% or 10％ or 10パーセント, three files or 3 files.
+// Each kind is compared within the document and the less common style is reported; neither is called right. Pure; the
+// words (counters, kanji numerals, percent units, number words) come from the language's lexicons.
+import { escapeRegExp } from "../orthography.js";
+import { QUOTATION_MARKS, isWithinAny, quotedSpans } from "../quoted-span.js";
+import { isCitation } from "./kutoten-consistency.js";
+const alternation = (words) => words
+    .toSorted((left, right) => right.length - left.length)
+    .map(escapeRegExp)
+    .join("|");
+const DIGIT_RUN = "[0-9０-９]+(?:[,，.．][0-9０-９]+)*";
+// ───────── counts in Japanese: 3件 or 三件 ─────────
+/**
+ * A number right before a counter. Not after a kanji, a kana prolonged mark or 第 (統一, 唯一, 第三回 are words or ordinals),
+ * nor after a letter or digit (part of a code).
+ */
+const countPattern = (words) => {
+    if (words.counters.length === 0 || words.kanji.length === 0)
+        return undefined;
+    const kanji = `[${words.kanji.map(escapeRegExp).join("")}]+`;
+    return new RegExp(`(?<![\\p{Script=Han}A-Za-z0-9０-９.．])(?:(${DIGIT_RUN}[万億]?)|(${kanji}))\\s?(?:${alternation(words.counters)})`, "gu");
+};
+/**
+ * 一 alone is the idiom (一つ, 一人, 一回) a document writes in kanji whatever its style for counts. A run of magnitudes only
+ * (万人, 百人一首, 千人) is as often a word as a number.
+ */
+const MAGNITUDES = /^[一百千万億]+$/u;
+const isIdiomNumeral = (numeral) => MAGNITUDES.test(numeral);
+const countMarks = (text, base, words) => {
+    const pattern = countPattern(words);
+    if (pattern === undefined)
+        return [];
+    return [...text.matchAll(pattern)].flatMap((match) => {
+        const [written, digits, kanji] = match;
+        if (kanji !== undefined && isIdiomNumeral(kanji))
+            return [];
+        return [{ kind: "count", style: digits === undefined ? "kanji" : "digits", written, offset: base + match.index }];
+    });
+};
+// ───────── digit grouping: 1,000 or 1000 ─────────
+const GROUPED = /^[0-9０-９]{1,3}(?:[,，][0-9０-９]{3})+$/u;
+const DIGITS_AND_COMMAS = /[0-9０-９]+(?:[,，][0-9０-９]+)*/gu;
+const PLAIN_BIG = /^[1-9１-９][0-9０-９]{3,}$/u;
+/** Next to one of these, a run of digits is part of a code, a version, a date, a time, a phone number or a decimal. */
+const NOT_BEFORE = /[\p{L}\p{N}.,，．#＃/:：\-－_]/u;
+const NOT_AFTER = /[\p{N}.,，．:：\-－/]/u;
+/** Each number of four digits or more standing alone, grouped (1,000) or not (1000). */
+const bigNumbers = (text) => [...text.matchAll(DIGITS_AND_COMMAS)].flatMap((match) => {
+    const written = match[0];
+    const standsAlone = !NOT_BEFORE.test(text.charAt(match.index - 1)) && !NOT_AFTER.test(text.charAt(match.index + written.length));
+    return standsAlone && (GROUPED.test(written) || PLAIN_BIG.test(written)) ? [{ written, index: match.index }] : [];
+});
+const CURRENCY_BEFORE = /[$¥￥€£]\s?$/u;
+/** A capitalised word or abbreviation right before (H.R. 5376, Form 1040, No. 12345): the number is a label, never grouped. */
+const isLabelBefore = (before) => /\s$/u.test(before) && /^\p{Lu}/u.test(before.trimEnd().split(/\s/u).at(-1) ?? "");
+/** A four-digit number in this range is most often a year (2019 annual report), which is never grouped. */
+const YEAR_LIKE = { from: 1000, to: 2199 };
+const FOUR_DIGITS = 4;
+const isYearLike = (digits) => {
+    const value = Number(digits.normalize("NFKC"));
+    return digits.length === FOUR_DIGITS && value >= YEAR_LIKE.from && value <= YEAR_LIKE.to;
+};
+/** What follows a quantity: a counter or 円 / 万 (Japanese), or a word in lower case (10000 users). */
+/** A counter or 円 / 万 / 億 right after (with at most one space): the number counts something in Japanese. */
+const countedAfter = (after, words) => {
+    const next = after.replace(/^\s/u, "");
+    return /^[万億円]/u.test(next) || words.counters.some((counter) => next.startsWith(counter));
