chaffjs 0.27.1 → 0.29.0

package/dist/ui-labels.js

@@ -0,0 +1,42 @@
+// The same label on the screen written two ways in one manual (**Start Restore** and **Start restore**, ［復元を開始］ and
+// ［復元開始］). Pure: reads the labels a manual marks as labels, bold or in ［］, and points at the form written less often.
+import { escapeRegExp } from "./orthography.js";
+/** What a manual marks as a label on the screen: bold (** or __), or the full-width brackets Japanese manuals use. */
+const MARKED = /\*\*([^*\n]{1,40})\*\*|__([^_\n]{1,40})__|［([^］\n]{1,40})］/gu;
+const SPACES = /\s+/gu;
+const LETTER = /\p{L}/u;
+/** Each label marked in the source whose text is prose (not covered as code in prose) and holds a letter. */
+export const labelsIn = (source, prose) => [...source.matchAll(MARKED)].flatMap((match) => {
+    const surface = (match[1] ?? match[2] ?? match[3] ?? "").trim();
+    const inner = match.index + match[0].indexOf(surface);
+    const covered = prose.slice(inner, inner + surface.length) !== surface;
+    return covered || !LETTER.test(surface) ? [] : [{ surface, offset: inner }];
+});
+/**
+ * Two forms of one label differ only in case, spaces, or a word a label is written with or without (the label-particle
+ * lexicon: 復元を開始 and 復元開始).
+ */
+const keyOf = (surface, particles) => particles.reduce((text, particle) => text.replace(joining(particle), ""), surface.replace(SPACES, "")).toLowerCase();
+/** A particle between two kanji or katakana (復元を開始), not one inside a word (その他). */
+const joining = (particle) => new RegExp(`(?<=[\\p{Script=Han}\\p{Script=Katakana}ー])${escapeRegExp(particle)}(?=[\\p{Script=Han}\\p{Script=Katakana}])`, "gu");
+/** Two forms that differ only in the case of their first letter (Note and note): emphasis at a sentence's start, not two labels. */
+const onlyFirstLetterDiffers = (forms) => new Set(forms.map((form) => `${form.charAt(0).toLowerCase()}${form.slice(1)}`)).size === 1;
+const groupBy = (labels, keyFor) => labels.reduce((groups, label) => groups.set(keyFor(label), [...(groups.get(keyFor(label)) ?? []), label]), new Map());
+/**
+ * Each label written in a form other than the one its key is written in most; on a tie, the later form. A label written
+ * one way only says nothing.
+ */
+export const labelVariants = (labels, particles = []) => {
+    const byKey = groupBy(labels, (label) => keyOf(label.surface, particles));
+    return [...byKey.values()].flatMap((group) => {
+        const counts = groupBy(group, (label) => label.surface);
+        if (counts.size < 2 || onlyFirstLetterDiffers([...counts.keys()]))
+            return [];
+        const [first] = counts;
+        if (first === undefined)
+            return [];
+        const usual = [...counts].reduce((best, entry) => (entry[1].length > best[1].length ? entry : best), first)[0];
+        return group.filter((label) => label.surface !== usual).map((label) => ({ label, usual }));
+    });
+};
+//# sourceMappingURL=ui-labels.js.map

package/dist/ui-labels.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ui-labels.js","sourceRoot":"","sources":["../src/ui-labels.ts"],"names":[],"mappings":"AAAA,oHAAoH;AACpH,sHAAsH;AAEtH,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAKhD,sHAAsH;AACtH,MAAM,MAAM,GAAG,8DAA8D,CAAC;AAC9E,MAAM,MAAM,GAAG,OAAO,CAAC;AACvB,MAAM,MAAM,GAAG,QAAQ,CAAC;AAExB,6GAA6G;AAC7G,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAE,KAAa,EAAW,EAAE,CACjE,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IAC7C,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IAChE,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IACtD,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,KAAK,OAAO,CAAC;IACvE,OAAO,OAAO,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;AAC9E,CAAC,CAAC,CAAC;AAEL;;;GAGG;AACH,MAAM,KAAK,GAAG,CAAC,OAAe,EAAE,SAA4B,EAAU,EAAE,CACtE,SAAS,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC,EAAE,OAAO,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;AAEvH,qFAAqF;AACrF,MAAM,OAAO,GAAG,CAAC,QAAgB,EAAU,EAAE,CAC3C,IAAI,MAAM,CAAC,8CAA8C,YAAY,CAAC,QAAQ,CAAC,2CAA2C,EAAE,IAAI,CAAC,CAAC;AAEpI,oIAAoI;AACpI,MAAM,sBAAsB,GAAG,CAAC,KAAwB,EAAW,EAAE,CACnE,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC;AAE7F,MAAM,OAAO,GAAG,CAAC,MAAwB,EAAE,MAAgC,EAAwB,EAAE,CACnG,MAAM,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,IAAI,GAAG,EAAmB,CAAC,CAAC;AAEzI;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,MAAwB,EAAE,YAA+B,EAAE,EAAkB,EAAE;IAC3G,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,CAAC;IAC1E,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;QAC3C,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QACxD,IAAI,MAAM,CAAC,IAAI,GAAG,CAAC,IAAI,sBAAsB,CAAC,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;YAAE,OAAO,EAAE,CAAC;QAC7E,MAAM,CAAC,KAAK,CAAC,GAAG,MAAM,CAAC;QACvB,IAAI,KAAK,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACnC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAC/G,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,KAAK,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;IAC7F,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/undefined-call.d.ts

@@ -0,0 +1,13 @@
+/** A call to a name the page does not define, where it is written, and the documented name it is likely a slip of. */
+export type UndefinedCall = {
+    readonly name: string;
+    readonly offset: number;
+    readonly documented: string;
+};
+/**
+ * Every call in a code block to a name written nowhere else on the page, when a documented name is its stem
+ * (wrap → wrapText). A name used twice (defined and called, or called twice) is the example's own. headings are the page's
+ * Markdown headings; globals the language functions every page may call.
+ */
+export declare const undefinedCalls: (source: string, headings: readonly string[], globals?: readonly string[]) => UndefinedCall[];
+//# sourceMappingURL=undefined-call.d.ts.map

package/dist/undefined-call.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"undefined-call.d.ts","sourceRoot":"","sources":["../src/undefined-call.ts"],"names":[],"mappings":"AAKA,sHAAsH;AACtH,MAAM,MAAM,aAAa,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;CAAE,CAAC;AAuC5G;;;;GAIG;AACH,eAAO,MAAM,cAAc,GAAI,QAAQ,MAAM,EAAE,UAAU,SAAS,MAAM,EAAE,EAAE,UAAS,SAAS,MAAM,EAAO,KAAG,aAAa,EAc1H,CAAC"}

