fix: normalize case before calculating similarity scores

CHANGELOG.md

@@ -26,6 +26,7 @@
 - Release publishing now validates the tag/version and runs formatting, lint, and tests before publishing the verified build.
 
 ### Fixed
+- Calculate `similarityScore` lengths and edit distance after lowercasing so Unicode case expansion cannot produce negative scores or incorrect similarity ratios.
 - Prevent invalid levels from hanging `tashfeer`, `tashfeerBannedWords`, and `toOldArabicAndTashfeerBannedWords`: non-number values throw `TypeError`; negative, fractional, and non-finite numbers throw `RangeError`. Random-index generation handles empty input and caps the requested count at the word length ([#35](https://github.com/Seen-Arabic/Arabic-Services-JavaScript/issues/35), [PR #36](https://github.com/Seen-Arabic/Arabic-Services-JavaScript/pull/36)).
 - Attempt browser bundle uploads after a successful build even if npm publishing fails ([commit 598035a](https://github.com/Seen-Arabic/Arabic-Services-JavaScript/commit/598035a)).
 

src/utils/string.util.ts

@@ -12,13 +12,17 @@ export function setCharAt(str: string, index: number, chr: string): string {
 }
 
 /**
- * Calculates a similarity score between two strings, potentially based on their edit distance.
+ * Calculates a case-insensitive similarity score using Levenshtein distance over UTF-16 code units.
+ * Both the distance and string lengths are measured after lowercasing.
  *
  * @param {string} s1 - The first string to compare.
  * @param {string} s2 - The second string to compare.
  * @returns {number} A similarity score representing how similar the two strings are. The score is between 0 and 1.
  */
 export function similarityScore(s1: string, s2: string): number {
+	s1 = s1.toLowerCase();
+	s2 = s2.toLowerCase();
+
 	let longer = s1;
 	let shorter = s2;
 	// swap them if s1 is bigger than s2
@@ -45,9 +49,6 @@ export function similarityScore(s1: string, s2: string): number {
  * @returns {number} The edit distance between the two strings.
  */
 function editDistance(s1: string, s2: string): number {
-	s1 = s1.toLowerCase();
-	s2 = s2.toLowerCase();
-
 	const costs = new Array<number>();
 	for (let i = 0; i <= s1.length; i++) {
 		let lastValue = i;

tests/scripts.test.ts

@@ -286,6 +286,17 @@ describe('#removeArabicAffixes', () => {
 });
 
 describe('#similarityScore', () => {
+	it.each([
+		['İ', 'a', 0],
+		['İ', 'i', 0.5],
+		['İ', '', 0],
+		['İ', 'i\u0307', 1],
+		['A', 'a', 1],
+	])('should score %j and %j as %s after case conversion in either order', (left, right, expected) => {
+		expect(ArabicServices.utils.similarityScore(left, right)).toBe(expected);
+		expect(ArabicServices.utils.similarityScore(right, left)).toBe(expected);
+	});
+
 	it('should return 1 for identical strings', () => {
 		const s1 = 'مرحبا';
 		const s2 = 'مرحبا';