+};
+/** What follows a quantity: a Japanese counter, or a word in lower case (10000 users). */
+const quantityAfter = (after, words) => /^\s?[a-z]/u.test(after) || countedAfter(after, words);
+/** How far back to look for a currency sign or a label word. The sentence may start right after one (H.R. ends a sentence). */
+const BEFORE_REACH = 16;
+const groupingMarks = (text, base, words, source) => bigNumbers(text).flatMap(({ written, index }) => {
+    const grouped = GROUPED.test(written);
+    const before = source.slice(Math.max(0, base + index - BEFORE_REACH), base + index);
+    const after = text.slice(index + written.length);
+    const currency = CURRENCY_BEFORE.test(before);
+    if (!currency && !quantityAfter(after, words))
+        return [];
+    if (!grouped && isLabelBefore(before))
+        return [];
+    if (!grouped && !currency && isYearLike(written) && !countedAfter(after, words))
+        return [];
+    return [{ kind: "grouping", style: grouped ? "grouped" : "plain", written, offset: base + index }];
+});
+// ───────── percent: 10% or 10％ or 10パーセント, 10% or 10 percent ─────────
+const percentMarks = (text, base, words) => {
+    if (words.percentUnits.length === 0)
+        return [];
+    const pattern = new RegExp(`(?<![\\p{L}\\p{N}.．,，])(${DIGIT_RUN})\\s?(${alternation(words.percentUnits)})(?![A-Za-z])`, "gu");
+    return [...text.matchAll(pattern)].map((match) => ({
+        kind: "percent",
+        style: (match[2] ?? "").toLowerCase(),
+        written: match[0],
+        offset: base + match.index,
+    }));
+};
+// ───────── counts in English: three files or 3 files ─────────
+const SMALL_FROM = 2;
+const LARGE_FROM = 10;
+const valueOf = (surface, words) => {
+    if (/^\d+$/u.test(surface))
+        return Number(surface);
+    const at = words.indexOf(surface.toLowerCase());
+    return at === -1 ? undefined : at + 1;
+};
+const isCountedNoun = (token, figureNouns) => token?.pos === "NOUN" && token.features?.["Number"] === "Plur" && !figureNouns.includes(token.surface.toLowerCase());
+/**
+ * A number before a plural noun (three files, 3 files), from two to the last number word the language lists. One is a
+ * pronoun as often as a count. A number word that opens the sentence is spelt out by every style, so it is not counted.
+ */
+const englishCountMarks = (sentence, words) => {
+    const tokens = sentence.tokens ?? [];
+    return tokens.flatMap((token, index) => {
+        const value = valueOf(token.surface, words.countWords);
+        if (value === undefined || value < SMALL_FROM || value > words.countWords.length || index === 0 || !isCountedNoun(tokens[index + 1], words.figureNouns))
+            return [];
+        const style = /^\d+$/u.test(token.surface) ? "digits" : "words";
+        return [{ kind: value < LARGE_FROM ? "count-small" : "count-large", style, written: token.surface, offset: sentence.span.start + token.span.start }];
+    });
+};
+// ───────── the document ─────────
+/** Every number mark in a sentence, less those inside a quotation. */
+export const numberMarksIn = (sentence, words, language, source) => {
+    const base = sentence.span.start;
+    const marks = [
+        ...countMarks(sentence.text, base, words),
+        ...groupingMarks(sentence.text, base, words, source),
+        ...percentMarks(sentence.text, base, words),
+        ...(language === "ja" ? [] : englishCountMarks(sentence, words)),
+    ];
+    const quoted = quotedSpans(sentence.text, QUOTATION_MARKS);
+    return marks.filter((mark) => !isWithinAny(quoted, { start: mark.offset - base, end: mark.offset - base + mark.written.length }));
+};
+const PERCENT = 100;
+/**
+ * The marks not in the document's usual style, when they are few enough to be slips: at most limitPercent of the kind.
+ * The usual style is the most common one; on a tie, the one the document used first.