package/dist/undefined-call.js

@@ -0,0 +1,55 @@
+// A function called in a code example that the page never defines, where the page documents one whose name it extends:
+// wrapText( on a page about wrap, wrapLines( after import { wrap }. Pure; reads the Markdown source. The global functions
+// of programming languages (parseInt, setTimeout) come from a lexicon, so a page about parse may call parseInt.
+import { codeFences } from "./detectors/code-fences.js";
+const CALL = /(?<![\w$.])([A-Za-z_$][\w$]*)[ \t]*\(/gu;
+const WORD = /[A-Za-z_$][\w$]*/gu;
+const LEADING_WORD = /^[A-Za-z_$][\w$]*/u;
+/** import { wrap, type Options, measure as m } from "wrapkit" (JavaScript) and from wrapkit import wrap, measure (Python). */
+const IMPORT_BRACES = /\bimport[ \t]*\{([^}]*)\}/gu;
+const FROM_IMPORT = /\bfrom[ \t]+[\w.]+[ \t]+import[ \t]+([\w, \t]+)/gu;
+const TYPE_MODIFIER = /^type[ \t]+/u;
+/** A name right after one of these is being declared, not called: function wrapText(, def wrap_text(. */
+const DECLARING = /\b(?:function\*?|def|fn|func|class|new)[ \t]+$/u;
+/** Where a line comment starts: a call after it is commented out. */
+const LINE_COMMENT = /\/\/|(?:^|[ \t])#/u;
+/** The shortest documented name a call is compared with: shorter ones (id, get) are prefixes of too many names. */
+const MIN_DOCUMENTED = 4;
+/** Each specifier's own name: `type Options` is Options, `measure as m` is measure. */
+const specifiers = (list) => list.split(",").flatMap((specifier) => LEADING_WORD.exec(specifier.trim().replace(TYPE_MODIFIER, ""))?.[0] ?? []);
+const importedNames = (source) => [...source.matchAll(IMPORT_BRACES), ...source.matchAll(FROM_IMPORT)].flatMap((match) => specifiers(match[1] ?? ""));
+/** Names the page documents: the first word of a heading (## wrap, ### `wrap()`) and the names it imports. */
+const documentedNames = (source, headings) => {
+    const fromHeadings = headings.flatMap((heading) => LEADING_WORD.exec(heading.replace(/[`*]/gu, "").trim())?.[0] ?? []);
+    return new Set([...fromHeadings, ...importedNames(source)].filter((name) => name.length >= MIN_DOCUMENTED));
+};
+const countWords = (source) => [...source.matchAll(WORD)].reduce((counts, match) => counts.set(match[0], (counts.get(match[0]) ?? 0) + 1), new Map());
+/** Whether the call at `at` in `code` is declared there or commented out. */
+const isDeclaredOrCommented = (code, at) => {
+    const lineStart = code.lastIndexOf("\n", at - 1) + 1;
+    const before = code.slice(lineStart, at);
+    return DECLARING.test(before) || LINE_COMMENT.test(before);
+};
+/**
+ * Every call in a code block to a name written nowhere else on the page, when a documented name is its stem
+ * (wrap → wrapText). A name used twice (defined and called, or called twice) is the example's own. headings are the page's
+ * Markdown headings; globals the language functions every page may call.
+ */
+export const undefinedCalls = (source, headings, globals = []) => {
+    const documented = documentedNames(source, headings);
+    if (documented.size === 0)
+        return [];
+    const counts = countWords(source);
+    const known = new Set([...documented, ...globals]);
+    return codeFences(source).flatMap((fence) => {
+        const code = source.slice(fence.start, fence.end);
+        return [...code.matchAll(CALL)].flatMap((match) => {
+            const name = match[1] ?? "";
+            const stem = [...documented].find((candidate) => name !== candidate && name.startsWith(candidate));
+            if (stem === undefined || known.has(name) || (counts.get(name) ?? 0) > 1 || isDeclaredOrCommented(code, match.index))
+                return [];
+            return [{ name, offset: fence.start + match.index, documented: stem }];
+        });
+    });
+};
+//# sourceMappingURL=undefined-call.js.map

package/dist/undefined-call.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"undefined-call.js","sourceRoot":"","sources":["../src/undefined-call.ts"],"names":[],"mappings":"AAAA,uHAAuH;AACvH,0HAA0H;AAC1H,gHAAgH;AAChH,OAAO,EAAE,UAAU,EAAE,MAAM,4BAA4B,CAAC;AAKxD,MAAM,IAAI,GAAG,yCAAyC,CAAC;AACvD,MAAM,IAAI,GAAG,oBAAoB,CAAC;AAClC,MAAM,YAAY,GAAG,oBAAoB,CAAC;AAC1C,8HAA8H;AAC9H,MAAM,aAAa,GAAG,6BAA6B,CAAC;AACpD,MAAM,WAAW,GAAG,mDAAmD,CAAC;AACxE,MAAM,aAAa,GAAG,cAAc,CAAC;AACrC,yGAAyG;AACzG,MAAM,SAAS,GAAG,iDAAiD,CAAC;AACpE,qEAAqE;AACrE,MAAM,YAAY,GAAG,oBAAoB,CAAC;AAC1C,mHAAmH;AACnH,MAAM,cAAc,GAAG,CAAC,CAAC;AAEzB,uFAAuF;AACvF,MAAM,UAAU,GAAG,CAAC,IAAY,EAAY,EAAE,CAC5C,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAEpH,MAAM,aAAa,GAAG,CAAC,MAAc,EAAY,EAAE,CACjD,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,aAAa,CAAC,EAAE,GAAG,MAAM,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;AAEtH,8GAA8G;AAC9G,MAAM,eAAe,GAAG,CAAC,MAAc,EAAE,QAA2B,EAAe,EAAE;IACnF,MAAM,YAAY,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IACvH,OAAO,IAAI,GAAG,CAAC,CAAC,GAAG,YAAY,EAAE,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,IAAI,cAAc,CAAC,CAAC,CAAC;AAC9G,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,MAAc,EAAuB,EAAE,CACzD,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,GAAG,EAAkB,CAAC,CAAC;AAEzI,6EAA6E;AAC7E,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,EAAU,EAAW,EAAE;IAClE,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IACrD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC;IACzC,OAAO,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC7D,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,MAAc,EAAE,QAA2B,EAAE,UAA6B,EAAE,EAAmB,EAAE;IAC9H,MAAM,UAAU,GAAG,eAAe,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;IACrD,IAAI,UAAU,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACrC,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,UAAU,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC;IACnD,OAAO,UAAU,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;QAC1C,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;QAClD,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;YAChD,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC5B,MAAM,IAAI,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,CAAC;YACnG,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,qBAAqB,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC;gBAAE,OAAO,EAAE,CAAC;YAChI,OAAO,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC;QACzE,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/unnamed-speaker.d.ts

