From e4e6bd5335a175276e07568e0b46e26988b9501d Mon Sep 17 00:00:00 2001
From: voduylinhnashtech <142197011+voduylinhnashtech@users.noreply.github.com>
Date: Thu, 3 Oct 2024 13:30:43 +0700
Subject: [PATCH] Update kgsim.go

---
 chipper/pkg/wirepod/ttr/kgsim.go | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/chipper/pkg/wirepod/ttr/kgsim.go b/chipper/pkg/wirepod/ttr/kgsim.go
index 9b0b9854..8760a2e9 100644
--- a/chipper/pkg/wirepod/ttr/kgsim.go
+++ b/chipper/pkg/wirepod/ttr/kgsim.go
@@ -66,7 +66,27 @@ func Remember(user, ai openai.ChatCompletionMessage, esn string) {
 }
 
 func isMn(r rune) bool {
-	return unicode.Is(unicode.Mn, r) // Mn: nonspacing marks
+	// Remove the characters that are not related to Vietnamese.
+	// Retain the tonal marks and diacritics such as the circumflex, ơ, and ư in Vietnamese.
+	keepMarks := []rune{
+		'\u0300', // Dấu huyền
+		'\u0301', // Dấu sắc
+		'\u0303', // Dấu ngã
+		'\u0309', // Dấu hỏi
+		'\u0323', // Dấu nặng
+		'\u0302', // Dấu mũ (â, ê, ô)
+		'\u031B', // Dấu ơ và ư
+		'\u0306', // Dấu trầm
+	}
+	if unicode.Is(unicode.Mn, r) {
+		for _, mark := range keepMarks {
+			if r == mark {
+				return false
+			}
+		}
+		return true
+	}
+	return false
 }
 
 func removeSpecialCharacters(str string) string {