+ */
+export const minorityStyles = (marks, limitPercent) => {
+    const counts = new Map();
+    marks.forEach((mark) => counts.set(mark.style, (counts.get(mark.style) ?? 0) + 1));
+    if (counts.size < 2)
+        return { odd: [], usual: undefined };
+    const most = Math.max(...counts.values());
+    const usual = marks.find((mark) => counts.get(mark.style) === most);
+    const odd = marks.filter((mark) => mark.style !== usual?.style);
+    return odd.length * PERCENT > limitPercent * marks.length ? { odd: [], usual } : { odd, usual };
+};
+const wordsOf = (doc) => {
+    const patterns = (id) => (doc.lexicons[id] ?? []).map((entry) => entry.pattern);
+    return {
+        counters: patterns("numeral-counter"),
+        kanji: patterns("numeral-kanji"),
+        percentUnits: patterns("percent-unit"),
+        countWords: patterns("count-number"),
+        figureNouns: patterns("figure-noun"),
+    };
+};
+const KINDS = ["count", "grouping", "percent", "count-small", "count-large"];
+const findingOf = (mark, usual, count, of, limit) => ({
+    rule: "number-style-consistency",
+    severity: "info",
+    line: 0,
+    column: 0,
+    quote: mark.written,
+    values: { written: mark.written, usual: usual.written, count, of, limit, offset: mark.offset },
+    variant: mark.kind,
+});
+/** For each kind of number, the marks written in the document's less common style. */
+export const numberStyleConsistency = (doc, options) => {
+    const words = wordsOf(doc);
+    const marks = doc.sentences
+        .filter((sentence) => sentence.embeddedLanguage === undefined && !isCitation(sentence, doc.source))
+        .flatMap((sentence) => numberMarksIn(sentence, words, doc.language, doc.source));
+    return KINDS.flatMap((kind) => {
+        const ofKind = marks.filter((mark) => mark.kind === kind);
+        const { odd, usual } = minorityStyles(ofKind, options.limit);
+        return usual === undefined ? [] : odd.map((mark) => findingOf(mark, usual, odd.length, ofKind.length, options.limit));
+    });
+};
+//# sourceMappingURL=number-style.js.map

package/dist/detectors/number-style.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"number-style.js","sourceRoot":"","sources":["../../src/detectors/number-style.ts"],"names":[],"mappings":"AAAA,6GAA6G;AAC7G,sHAAsH;AACtH,mGAAmG;AACnG,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,eAAe,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAC9E,OAAO,EAAE,UAAU,EAAE,MAAM,0BAA0B,CAAC;AAkBtD,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAU,EAAE,CACvD,KAAK;KACF,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;KACrD,GAAG,CAAC,YAAY,CAAC;KACjB,IAAI,CAAC,GAAG,CAAC,CAAC;AAEf,MAAM,SAAS,GAAG,+BAA+B,CAAC;AAElD,mDAAmD;AAEnD;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,KAAkB,EAAsB,EAAE;IAC9D,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC9E,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC;IAC7D,OAAO,IAAI,MAAM,CAAC,2CAA2C,SAAS,WAAW,KAAK,YAAY,WAAW,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;AAC1I,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,GAAG,aAAa,CAAC;AAEjC,MAAM,cAAc,GAAG,CAAC,OAAe,EAAW,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;AAE9E,MAAM,UAAU,GAAG,CAAC,IAAY,EAAE,IAAY,EAAE,KAAkB,EAAgB,EAAE;IAClF,MAAM,OAAO,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IACpC,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACrC,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAgB,EAAE;QACjE,MAAM,CAAC,OAAO,EAAE,MAAM,EAAE,KAAK,CAAC,GAAG,KAAK,CAAC;QACvC,IAAI,KAAK,KAAK,SAAS,IAAI,cAAc,CAAC,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAC5D,OAAO,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,GAAG,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;IACpH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,oDAAoD;AAEpD,MAAM,OAAO,GAAG,sCAAsC,CAAC;AAEvD,MAAM,iBAAiB,GAAG,+BAA+B,CAAC;AAC1D,MAAM,SAAS,GAAG,yBAAyB,CAAC;AAE5C,uHAAuH;AACvH,MAAM,UAAU,GAAG,4BAA4B,CAAC;AAChD,MAAM,SAAS,GAAG,oBAAoB,CAAC;AAIvC,wFAAwF;AACxF,MAAM,UAAU,GAAG,CAAC,IAAY,EAAc,EAAE,CAC9C,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAc,EAAE;IAClE,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACzB,MAAM,WAAW,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC;IACjI,OAAO,WAAW,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AACpH,CAAC,CAAC,CAAC;AAEL,MAAM,eAAe,GAAG,cAAc,CAAC;AAEvC,+HAA+H;AAC/H,MAAM,aAAa,GAAG,CAAC,MAAc,EAAW,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAEtI,2GAA2G;AAC3G,MAAM,SAAS,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC;AAE3C,MAAM,WAAW,GAAG,CAAC,CAAC;AAEtB,MAAM,UAAU,GAAG,