@@ -0,0 +1,23 @@
+import type { LexiconEntry } from "./plugin.ts";
+export type SpeakerSentence = {
+    readonly start: number;
+    readonly text: string;
+};
+export type UnnamedSpeakerQuote = {
+    readonly offset: number;
+    readonly quote: string;
+    readonly speaker: string;
+};
+/** The word list unnamed-speaker, by group, and quote-attribution with the list's own cues (group cue), longest first. */
+export type SpeakerWords = {
+    readonly roles: readonly string[];
+    readonly afterRole: readonly string[];
+    readonly honorifics: readonly string[];
+    readonly anaphors: readonly string[];
+    readonly afterQuote: readonly string[];
+    readonly cues: readonly LexiconEntry[];
+};
+export declare const speakerWords: (speaker: readonly LexiconEntry[], cues: readonly LexiconEntry[]) => SpeakerWords;
+/** Each quotation in the sentence given to a role with no name. */
+export declare const unnamedSpeakerQuotes: (sentence: SpeakerSentence, words: SpeakerWords) => UnnamedSpeakerQuote[];
+//# sourceMappingURL=unnamed-speaker.d.ts.map

package/dist/unnamed-speaker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"unnamed-speaker.d.ts","sourceRoot":"","sources":["../src/unnamed-speaker.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,YAAY,EAAQ,MAAM,aAAa,CAAC;AAEtD,MAAM,MAAM,eAAe,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAEhF,MAAM,MAAM,mBAAmB,GAAG;IAAE,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAAC;AAEhH,0HAA0H;AAC1H,MAAM,MAAM,YAAY,GAAG;IACzB,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAC;IAClC,QAAQ,CAAC,SAAS,EAAE,SAAS,MAAM,EAAE,CAAC;IACtC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,IAAI,EAAE,SAAS,YAAY,EAAE,CAAC;CACxC,CAAC;AAEF,eAAO,MAAM,YAAY,GAAI,SAAS,SAAS,YAAY,EAAE,EAAE,MAAM,SAAS,YAAY,EAAE,KAAG,YAY9F,CAAC;AAsLF,mEAAmE;AACnE,eAAO,MAAM,oBAAoB,GAAI,UAAU,eAAe,EAAE,OAAO,YAAY,KAAG,mBAAmB,EAQxG,CAAC"}

