From 2bd9dea91d67440f43c68a4eceef2fc68ad4f384 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Enrique=20Jim=C3=A9nez?= <enrique.jimenez@lmu.de>
Date: Sat, 24 Aug 2024 10:23:04 +0000
Subject: [PATCH] Implement new logic for accent conversion in search field

---
 src/fragmentarium/domain/normalizeAccents.test.ts | 2 ++
 src/fragmentarium/domain/normalizeAccents.ts      | 9 +++++++--
 2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/src/fragmentarium/domain/normalizeAccents.test.ts b/src/fragmentarium/domain/normalizeAccents.test.ts
index 5096ef8b3..c90b5bd14 100644
--- a/src/fragmentarium/domain/normalizeAccents.test.ts
+++ b/src/fragmentarium/domain/normalizeAccents.test.ts
@@ -14,6 +14,8 @@ test.each([
   ['PÈL', 'PEL₃'],
   ['LÚ', 'LU₂'],
   ['ÀMA', 'AMA₃'],
+  ['ÀM]A', 'AM]A₃'],
+  ['ÀM[A', 'AM[A₃'],
   ['ÀMA.KÚR', 'AMA₃.KUR₂'],
   ['LÌ gí-la-ka₃', 'LI₃ gi₂-la-ka₃'],
 ])('%s', (character, replacement) => {
diff --git a/src/fragmentarium/domain/normalizeAccents.ts b/src/fragmentarium/domain/normalizeAccents.ts
index 338db8879..7106c42f5 100644
--- a/src/fragmentarium/domain/normalizeAccents.ts
+++ b/src/fragmentarium/domain/normalizeAccents.ts
@@ -10,8 +10,10 @@ const specialTransliterationCharactersAsString = escapeRegExp(
   specialTransliterationCharacters
 )
 
+const spaceCharacters = '[^\\{\\}\\-\\.:,\\+\\s\\n]'
+
 const accentExpression = new RegExp(
-  `(${charactersWithAccentsAsString})(${specialTransliterationCharactersAsString}|)\\w*`,
+  `(${charactersWithAccentsAsString})(${specialTransliterationCharactersAsString}|)${spaceCharacters}*`,
   'g'
 )
 
@@ -25,6 +27,9 @@ export default function normalizeAccents(userInput: string): string {
       .split('')
       .map((character) => charactersWithAccents[character]?.letter ?? character)
 
-    return characterWithoutAccent.concat(subindex).join('')
+    // Check if the characterWithoutAccent contains an index
+    const index = subindex.length > 0 ? subindex.join('') : ''
+
+    return characterWithoutAccent.concat(index).join('')
   })
 }