CAAC,MAAc,EAAW,EAAE;IAC7C,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;IAC/C,OAAO,MAAM,CAAC,MAAM,KAAK,WAAW,IAAI,KAAK,IAAI,SAAS,CAAC,IAAI,IAAI,KAAK,IAAI,SAAS,CAAC,EAAE,CAAC;AAC3F,CAAC,CAAC;AAEF,qGAAqG;AACrG,4GAA4G;AAC5G,MAAM,YAAY,GAAG,CAAC,KAAa,EAAE,KAAkB,EAAW,EAAE;IAClE,MAAM,IAAI,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IACvC,OAAO,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC;AAC5F,CAAC,CAAC;AAEF,0FAA0F;AAC1F,MAAM,aAAa,GAAG,CAAC,KAAa,EAAE,KAAkB,EAAW,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,YAAY,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;AAE7H,+HAA+H;AAC/H,MAAM,YAAY,GAAG,EAAE,CAAC;AAExB,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,IAAY,EAAE,KAAkB,EAAE,MAAc,EAAgB,EAAE,CACrG,UAAU,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,EAAgB,EAAE;IAC5D,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,GAAG,KAAK,GAAG,YAAY,CAAC,EAAE,IAAI,GAAG,KAAK,CAAC,CAAC;IACpF,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IACjD,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAC9C,IAAI,CAAC,QAAQ,IAAI,CAAC,aAAa,CAAC,KAAK,EAAE,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACzD,IAAI,CAAC,OAAO,IAAI,aAAa,CAAC,MAAM,CAAC;QAAE,OAAO,EAAE,CAAC;IACjD,IAAI,CAAC,OAAO,IAAI,CAAC,QAAQ,IAAI,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAC3F,OAAO,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,GAAG,KAAK,EAAE,CAAC,CAAC;AACrG,CAAC,CAAC,CAAC;AAEL,wEAAwE;AAExE,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,IAAY,EAAE,KAAkB,EAAgB,EAAE;IACpF,IAAI,KAAK,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAC/C,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,2BAA2B,SAAS,SAAS,WAAW,CAAC,KAAK,CAAC,YAAY,CAAC,eAAe,EAAE,IAAI,CAAC,CAAC;IAC9H,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACjD,IAAI,EAAE,SAAS;QACf,KAAK,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE;QACrC,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;QACjB,MAAM,EAAE,IAAI,GAAG,KAAK,CAAC,KAAK;KAC3B,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,gEAAgE;AAEhE,MAAM,UAAU,GAAG,CAAC,CAAC;AACrB,MAAM,UAAU,GAAG,EAAE,CAAC;AAEtB,MAAM,OAAO,GAAG,CAAC,OAAe,EAAE,KAAwB,EAAsB,EAAE;IAChF,IAAI,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC;IACnD,MAAM,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC;IAChD,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;AACxC,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,KAAwB,EAAE,WAA8B,EAAW,EAAE,CAC1F,KAAK,EAAE,GAAG,KAAK,MAAM,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC,QAAQ,CAAC,KAAK,MAAM,IAAI,CAAC,WAAW,CAAC,QAAQ,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC;AAEvH;;;GAGG;AACH,MAAM,iBAAiB,GAAG,CAAC,QAAkB,EAAE,KAAkB,EAAgB,EAAE;IACjF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAgB,EAAE;QACnD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC;QACvD,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,GAAG,UAAU,IAAI,KAAK,GAAG,KAAK,CAAC,UAAU,CAAC,MAAM,IAAI,KAAK,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,WAAW,CAAC;YACrJ,OAAO,EAAE,CAAC;QACZ,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;QAChE,OAAO,CAAC,EAAE,IAAI,EAAE,KAAK,GAAG,UAAU,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,aAAa,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;IACvJ,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,mCAAmC;AAEnC,sEAAsE;AACtE,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,QAAkB,EAAE,KAAkB,EAAE,QAAgB,EAAE,MAAc,EAAgB,EAAE;IACtH,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;IACjC,MAAM,KAAK,GAAG;QACZ,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,CAAC;QACzC,GAAG,aAAa,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,CAAC;QACpD,GAAG,YAAY,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,CAAC;QAC3C,GAAG,CAAC,QAAQ,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,iBAAiB,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;KACjE,CAAC;IACF,MAAM,MAAM,GAAoB,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,eAAe,CAAC,CAAC;IAC5E,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;AACpI,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,GAAG,CAAC;AAEpB;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,KAA4B,EAAE,YAAoB,EAA0E,EAAE;IAC3J,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IACnF,IAAI,MAAM,CAAC,IAAI,GAAG,CAAC;QAAE,OAAO,EAAE,GAAG,EAAE,EAAE,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;IAC1D,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;IAC1C,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC;IACpE,MAAM,GAAG,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,KAAK,KAAK,EAAE,KAAK,CAAC,CAAC;IAChE,OAAO,GAAG,CAAC,MAAM,GAAG,OAAO,GAAG,YAAY,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC;AAClG,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAe,EAAE;IAClD,MAAM,QAAQ,GAAG,CAAC,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;IAClG,OAAO;QACL,QAAQ,EAAE,QAAQ,CAAC,iBAAiB,CAAC;QACrC,KAAK,EAAE,QAAQ,CAAC,eAAe,CAAC;QAChC,YAAY,EAAE,QAAQ,CAAC,cAAc,CAAC;QACtC,UAAU,EAAE,QAAQ,CAAC,cAAc,CAAC;QACpC,WAAW,EAAE,QAAQ,CAAC,aAAa,CAAC;KACrC,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,KAAK,GAA0B,CAAC,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,aAAa,EAAE,aAAa,CAAC,CAAC;AAEpG,MAAM,SAAS,GAAG,CAAC,IAAgB,EAAE,KAAiB,EAAE,KAAa,EAAE,EAAU,EAAE,KAAa,EAAW,EAAE,CAAC,CAAC;IAC7G,IAAI,EAAE,0BAA0B;IAChC,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,IAAI,CAAC,OAAO;IACnB,MAAM,EAAE,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE;IAC9F,OAAO,EAAE,IAAI,CAAC,IAAI;CACnB,CAAC,CAAC;AAEH,sFAAsF;AACtF,MAAM,CAAC,MAAM,sBAAsB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC1E,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;IAC3B,MAAM,KAAK,GAAG,GAAG,CAAC,SAAS;SACxB,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,gBAAgB,KAAK,SAAS,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;SAClG,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,EAAE,KAAK,EAAE,GAAG,CAAC,QAAQ,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;IACnF,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAC5B,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;QAC1D,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,cAAc,CAAC,MAAM,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC;QAC7D,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC;IACxH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/detectors/number-unit.d.ts

@@ -0,0 +1,35 @@
+import type { Detector } from "../plugin.ts";
+/** One value of a kind: where its number starts, as written, and its unit ("" when bare). */
+export type Measured = {
+    readonly kind: string;
+    readonly name: string;
+    readonly offset: number;
+    readonly written: string;
+    readonly unit: string;
+};
+export type UnitGap = {
+    readonly value: Measured;
+    readonly unit: string;
+};
+/** The lexicons that name units. */
+export declare const UNIT_LEXICONS: readonly ["value-unit", "measure-unit", "percent-unit", "currency-code"];
+/**
+ * The unit of a lone number, signed or not, with a currency mark before it or a unit after it (¥300, 1,200円, -3kg, 45%,
+ * 3 kg): the mark or the unit, "" when the number is bare, undefined when the text is not one number. Read after NFKC, so
+ * full-width digits count. A word after the number is a unit only if a unit lexicon lists it: "3 fixed" is not a value.
+ */
+export declare const unitOf: (text: string, units: ReadonlySet<string>) => string | undefined;
+type LabelLine = {
+    readonly label: string;
+    readonly value: string;
+    readonly at: number;
+};
+/** A line that names a field and gives its value (「予算：300万円」 "**Budget:** $300"), list marker and bold allowed. */
+export declare const labelLine: (line: string, start: number) => LabelLine | undefined;
+/** Every number a table cell or label line gives, by kind. */
+export declare const measuredValues: (source: string, prose: string, units: ReadonlySet<string>) => Measured[];
+/** Bare numbers among values of one kind that otherwise all carry one unit, in document order. */
+export declare const unitGaps: (values: readonly Measured[]) => UnitGap[];
+export declare const numberWithoutUnit: Detector;
+export {};
+//# sourceMappingURL=number-unit.d.ts.map

package/dist/detectors/number-unit.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"number-unit.d.ts","sourceRoot":"","sources":["../../src/detectors/number-unit.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAE,QAAQ,EAA0B,MAAM,cAAc,CAAC;AAMrE,6FAA6F;AAC7F,MAAM,MAAM,QAAQ,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAElJ,MAAM,MAAM,OAAO,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAE1E,oCAAoC;AACpC,eAAO,MAAM,aAAa,0EAA2E,CAAC;AAmBtG;;;;GAIG;AACH,eAAO,MAAM,MAAM,GAAI,MAAM,MAAM,EAAE,OAAO,WAAW,CAAC,MAAM,CAAC,KAAG,MAAM,GAAG,SAS1E,CAAC;AAQF,KAAK,SAAS,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,EAAE,EAAE,MAAM,CAAA;CAAE,CAAC;AAEzF,mHAAmH;AACnH,eAAO,MAAM,SAAS,GAAI,MAAM,MAAM,EAAE,OAAO,MAAM,KAAG,SAAS,GAAG,SAUnE,CAAC;AA2CF,8DAA8D;AAC9D,eAAO,MAAM,cAAc,GAAI,QAAQ,MAAM,EAAE,OAAO,MAAM,EAAE,OAAO,WAAW,CAAC,MAAM,CAAC,KAAG,QAAQ,EAGlG,CAAC;AAYF,kGAAkG;AAClG,eAAO,MAAM,QAAQ,GAAI,QAAQ,SAAS,QAAQ,EAAE,KAAG,OAAO,EAG7D,CAAC;AAKF,eAAO,MAAM,iBAAiB,EAAE,QAQ3B,CAAC"}

package/dist/detectors/number-unit.js