package/dist/unnamed-speaker.js

@@ -0,0 +1,176 @@
+// A quotation whose speaker, in the words right beside it, is a role and not a name (業界のアナリストは「…」と述べた,
+// "…," said an analyst). The words come from the language's word lists: unnamed-speaker (roles, what may follow a role,
+// titles, words pointing back, the quoting particle) and quote-attribution (said, according to, と述べ). Pure.
+import { QUOTATION_MARKS, quotedSpans } from "./quoted-span.js";
+export const speakerWords = (speaker, cues) => {
+    const group = (name) => speaker.filter((entry) => entry.group === name && entry.pattern !== "").map((entry) => entry.pattern);
+    return {
+        roles: group("role"),
+        afterRole: group("after-role"),
+        honorifics: group("honorific"),
+        anaphors: group("anaphor"),
+        afterQuote: group("after-quote"),
+        cues: [...cues, ...speaker.filter((entry) => entry.group === "cue")]
+            .filter((cue) => cue.pattern !== "")
+            .toSorted((left, right) => right.pattern.length - left.pattern.length),
+    };
+};
+/** Shorter than this, a quotation is a term or a label (「保存」, "Save"), not someone's words. */
+const MIN_QUOTE_CHARS = 10;
+const LATIN = /\p{Script=Latin}/u;
+const LETTER = /\p{L}/u;
+const UPPER_START = /^\p{Lu}/u;
+const LOWER_START = /^\p{Ll}/u;
+const WORD = /[\p{L}\p{N}'’-]+/gu;
+/** Words, and the commas between them: a comma ends a run of capitalised words (an analyst at Acme, Hanako Mori). */
+const WORD_OR_COMMA = /[\p{L}\p{N}'’-]+|[,，]/gu;
+/** A Japanese title is not one when a kanji follows it (氏名, 様々). */
+const JOINED_KANJI = /^[\p{Script=Han}々]/u;
+/** How far from the quotation the speaker is looked for, on each side: keeps a long sentence of many quotations linear. */
+const SPEAKER_REACH = 200;
+/** Where the clause holding the speaker before a quotation starts: after these marks. */
+const CLAUSE_MARKS_BEFORE = /[。．！？!?;；「」『』“”"、，]/u;
+/** Where the speaker's phrase after a quotation ends. A comma does not end it in English (said Hanako Mori, an analyst). */
+const CLAUSE_MARKS_AFTER = /[。．！？.!?;；「」『』“”"]/u;
+/** Where a phrase after the quoting particle (「…」と業界関係者は話す) ends. */
+const CLAUSE_MARKS_PARTICLE = /[。．！？.!?;；「」『』“”"、，]/u;
+/** Where a speaker standing before its cue ("…," an analyst said) must not reach past. */
+const CLAUSE_MARKS_INNER = /[,，、。．！？.!?;；:：]/u;
+/** Spaces and marks that join a speaker to a quotation or a cue, and are not part of either. */
+const GLUE = /[\s,，、:：—–-]/u;
+const PHRASE_END = /^[,，、。．.;；:：)）]/u;
+/** A Latin word is matched whole and without case; a Japanese one as written, inside any text. */
+const matchesAt = (text, at, word) => {
+    if (!LATIN.test(word))
+        return text.startsWith(word, at);
+    return (text.slice(at, at + word.length).toLowerCase() === word.toLowerCase() && !LETTER.test(text.charAt(at - 1)) && !LETTER.test(text.charAt(at + word.length)));
+};
+const positionsOf = (text, word) => word === "" ? [] : [...Array(Math.max(0, text.length - word.length + 1)).keys()].filter((at) => matchesAt(text, at, word));
+const startsWithWord = (text, word) => matchesAt(text, 0, word);
+const endsWithWord = (text, word) => text.length >= word.length && matchesAt(text, text.length - word.length, word);
+const afterLast = (text, marks) => {
+    const cut = [...text].reduce((last, char, index) => (marks.test(char) ? index : last), -1);
+    return [...text].slice(cut + 1).join("");
+};
+const beforeFirst = (text, marks) => {
+    const at = [...text].findIndex((char) => marks.test(char));
+    return at < 0 ? text : [...text].slice(0, at).join("");
+};
+/** The text without glue at either end. */
+const clean = (text) => {
+    const chars = [...text];
+    const first = chars.findIndex((char) => !GLUE.test(char));
+    return first < 0 ? "" : chars.slice(first, chars.findLastIndex((char) => !GLUE.test(char)) + 1).join("");
+};
+/** The clause before the quotation: the speaker, with a cue at its end (An analyst said, 関係者によれば) or before it (according to). */
+const speakerBefore = (before, words) => {
+    const clause = clean(afterLast(clean(before), CLAUSE_MARKS_BEFORE));
+    const cues = words.cues.filter((cue) => cue.position !== "after");
+    const ending = cues.find((cue) => endsWithWord(clause, cue.pattern));
+    if (ending !== undefined)
+        return { speaker: clean(clause.slice(0, clause.length - ending.pattern.length)), attributed: true };
+    const leads = cues
+        .filter((cue) => cue.position === "before")
+        .flatMap((cue) => positionsOf(clause, cue.pattern).map((at) => at + cue.pattern.length))
+        .toSorted((left, right) => right - left);
+    const lead = leads[0];
+    if (lead !== undefined)
+        return { speaker: clean(clause.slice(lead)), attributed: true };
+    return { speaker: clause, attributed: false };
+};
+/** The words after the quotation: a quoting particle (「…」と業界関係者は), a cue then the speaker (said an analyst), or the speaker then a cue. */
+const speakerAfter = (after, words) => {
+    const rest = clean(after);
+    const particle = words.afterQuote.find((word) => startsWithWord(rest, word));
+    if (particle !== undefined)
+        return { speaker: clean(beforeFirst(clean(rest.slice(particle.length)), CLAUSE_MARKS_PARTICLE)), attributed: true, leavesSpeakerBefore: true };
+    const cues = words.cues.filter((cue) => cue.position !== "before");
+    const leading = cues.find((cue) => startsWithWord(rest, cue.pattern));
+    if (leading !== undefined) {
+        const speaker = clean(beforeFirst(rest.slice(leading.pattern.length), CLAUSE_MARKS_AFTER));
+        return { speaker, attributed: true, leavesSpeakerBefore: speaker === "" };
+    }
+    const clause = beforeFirst(rest, CLAUSE_MARKS_INNER);
+    const cueAt = cues.flatMap((cue) => positionsOf(clause, cue.pattern)).toSorted((left, right) => left - right)[0];
+    if (cueAt !== undefined)
+        return { speaker: clean(clause.slice(0, cueAt)), attributed: true, leavesSpeakerBefore: cueAt === 0 };
+    return { speaker: "", attributed: false };
+};
+const isCapitalised = (word) => word !== undefined && UPPER_START.test(word);
+const isRoleWord = (word, words) => word !== undefined && words.roles.some((role) => LATIN.test(role) && role.toLowerCase() === word.toLowerCase());
+const isAfterRoleWord = (word, words) => word !== undefined && words.afterRole.some((entry) => entry.toLowerCase() === word.toLowerCase());
+/**
+ * A run of capitalised words that is a person's name (Hanako Mori, analyst Hanako Mori). Not one: a run after a word like
+ * "at" (an analyst at Minato Research), a run right before a role (an Acme spokesperson), or, when the phrase opens the
+ * clause, its first word alone ("An analyst", "Experts"). After a cue it is a name: said Mori, an analyst.
+ */
+const holdsCapitalisedName = (phrase, words, opensSentence) => {
+    const tokens = [...phrase.matchAll(WORD_OR_COMMA)].map((match) => match[0]);
+    return tokens.some((token, index) => {
+        if (!isCapitalised(token) || isCapitalised(tokens[index - 1]))
+            return false;
+        const runEnd = tokens.findIndex((later, at) => at > index && !isCapitalised(later));
+        const end = runEnd < 0 ? tokens.length : runEnd;
+        if (isAfterRoleWord(tokens[index - 1], words) || isRoleWord(tokens[end], words))
+            return false;
+        const next = tokens[1];
+        return !(opensSentence && index === 0 && end === 1 && (next === undefined || LOWER_START.test(next)));
+    });
+};
+const holdsTitle = (phrase, title) => positionsOf(phrase, title).some((at) => LATIN.test(title) || !JOINED_KANJI.test(phrase.slice(at + title.length)));
+const isNamed = (phrase, words, opensSentence) => words.honorifics.some((title) => holdsTitle(phrase, title)) || holdsCapitalisedName(phrase, words, opensSentence);
+const lastWord = (text) => [...text.matchAll(WORD)].map((match) => match[0]).at(-1) ?? "";
+/** Whether the role at this place is the head of the phrase: nothing after it, or only a word like は or "at". */
+const headsPhrase = (phrase, at, role, words) => {
+    const after = phrase.slice(at + role.length).trimStart();
+    return after === "" || PHRASE_END.test(after) || words.afterRole.some((word) => startsWithWord(after, word));
+};
+const pointsBack = (phrase, at, words) => {
+    const before = phrase.slice(0, at).trimEnd();
+    return words.anaphors.some((word) => (LATIN.test(word) ? lastWord(before).toLowerCase() === word.toLowerCase() : before.endsWith(word)));
+};
+/**
+ * The phrase as the message shows it, when a role heads it and does not point back at someone named before (同アナリスト,
+ * the analyst). A Japanese phrase is shown up to its role (業界関係者は話す as 業界関係者); an English one whole.
+ */
+const unnamedRoleIn = (phrase, words) => {
+    const heads = words.roles.flatMap((role) => positionsOf(phrase, role)
+        .filter((at) => headsPhrase(phrase, at, role, words) && !pointsBack(phrase, at, words))
+        .map((at) => (LATIN.test(role) ? phrase : phrase.slice(0, at + role.length))));
+    return heads[0];
+};
+/** The phrases that may be the speaker: the attributed side's, and the one before when a cue after leaves the speaker there. */
+const speakerPhrases = (before, after, beforeOpens) => {
+    const useBefore = before.attributed || (after.attributed && after.leavesSpeakerBefore === true);
+    return [
+        ...(useBefore ? [{ text: before.speaker, opensSentence: beforeOpens }] : []),
+        ...(after.attributed ? [{ text: after.speaker, opensSentence: false }] : []),
+    ].filter((phrase) => phrase.text !== "");
+};
+/** The phrase that gives this quotation to a role and to no name, if any. */
+const unnamedSpeakerOf = (text, span, words) => {
+    const reach = Math.max(0, span.start - 1 - SPEAKER_REACH);
+    const before = speakerBefore(text.slice(reach, span.start - 1), words);
+    const after = speakerAfter(text.slice(span.end + 1, span.end + 1 + SPEAKER_REACH), words);
+    const phrases = speakerPhrases(before, after, reach === 0 &&
+        text
+            .slice(0, span.start - 1)
+            .trim()
+            .startsWith(before.speaker));
+    if (phrases.some((phrase) => isNamed(phrase.text, words, phrase.opensSentence)))
+        return undefined;
+    return phrases.map((phrase) => unnamedRoleIn(phrase.text, words)).find((speaker) => speaker !== undefined);
+};
+/** Each quotation in the sentence given to a role with no name. */
+export const unnamedSpeakerQuotes = (sentence, words) => {
+    if (words.roles.length === 0)
+        return [];
+    return quotedSpans(sentence.text, QUOTATION_MARKS).flatMap((span) => {
+        const quote = clean(sentence.text.slice(span.start, span.end));
+        if ([...quote].length < MIN_QUOTE_CHARS)
+            return [];
+        const speaker = unnamedSpeakerOf(sentence.text, span, words);
+        return speaker === undefined ? [] : [{ offset: sentence.start + span.start, quote, speaker }];
+    });
+};
+//# sourceMappingURL=unnamed-speaker.js.map

