@sushichan044/kg-core 0.11.0 → 0.11.2

package/dist/{question-or-exclamation-spacing-C5cFLeKr.mjs → question-or-exclamation-spacing-D7TSDUG8.mjs}

@@ -180,6 +180,26 @@ function assertNever(value) {
 	throw new TypeError(`unhandled variant: ${JSON.stringify(value)}`);
 }
 //#endregion
+//#region src/internal/japanese-punctuation.ts
+/**
+* JLReq cl-04 occupies a full em, including compound marks whose Unicode East Asian Width is
+* neutral.
+*/
+const DIVIDING_PUNCTUATION = "！？‼⁇⁈⁉";
+/**
+* The marks a short run of which is set as one tate-chu-yoko cell, `！？` reading as `⁉`.
+*/
+const TATE_CHU_YOKO_MARKS = "!?！？";
+const SENTENCE_ENDING_MARKS = `${DIVIDING_PUNCTUATION}♡♪☆`;
+const validAdditionalSentenceEndingMarks = (value) => /^[\p{S}]*$/u.test(value);
+/**
+* Match one symbol grapheme, allowing its text or emoji variation selector.
+*/
+function isSentenceEndingMark(grapheme, additionalMarks = "") {
+	const symbol = grapheme.replace(/\p{Variation_Selector}/gu, "");
+	return /^.$/u.test(symbol) && (SENTENCE_ENDING_MARKS + additionalMarks).includes(symbol);
+}
+//#endregion
 //#region src/internal/japanese-brackets.ts
 /**
 * Brackets and quotes a Japanese paragraph may open with. Halfwidth forms are included because a
@@ -198,17 +218,6 @@ function characterClass(characters) {
 	return characters.replaceAll(/[\\\]^-]/gu, "\\$&");
 }
 //#endregion
-//#region src/internal/japanese-punctuation.ts
-/**
-* JLReq cl-04 occupies a full em, including compound marks whose Unicode East Asian Width is
-* neutral.
-*/
-const DIVIDING_PUNCTUATION = "！？‼⁇⁈⁉";
-/**
-* The marks a short run of which is set as one tate-chu-yoko cell, `！？` reading as `⁉`.
-*/
-const TATE_CHU_YOKO_MARKS = "!?！？";
-//#endregion
 //#region src/internal/japanese-spaces.ts
 const LEADING_SPACES = /^[　 \t]+/u;
 /**
@@ -223,14 +232,7 @@ function leadingSpaces(text) {
 /**
 * Runs, not single marks: `！？` is one mark and takes one gap, so the gap belongs after the run.
 */
-const MARKS = new RegExp(`[${characterClass(DIVIDING_PUNCTUATION)}]+`, "gu");
-function analyzeMatch(line, marks) {
-	const start = marks.index;
-	if (start === void 0) return void 0;
-	const marksRange = {
-		start,
-		end: start + marks[0].length
-	};
+function analyzeMatch(line, marksRange) {
 	const next = line[marksRange.end];
 	if (next === void 0 || "」』〗〉》）)”\"’'］]〕｝}＞>".includes(next)) return void 0;
 	const spaces = leadingSpaces(line.slice(marksRange.end));
@@ -264,8 +266,22 @@ function analyzeMatch(line, marks) {
 * line. Runs at the end of the line or immediately before a closing bracket need no gap and are
 * omitted from the result.
 */
-function questionOrExclamationSpacings(line) {
-	return [...line.matchAll(MARKS)].map((marks) => analyzeMatch(line, marks)).filter((spacing) => spacing !== void 0);
+function questionOrExclamationSpacings(line, additionalSentenceEndingMarks = "") {
+	const result = [];
+	let marks;
+	const flush = () => {
+		if (marks === void 0) return;
+		const spacing = analyzeMatch(line, marks);
+		if (spacing !== void 0) result.push(spacing);
+		marks = void 0;
+	};
+	for (const { segment, index } of graphemeSegmenter.segment(line)) if (isSentenceEndingMark(segment, additionalSentenceEndingMarks)) marks = {
+		start: marks?.start ?? index,
+		end: index + segment.length
+	};
+	else flush();
+	flush();
+	return result;
 }
 //#endregion
-export { nonNegativeInteger as _, CLOSING_BRACKETS as a, assertNever as c, ManuscriptResult as d, ManuscriptDiagnostic as f, TextRange as g, ManuscriptRange as h, TATE_CHU_YOKO_MARKS as i, graphemeSegmenter as l, NamespacedId as m, leadingSpaces as n, OPENING_BRACKETS as o, DiagnosticSeverity as p, DIVIDING_PUNCTUATION as r, characterClass as s, questionOrExclamationSpacings as t, ValidationIssue as u, readonlyArray as v, readonlyObject as y };
+export { ManuscriptRange as _, characterClass as a, readonlyArray as b, isSentenceEndingMark as c, graphemeSegmenter as d, ValidationIssue as f, NamespacedId as g, DiagnosticSeverity as h, OPENING_BRACKETS as i, validAdditionalSentenceEndingMarks as l, ManuscriptDiagnostic as m, leadingSpaces as n, SENTENCE_ENDING_MARKS as o, ManuscriptResult as p, CLOSING_BRACKETS as r, TATE_CHU_YOKO_MARKS as s, questionOrExclamationSpacings as t, assertNever as u, TextRange as v, readonlyObject as x, nonNegativeInteger as y };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sushichan044/kg-core",
-  "version": "0.11.0",
+  "version": "0.11.2",
   "description": "Framework-independent parsing, composition, and proofreading for Japanese manuscripts",
   "keywords": [
     "japanese",