@@ -0,0 +1,121 @@
+import { readMarkdown } from "../markdown-read.js";
+import { spanOf } from "../markdown-node.js";
+import { eachPreOrder } from "../tree-walk.js";
+import { quoteAt } from "./structure-tree.js";
+/** The lexicons that name units. */
+export const UNIT_LEXICONS = ["value-unit", "measure-unit", "percent-unit", "currency-code"];
+const CURRENCY = new Set(["$", "¥", "€", "£"]);
+const SIGN = new Set(["-", "−"]);
+const MULTIPLIER = new Set(["万", "億", "千"]);
+const NUMBER = /^\d[\d,]*(?:\.\d+)?/u;
+/** What may follow a number as its unit: a short run with no digit, space or separator ("1.2.3" and "2026-10-14" are not values). */
+const UNIT_SHAPE = /^[^\d\s.,:|()[\]-]{1,8}$/u;
+/** A label names something in words: 10:30 is a time, not the label "10". */
+const HAS_LETTER = /\p{L}/u;
+const LIST_MARKER = /^\s*(?:[-*+]|\d+[.)])\s+/u;
+const COLON = /[:：]/u;
+const MAX_LABEL = 20;
+const isUnit = (word, units) => units.has(word) || units.has(word.toLowerCase());
+const withoutFirst = (text, marks) => (marks.has(text.charAt(0)) ? text.slice(1).trimStart() : text);
+/**
+ * The unit of a lone number, signed or not, with a currency mark before it or a unit after it (¥300, 1,200円, -3kg, 45%,
+ * 3 kg): the mark or the unit, "" when the number is bare, undefined when the text is not one number. Read after NFKC, so
+ * full-width digits count. A word after the number is a unit only if a unit lexicon lists it: "3 fixed" is not a value.
+ */
+export const unitOf = (text, units) => {
+    const value = text.normalize("NFKC").trim();
+    const currency = CURRENCY.has(value.charAt(0)) ? value.charAt(0) : "";
+    const unsigned = withoutFirst(value.slice(currency.length).trimStart(), SIGN);
+    const number = NUMBER.exec(unsigned)?.[0];
+    if (number === undefined)
+        return undefined;
+    const suffix = withoutFirst(unsigned.slice(number.length).trimStart(), MULTIPLIER);
+    if (suffix === "")
+        return currency;
+    return currency === "" && UNIT_SHAPE.test(suffix) && isUnit(suffix, units) ? suffix : undefined;
+};
+const measured = (kind, name, text, offset, units) => {
+    const written = text.trim();
+    const unit = unitOf(written, units);
+    return unit === undefined ? [] : [{ kind, name, offset: offset + Math.max(0, text.indexOf(written)), written, unit }];
+};
+/** A line that names a field and gives its value (「予算：300万円」 "**Budget:** $300"), list marker and bold allowed. */
+export const labelLine = (line, start) => {
+    const marker = LIST_MARKER.exec(line)?.[0].length ?? 0;
+    const body = line.slice(marker);
+    const colon = body.search(COLON);
+    if (colon <= 0)
+        return undefined;
+    const label = body.slice(0, colon).replaceAll("**", "").trim();
+    const rest = body.slice(colon + 1).replace(/^\*\*/u, "");
+    const value = rest.trim();
+    if (label.length > MAX_LABEL || label.includes("|") || !HAS_LETTER.test(label) || value === "")
+        return undefined;
+    return { label, value, at: start + marker + (body.length - rest.length) + rest.indexOf(value) };
+};
+const cellText = (source, cell) => {
+    const span = cell === undefined ? undefined : spanOf(cell);
+    if (span === undefined)
+        return undefined;
+    const raw = source.slice(span.start, span.end);
+    const lead = raw.search(/[^|\s]/u);
+    return lead === -1 ? undefined : { text: raw.slice(lead).replace(/\|\s*$/u, ""), start: span.start + lead };
+};
+const tableValues = (source, table, index, units) => {
+    const [header, ...body] = table.children ?? [];
+    return (header?.children ?? []).flatMap((headCell, column) => {
+        const name = cellText(source, headCell)?.text.trim() ?? "";
+        return body.flatMap((row) => {
+            const cell = cellText(source, row.children?.[column]);
+            return cell === undefined || name === "" ? [] : measured(`table ${String(index)}: ${name}`, name, cell.text, cell.start, units);
+        });
+    });
+};
+const tablesOf = (source) => {
+    const found = [];
+    eachPreOrder(readMarkdown(source).root, (node) => {
+        if (node.type === "table")
+            found.push(node);
+    });
+    return found;
+};
+/** The lines of a text with where each starts. */
+const linesWithStarts = (text) => text.split("\n").reduce((lines, line) => {
+    const last = lines.at(-1);
+    return [...lines, { line, start: last === undefined ? 0 : last.start + last.line.length + 1 }];
+}, []);
+/** The label lines of the prose (code masked), keyed by the label as written, so one label is one kind across the document. */