package/dist/unnamed-speaker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"unnamed-speaker.js","sourceRoot":"","sources":["../src/unnamed-speaker.ts"],"names":[],"mappings":"AAAA,uGAAuG;AACvG,wHAAwH;AACxH,4GAA4G;AAC5G,OAAO,EAAE,eAAe,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAiBhE,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,OAAgC,EAAE,IAA6B,EAAgB,EAAE;IAC5G,MAAM,KAAK,GAAG,CAAC,IAAY,EAAY,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,IAAI,IAAI,KAAK,CAAC,OAAO,KAAK,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAChJ,OAAO;QACL,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC;QACpB,SAAS,EAAE,KAAK,CAAC,YAAY,CAAC;QAC9B,UAAU,EAAE,KAAK,CAAC,WAAW,CAAC;QAC9B,QAAQ,EAAE,KAAK,CAAC,SAAS,CAAC;QAC1B,UAAU,EAAE,KAAK,CAAC,aAAa,CAAC;QAChC,IAAI,EAAE,CAAC,GAAG,IAAI,EAAE,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;aACjE,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,KAAK,EAAE,CAAC;aACnC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC;KACzE,CAAC;AACJ,CAAC,CAAC;AAEF,+FAA+F;AAC/F,MAAM,eAAe,GAAG,EAAE,CAAC;AAE3B,MAAM,KAAK,GAAG,mBAAmB,CAAC;AAClC,MAAM,MAAM,GAAG,QAAQ,CAAC;AACxB,MAAM,WAAW,GAAG,UAAU,CAAC;AAC/B,MAAM,WAAW,GAAG,UAAU,CAAC;AAC/B,MAAM,IAAI,GAAG,oBAAoB,CAAC;AAClC,qHAAqH;AACrH,MAAM,aAAa,GAAG,yBAAyB,CAAC;AAChD,oEAAoE;AACpE,MAAM,YAAY,GAAG,qBAAqB,CAAC;AAC3C,2HAA2H;AAC3H,MAAM,aAAa,GAAG,GAAG,CAAC;AAC1B,yFAAyF;AACzF,MAAM,mBAAmB,GAAG,sBAAsB,CAAC;AACnD,4HAA4H;AAC5H,MAAM,kBAAkB,GAAG,qBAAqB,CAAC;AACjD,qEAAqE;AACrE,MAAM,qBAAqB,GAAG,uBAAuB,CAAC;AACtD,0FAA0F;AAC1F,MAAM,kBAAkB,GAAG,mBAAmB,CAAC;AAC/C,gGAAgG;AAChG,MAAM,IAAI,GAAG,eAAe,CAAC;AAC7B,MAAM,UAAU,GAAG,kBAAkB,CAAC;AAEtC,kGAAkG;AAClG,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,EAAU,EAAE,IAAY,EAAW,EAAE;IACpE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACxD,OAAO,CACL,IAAI,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,WAAW,EAAE,KAAK,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAC1J,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,IAAY,EAAY,EAAE,CAC3D,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,EAAE,EAAE,IAAI,CAAC,CAAC,CAAC;AAE7H,MAAM,cAAc,GAAG,CAAC,IAAY,EAAE,IAAY,EAAW,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC;AAEzF,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,IAAY,EAAW,EAAE,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,IAAI,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;AAE7I,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE;IACxD,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC3F,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC3C,CAAC,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE;IAC1D,MAAM,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAC3D,OAAO,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACzD,CAAC,CAAC;AAEF,2CAA2C;AAC3C,MAAM,KAAK,GAAG,CAAC,IAAY,EAAU,EAAE;IACrC,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,CAAC;IACxB,MAAM,KAAK,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAC1D,OAAO,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC3G,CAAC,CAAC;AAKF,kIAAkI;AAClI,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,KAAmB,EAAQ,EAAE;IAClE,MAAM,MAAM,GAAG,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,mBAAmB,CAAC,CAAC,CAAC;IACpE,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,QAAQ,KAAK,OAAO,CAAC,CAAC;IAClE,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,YAAY,CAAC,MAAM,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;IACrE,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC;IAC9H,MAAM,KAAK,GAAG,IAAI;SACf,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,QAAQ,KAAK,QAAQ,CAAC;SAC1C,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;SACvF,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC;IAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACtB,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC;IACxF,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,CAAC;AAChD,CAAC,CAAC;AAEF,2IAA2I;AAC3I,MAAM,YAAY,GAAG,CAAC,KAAa,EAAE,KAAmB,EAAQ,EAAE;IAChE,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;IAC1B,MAAM,QAAQ,GAAG,KAAK,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;IAC7E,IAAI,QAAQ,KAAK,SAAS;QACxB,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,EAAE,qBAAqB,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,mBAAmB,EAAE,IAAI,EAAE,CAAC;IACjJ,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC;IACnE,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;IACtE,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;QAC1B,MAAM,OAAO,GAAG,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,kBAAkB,CAAC,CAAC,CAAC;QAC3F,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,mBAAmB,EAAE,OAAO,KAAK,EAAE,EAAE,CAAC;IAC5E,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,EAAE,kBAAkB,CAAC,CAAC;IACrD,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IACjH,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,EAAE,OAAO,EAAE,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,mBAAmB,EAAE,KAAK,KAAK,CAAC,EAAE,CAAC;IAC/H,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,UAAU,EAAE,KAAK,EAAE,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,IAAwB,EAAW,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAE1G,MAAM,UAAU,GAAG,CAAC,IAAwB,EAAE,KAAmB,EAAW,EAAE,CAC5E,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,WAAW,EAAE,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;AAElH,MAAM,eAAe,GAAG,CAAC,IAAwB,EAAE,KAAmB,EAAW,EAAE,CACjF,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,WAAW,EAAE,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;AAEpG;;;;GAIG;AACH,MAAM,oBAAoB,GAAG,CAAC,MAAc,EAAE,KAAmB,EAAE,aAAsB,EAAW,EAAE;IACpG,MAAM,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5E,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAClC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,IAAI,aAAa,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QAC5E,MAAM,MAAM,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,KAAK,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC;QACpF,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;QAChD,IAAI,eAAe,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,IAAI,UAAU,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC;QAC9F,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QACvB,OAAO,CAAC,CAAC,aAAa,IAAI,KAAK,KAAK,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACxG,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,KAAa,EAAW,EAAE,CAC5D,WAAW,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;AAEpH,MAAM,OAAO,GAAG,CAAC,MAAc,EAAE,KAAmB,EAAE,aAAsB,EAAW,EAAE,CACvF,KAAK,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,IAAI,oBAAoB,CAAC,MAAM,EAAE,KAAK,EAAE,aAAa,CAAC,CAAC;AAEpH,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;AAE1G,iHAAiH;AACjH,MAAM,WAAW,GAAG,CAAC,MAAc,EAAE,EAAU,EAAE,IAAY,EAAE,KAAmB,EAAW,EAAE;IAC7F,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,SAAS,EAAE,CAAC;IACzD,OAAO,KAAK,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAC/G,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,EAAU,EAAE,KAAmB,EAAW,EAAE;IAC9E,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IAC7C,OAAO,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,WAAW,EAAE,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAC3I,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,KAAmB,EAAsB,EAAE;IAChF,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CACzC,WAAW,CAAC,MAAM,EAAE,IAAI,CAAC;SACtB,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,EAAE,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,EAAE,EAAE,KAAK,CAAC,CAAC;SACtF,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAChF,CAAC;IACF,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CAAC;AAIF,gIAAgI;AAChI,MAAM,cAAc,GAAG,CAAC,MAAY,EAAE,KAAW,EAAE,WAAoB,EAAY,EAAE;IACnF,MAAM,SAAS,GAAG,MAAM,CAAC,UAAU,IAAI,CAAC,KAAK,CAAC,UAAU,IAAI,KAAK,CAAC,mBAAmB,KAAK,IAAI,CAAC,CAAC;IAChG,OAAO;QACL,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,OAAO,EAAE,aAAa,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5E,GAAG,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,OAAO,EAAE,aAAa,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;KAC7E,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC;AAC3C,CAAC,CAAC;AAEF,6EAA6E;AAC7E,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAE,IAAU,EAAE,KAAmB,EAAsB,EAAE;IAC7F,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,GAAG,aAAa,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IACvE,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,EAAE,IAAI,CAAC,GAAG,GAAG,CAAC,GAAG,aAAa,CAAC,EAAE,KAAK,CAAC,CAAC;IAC1F,MAAM,OAAO,GAAG,cAAc,CAC5B,MAAM,EACN,KAAK,EACL,KAAK,KAAK,CAAC;QACT,IAAI;aACD,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC;aACxB,IAAI,EAAE;aACN,UAAU,CAAC,MAAM,CAAC,OAAO,CAAC,CAChC,CAAC;IACF,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,KAAK,EAAE,MAAM,CAAC,aAAa,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IAClG,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC;AAC7G,CAAC,CAAC;AAEF,mEAAmE;AACnE,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,QAAyB,EAAE,KAAmB,EAAyB,EAAE;IAC5G,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACxC,OAAO,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,eAAe,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAClE,MAAM,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;QAC/D,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC,MAAM,GAAG,eAAe;YAAE,OAAO,EAAE,CAAC;QACnD,MAAM,OAAO,GAAG,gBAAgB,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC;QAC7D,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,QAAQ,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC;IAChG,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/warning-action.d.ts

@@ -0,0 +1,22 @@
+export type WarningWords = {
+    /** What a warning opens with: Warning, Caution, 注意, 警告. */
+    readonly labels: readonly string[];
+    /**
+     * What makes a sentence an action. A Latin entry is matched as whole words anywhere (do not, make sure, must); another is
+     * matched at the end of the sentence (ください, ましょう), where Japanese puts what the reader is to do.
+     */
+    readonly actions: readonly string[];
+};
+/**
+ * Where the text after a paragraph's warning label starts, or undefined when the paragraph opens with no label. The label
+ * must stand alone: "Warning:", "注意：", "**Caution**", "> [!WARNING]", not "Warnings are logged" or 「注意深く」.
+ */
+export declare const warningTextStart: (paragraph: string, labels: readonly string[]) => number | undefined;
+/** Whether a sentence tells the reader what to do, by the words of the warning-action lexicon. */
+export declare const namesAction: (sentence: string, actions: readonly string[]) => boolean;
+/**
+ * Where the block quote holding an offset ends, or undefined when its line is not quoted. A GitHub alert (> [!WARNING])
+ * runs over every quoted line after it, blank quoted lines (>) included, so its action may be in a later paragraph.
+ */
+export declare const quoteBlockEnd: (source: string, offset: number) => number | undefined;
+//# sourceMappingURL=warning-action.d.ts.map

package/dist/warning-action.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"warning-action.d.ts","sourceRoot":"","sources":["../src/warning-action.ts"],"names":[],"mappings":"AAKA,MAAM,MAAM,YAAY,GAAG;IACzB,2DAA2D;IAC3D,QAAQ,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,CAAC;IACnC;;;OAGG;IACH,QAAQ,CAAC,OAAO,EAAE,SAAS,MAAM,EAAE,CAAC;CACrC,CAAC;AAWF;;;GAGG;AACH,eAAO,MAAM,gBAAgB,GAAI,WAAW,MAAM,EAAE,QAAQ,SAAS,MAAM,EAAE,KAAG,MAAM,GAAG,SAWxF,CAAC;AAaF,kGAAkG;AAClG,eAAO,MAAM,WAAW,GAAI,UAAU,MAAM,EAAE,SAAS,SAAS,MAAM,EAAE,KAAG,OAI1E,CAAC;AAIF;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,QAAQ,MAAM,EAAE,QAAQ,MAAM,KAAG,MAAM,GAAG,SAOvE,CAAC"}