+const labelValues = (prose, units) => linesWithStarts(prose).flatMap(({ line, start }) => {
+    const found = labelLine(line, start);
+    return found === undefined ? [] : measured(`label: ${found.label}`, found.label, found.value, found.at, units);
+});
+/** Every number a table cell or label line gives, by kind. */
+export const measuredValues = (source, prose, units) => [
+    ...tablesOf(source).flatMap((table, index) => tableValues(source, table, index, units)),
+    ...labelValues(prose, units),
+];
+/** Bare values must be fewer than those with the unit, so at least two carry it before a bare one reads as missing it. */
+const gapsOfKind = (values) => {
+    const withUnit = values.filter((value) => value.unit !== "");
+    const bare = values.filter((value) => value.unit === "");
+    const units = new Set(withUnit.map((value) => value.unit));
+    const unit = [...units][0];
+    if (unit === undefined || units.size !== 1 || bare.length >= withUnit.length)
+        return [];
+    return bare.map((value) => ({ value, unit }));
+};
+/** Bare numbers among values of one kind that otherwise all carry one unit, in document order. */
+export const unitGaps = (values) => {
+    const kinds = values.reduce((groups, value) => groups.set(value.kind, [...(groups.get(value.kind) ?? []), value]), new Map());
+    return [...kinds.values()].flatMap(gapsOfKind).toSorted((left, right) => left.value.offset - right.value.offset);
+};
+const unitsOf = (doc) => new Set(UNIT_LEXICONS.flatMap((name) => (doc.lexicons[name] ?? []).flatMap((entry) => [entry.pattern, entry.pattern.toLowerCase()])));
+export const numberWithoutUnit = (doc) => unitGaps(measuredValues(doc.source, doc.prose ?? doc.source, unitsOf(doc))).map((gap) => ({
+    rule: "number-without-unit",
+    severity: "warning",
+    line: 0,
+    column: 0,
+    quote: quoteAt(doc.source, gap.value.offset),
+    values: { value: gap.value.written, name: gap.value.name, unit: gap.unit, offset: gap.value.offset },
+}));
+//# sourceMappingURL=number-unit.js.map

package/dist/detectors/number-unit.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"number-unit.js","sourceRoot":"","sources":["../../src/detectors/number-unit.ts"],"names":[],"mappings":"AAIA,OAAO,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;AACnD,OAAO,EAAE,MAAM,EAAqB,MAAM,qBAAqB,CAAC;AAChE,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAO9C,oCAAoC;AACpC,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,YAAY,EAAE,cAAc,EAAE,cAAc,EAAE,eAAe,CAAU,CAAC;AAEtG,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAC/C,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AACjC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAC5C,MAAM,MAAM,GAAG,sBAAsB,CAAC;AACtC,qIAAqI;AACrI,MAAM,UAAU,GAAG,2BAA2B,CAAC;AAE/C,6EAA6E;AAC7E,MAAM,UAAU,GAAG,QAAQ,CAAC;AAC5B,MAAM,WAAW,GAAG,2BAA2B,CAAC;AAChD,MAAM,KAAK,GAAG,OAAO,CAAC;AACtB,MAAM,SAAS,GAAG,EAAE,CAAC;AAErB,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,KAA0B,EAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;AAEvH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,KAA0B,EAAU,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;AAE1I;;;;GAIG;AACH,MAAM,CAAC,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,KAA0B,EAAsB,EAAE;IACrF,MAAM,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;IAC5C,MAAM,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACtE,MAAM,QAAQ,GAAG,YAAY,CAAC,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,SAAS,EAAE,EAAE,IAAI,CAAC,CAAC;IAC9E,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC1C,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,SAAS,EAAE,EAAE,UAAU,CAAC,CAAC;IACnF,IAAI,MAAM,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACnC,OAAO,QAAQ,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC;AAClG,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,IAAY,EAAE,IAAY,EAAE,MAAc,EAAE,KAA0B,EAAc,EAAE;IACpH,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC5B,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IACpC,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;AACxH,CAAC,CAAC;AAIF,mHAAmH;AACnH,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,KAAa,EAAyB,EAAE;IAC9E,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IAChC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACjC,IAAI,KAAK,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IACjC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/D,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;IACzD,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC1B,IAAI,KAAK,CAAC,MAAM,GAAG,SAAS,