package/dist/warning-action.js

@@ -0,0 +1,58 @@
+// A warning (Warning:, 注意：, > [!WARNING]) that names a risk and gives the reader nothing to do. Pure: reads a paragraph's
+// text and its sentences; the labels and the words of an action come from the language's warning-label and warning-action
+// lexicons.
+import { escapeRegExp } from "./orthography.js";
+/** What may come before the label: a quote mark, bold, a GitHub alert's [! (> [!WARNING], **Warning:**). */
+const BEFORE_LABEL = /^(?:>[ \t]*)*(?:\*\*|__)?(?:\[!)?/u;
+/** What closes the label: the end of bold or of an alert, then a colon or an exclamation mark, or the end of the line. */
+const LABEL_CLOSE = /^(?:\*\*|__)?\]?[ \t]*/u;
+const LABEL_MARK = /^[:：!！](?:\*\*|__)?[ \t]*/u;
+const SENTENCE_END_MARKS = new Set(" \t\r\n。．.!！?？」』)）");
+const LATIN_START = /^[A-Za-z]/u;
+const LETTER = /\p{L}/u;
+/**
+ * Where the text after a paragraph's warning label starts, or undefined when the paragraph opens with no label. The label
+ * must stand alone: "Warning:", "注意：", "**Caution**", "> [!WARNING]", not "Warnings are logged" or 「注意深く」.
+ */
+export const warningTextStart = (paragraph, labels) => {
+    const lead = BEFORE_LABEL.exec(paragraph)?.[0].length ?? 0;
+    const rest = paragraph.slice(lead);
+    const label = labels.find((word) => rest.toLowerCase().startsWith(word.toLowerCase()));
+    if (label === undefined)
+        return undefined;
+    const after = rest.slice(label.length);
+    const close = LABEL_CLOSE.exec(after)?.[0].length ?? 0;
+    const mark = LABEL_MARK.exec(after.slice(close))?.[0].length;
+    if (mark !== undefined)
+        return lead + label.length + close + mark;
+    const next = after.slice(close);
+    return next === "" || next.startsWith("\n") || next.startsWith("\r\n") ? lead + label.length + close : undefined;
+};
+const SPACES = /\s+/gu;
+/** A word or phrase as whole words, with any run of spaces or a line break between its words ("Do\nnot"). */
+const hasWholeWord = (text, word) => {
+    const lowered = text.replaceAll(SPACES, " ").toLowerCase();
+    const target = word.toLowerCase();
+    return [...lowered.matchAll(new RegExp(escapeRegExp(target), "gu"))].some((match) => !LETTER.test(lowered[match.index - 1] ?? "") && !LETTER.test(lowered[match.index + target.length] ?? ""));
+};
+/** Whether a sentence tells the reader what to do, by the words of the warning-action lexicon. */
+export const namesAction = (sentence, actions) => {
+    const chars = [...sentence];
+    const ending = chars.slice(0, chars.findLastIndex((char) => !SENTENCE_END_MARKS.has(char)) + 1).join("");
+    return actions.some((word) => (LATIN_START.test(word) ? hasWholeWord(sentence, word) : ending.endsWith(word)));
+};
+const QUOTED = /^[ \t]*>/u;
+/**
+ * Where the block quote holding an offset ends, or undefined when its line is not quoted. A GitHub alert (> [!WARNING])
+ * runs over every quoted line after it, blank quoted lines (>) included, so its action may be in a later paragraph.
+ */
+export const quoteBlockEnd = (source, offset) => {
+    const lineStart = source.lastIndexOf("\n", offset - 1) + 1;
+    const lines = source.slice(lineStart).split("\n");
+    const quoted = lines.findIndex((line) => !QUOTED.test(line));
+    const count = quoted < 0 ? lines.length : quoted;
+    if (count === 0)
+        return undefined;
+    return lineStart + lines.slice(0, count).join("\n").length;
+};
+//# sourceMappingURL=warning-action.js.map

package/dist/warning-action.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"warning-action.js","sourceRoot":"","sources":["../src/warning-action.ts"],"names":[],"mappings":"AAAA,0HAA0H;AAC1H,0HAA0H;AAC1H,YAAY;AACZ,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAYhD,4GAA4G;AAC5G,MAAM,YAAY,GAAG,oCAAoC,CAAC;AAC1D,0HAA0H;AAC1H,MAAM,WAAW,GAAG,yBAAyB,CAAC;AAC9C,MAAM,UAAU,GAAG,4BAA4B,CAAC;AAChD,MAAM,kBAAkB,GAAG,IAAI,GAAG,CAAC,oBAAoB,CAAC,CAAC;AACzD,MAAM,WAAW,GAAG,YAAY,CAAC;AACjC,MAAM,MAAM,GAAG,QAAQ,CAAC;AAExB;;;GAGG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,SAAiB,EAAE,MAAyB,EAAsB,EAAE;IACnG,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC;IAC3D,MAAM,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACnC,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACvF,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACvC,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC;IACvD,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC7D,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,KAAK,GAAG,IAAI,CAAC;IAClE,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IAChC,OAAO,IAAI,KAAK,EAAE,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AACnH,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,OAAO,CAAC;AAEvB,6GAA6G;AAC7G,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,IAAY,EAAW,EAAE;IAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC;IAC3D,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IAClC,OAAO,CAAC,GAAG,OAAO,CAAC,QAAQ,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CACvE,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CACpH,CAAC;AACJ,CAAC,CAAC;AAEF,kGAAkG;AAClG,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,QAAgB,EAAE,OAA0B,EAAW,EAAE;IACnF,MAAM,KAAK,GAAG,CAAC,GAAG,QAAQ,CAAC,CAAC;IAC5B,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,aAAa,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,kBAAkB,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACzG,OAAO,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AACjH,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,WAAW,CAAC;AAE3B;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,MAAc,EAAsB,EAAE;IAClF,MAAM,SAAS,GAAG,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC;IAC3D,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClD,MAAM,MAAM,GAAG,KAAK,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAC7D,MAAM,KAAK,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC;IACjD,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAClC,OAAO,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;AAC7D,CAAC,CAAC"}

package/genres.yaml

@@ -351,6 +351,8 @@ genres:
       total-mismatch: normal
       dangling-figure-reference: normal
       date-range-reversed: normal
+      # A statute uses its defined terms as defined, like a contract. Its profile says a quoted term is the term mentioned.
+      defined-term-form: normal
   - id: legal/judgment
     name: { ja: 判決, en: Court decision }
     summary:

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "chaffjs",
-  "version": "0.27.1",
+  "version": "0.29.0",
   "description": "Prose validation harness. Finds what is hard to read, and never rewrites it for you.",
   "keywords": [
     "chaff",
@@ -87,8 +87,9 @@
   },
   "dependencies": {
     "@anthropic-ai/sdk": "^0.125.0",
-    "@chaffjs/lang-en": "0.24.1",
-    "@chaffjs/lang-ja": "0.26.1",
+    "@chaffjs/lang-en": "0.26.0",
+    "@chaffjs/lang-ja": "0.28.0",
+    "character-entities": "^2.0.2",
     "mdast-util-from-markdown": "^2.0.3",
     "mdast-util-frontmatter": "^2.0.1",
     "mdast-util-gfm-table": "^2.0.0",

package/profiles/statute.yaml

@@ -15,6 +15,9 @@ detect:
     min_lines: 3
 
 ja:
+  # 定義した語を括弧に入れるのは、語そのものを指すとき（「株主」とあるのは「社員」と読み替える、「個人情報」の定義）。
+  # 定義のあとの括弧を、定義し直しや付け忘れとして指さない（defined-term-form）。
+  quote_mentions_term: true
   # 番地の書き方。第 + 漢数字 + 単位、枝番号（の二）が続いてもよい。
   # 「前二項」「前三条」は直前の二つの項・三つの条を指す番地で、数量ではない。「前二年」は期間なので書かない。
   addresses:
@@ -64,3 +67,8 @@ ja:
       - "の[〇一二三四五六七八九十百千]+"
     # 並びの途中に挟まる括弧書き。
     aside: { open: （, close: ） }
+
+en:
+  # A statute quotes a defined term to talk about the term itself ("controller" has the same meaning as in Part 3,
+  # the definition of "the UK GDPR"), not to define it again (defined-term-form).
+  quote_mentions_term: true

package/rules/agreement-slip.yaml

@@ -5,7 +5,7 @@ how_to_find: agreement-slip
 requires: [pos]
 languages: [en]
 extra_word_lists: [singular-determiner, plural-determiner, invariant-noun, count-adjective, dependent-possessive, finite-auxiliary, misnumbered-phrase]
-severity: info
+severity: warning
 
 name:
   ja: 語の形が合っていない
@@ -31,7 +31,7 @@ how_to_fix:
   ja: 限定詞か名詞のどちらかに揃えてください（a significant change / significant changes）。所有の語なら、主語の語に直してください（Your can → You can、their is → there is）。
   en: Make the determiner and the noun agree ("a significant change" or "significant changes"). For a possessive, write the subject you meant ("Your can" becomes "You can", "their is" becomes "there is").
 
-levels: { strict: warning, normal: info, relaxed: info }
+levels: { strict: error, normal: warning, relaxed: info }
 
 use_for: [blog, business, technical, legal, docs, academic, literature, speech]
 off_for:

package/rules/amount-range-reversed.yaml

@@ -0,0 +1,52 @@
+id: amount-range-reversed
+layer: L3
+status: stable
+how_to_find: amount-range
+# The amounts are read with currency-notation and amount-multiplier (amount-short-scale adds "k"); the joints are
+# range-connector, range-opener, range-closer and amount-range-word; amount-change-word and amount-bound-label come next.
+severity: warning
+
+name:
+  ja: 金額の範囲の上限が下限より小さい
+  en: An amount range whose upper end is below its lower end
+
+why:
+  ja: 「月給30万円〜25万円」「下限30万円・上限25万円」のように、金額の範囲の終わりが始まりより小さいと、読み手はどちらの金額が正しいのか、範囲がいくらからいくらまでなのかを問い合わせることになります。求人の給与や見積もりの幅で、片方の金額だけを書き換えたときや、二つを入れ替えて書いたときに起きます。同じ通貨の二つの金額の間が範囲の記号（〜、–）だけのときと、「から」でつなぎ終わりの金額のあとに「まで」が続くときを範囲として比べます。単位を一度だけ書いた範囲（30〜25万円）も読みます。同じ行に下限と上限を書いた二つの金額も比べます。
+  en: 'When an amount range ends below where it starts ("$90,000–$70,000 a year", "minimum $90,000, maximum $70,000"), the reader has to ask which amount is right and what the range is. It happens in a salary or a quote when only one of the two amounts is edited, or when the two are written the wrong way round. Two amounts in one currency are read as a range when only a range mark (–, —) or "to" stands between them, including a range that writes its currency or its "k" once ("$90–70k"). Two amounts on one line labelled minimum and maximum are compared too.'
+
+message:
+  ja: 金額の範囲「{range}」の上限が下限より小さくなっています
+  en: 'The amount range "{range}" ends below where it starts'
+
+how_to_fix:
+  ja: 範囲の二つの金額を確かめて、書き間違えたほうを直すか、小さいほうから書いてください。
+  en: Check both amounts, and correct the wrong one or write the smaller one first.
+
+levels: { strict: error, normal: warning, relaxed: info }
+
+use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+
+group: facts
+summary:
+  ja: 「月給30万円〜25万円」のように、金額の範囲の終わりが始まりより小さい所
+  en: An amount range whose second amount is smaller than its first, such as "$90,000–$70,000"
+example:
+  ja:
+    before: |
+      給与は月給30万円〜25万円（経験により決定）です。
+    after: |
+      給与は月給25万円〜30万円（経験により決定）です。
+  en:
+    before: |
+      Salary: $90,000–$70,000 a year, depending on experience.
+    after: |
+      Salary: $70,000–$90,000 a year, depending on experience.
+
+not_flagged:
+  ja: 同じ文に値段を変える語のある二つの金額（「月額を5,000円から3,000円に値下げ」「改定」）、通貨の違う二つの金額、間に言葉のある二つの金額、計算の式（「5,000円-2,000円=3,000円」）は比べません。版の番号、ページ、日付のように通貨の無い数は読みません。単位を一度だけ書いた範囲は、二つの数が十倍以上違えば比べません（「5,000〜1万円」）。
+  en: 'Two amounts in a sentence with a word that changes a price ("the fee was reduced from $90 to $70", "cut", "dropped"), two amounts in different currencies, two amounts with words between them, and a sum ("$90 - $70 = $20") are not compared. Numbers without a currency, such as versions, pages and dates, are not read. A range that writes its "k" or its word of scale once is not compared when its two numbers are ten times apart or more.'
+level_meaning:
+  ja: 金額の範囲の終わりが始まりより小さいとき指摘
+  en: reported when an amount range ends below where it starts