IAAI,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,KAAK,KAAK,EAAE;QAAE,OAAO,SAAS,CAAC;IACjH,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;AAClG,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAE,IAA8B,EAAiE,EAAE;IACjI,MAAM,IAAI,GAAG,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3D,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACzC,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC;IAC/C,MAAM,IAAI,GAAG,GAAG,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC;IACnC,OAAO,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,GAAG,IAAI,EAAE,CAAC;AAC9G,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,MAAc,EAAE,KAAmB,EAAE,KAAa,EAAE,KAA0B,EAAc,EAAE;IACjH,MAAM,CAAC,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,KAAK,CAAC,QAAQ,IAAI,EAAE,CAAC;IAC/C,OAAO,CAAC,MAAM,EAAE,QAAQ,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE;QAC3D,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC;QAC3D,OAAO,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;YAC1B,MAAM,IAAI,GAAG,QAAQ,CAAC,MAAM,EAAE,GAAG,CAAC,QAAQ,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC;YACtD,OAAO,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,SAAS,MAAM,CAAC,KAAK,CAAC,KAAK,IAAI,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;QAClI,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAkB,EAAE;IAClD,MAAM,KAAK,GAAmB,EAAE,CAAC;IACjC,YAAY,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE;QAC/C,IAAI,IAAI,CAAC,IAAI,KAAK,OAAO;YAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC9C,CAAC,CAAC,CAAC;IACH,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF,kDAAkD;AAClD,MAAM,eAAe,GAAG,CAAC,IAAY,EAAuD,EAAE,CAC5F,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAsD,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;IAC3F,MAAM,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1B,OAAO,CAAC,GAAG,KAAK,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,CAAC;AACjG,CAAC,EAAE,EAAE,CAAC,CAAC;AAET,+HAA+H;AAC/H,MAAM,WAAW,GAAG,CAAC,KAAa,EAAE,KAA0B,EAAc,EAAE,CAC5E,eAAe,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE;IACjD,MAAM,KAAK,GAAG,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IACrC,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,KAAK,CAAC,KAAK,EAAE,EAAE,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC;AACjH,CAAC,CAAC,CAAC;AAEL,8DAA8D;AAC9D,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,MAAc,EAAE,KAAa,EAAE,KAA0B,EAAc,EAAE,CAAC;IACvG,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;IACvF,GAAG,WAAW,CAAC,KAAK,EAAE,KAAK,CAAC;CAC7B,CAAC;AAEF,0HAA0H;AAC1H,MAAM,UAAU,GAAG,CAAC,MAA2B,EAAa,EAAE;IAC5D,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC;IAC7D,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC;IACzD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IAC3D,MAAM,IAAI,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC3B,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,IAAI,QAAQ,CAAC,MAAM;QAAE,OAAO,EAAE,CAAC;IACxF,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAChD,CAAC,CAAC;AAEF,kGAAkG;AAClG,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,MAA2B,EAAa,EAAE;IACjE,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,IAAI,GAAG,EAAsB,CAAC,CAAC;IAClJ,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;AACnH,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAe,EAAE,CAClD,IAAI,GAAG,CAAC,aAAa,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;AAExI,MAAM,CAAC,MAAM,iBAAiB,GAAa,CAAC,GAAkB,EAAa,EAAE,CAC3E,QAAQ,CAAC,cAAc,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;IACxF,IAAI,EAAE,qBAAqB;IAC3B,QAAQ,EAAE,SAAS;IACnB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC;IAC5C,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE;CACrG,CAAC,CAAC,CAAC"}

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

@@ -0,0 +1,8 @@
+import type { Detector } from "../plugin.ts";
+/**
+ * One word written two ways in one document (出来る and できる, 引っ越し and 引越し, ウィンドウ and ウインドウ, e-mail and email),
+ * found without a list of words: the document's own words are grouped by reading or by an evened-out spelling, and the way the
+ * document writes less is pointed at. No way is called right; limit is the largest share, in percent, the minority may have.
+ */
+export declare const orthographicVariant: Detector;
+//# sourceMappingURL=orthographic-variant.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"orthographic-variant.d.ts","sourceRoot":"","sources":["../../src/detectors/orthographic-variant.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA2C,MAAM,cAAc,CAAC;AAwItF;;;;GAIG;AACH,eAAO,MAAM,mBAAmB,EAAE,QAUjC,CAAC"}