Levenshtein

fluffy-melli · Jan 14, 2025 · 65bff56 · 65bff56
1 parent ed8827f
commit 65bff56
Show file tree

Hide file tree

Showing 3 changed files with 45 additions and 12 deletions.
diff --git a/korcen.go b/korcen.go
@@ -487,7 +487,7 @@ func General(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -532,7 +532,7 @@ func Minor(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -723,7 +723,7 @@ func Sexual(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -780,7 +780,7 @@ func Belittle(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -805,7 +805,7 @@ func Race(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -838,7 +838,7 @@ func Parent(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -882,7 +882,7 @@ func Politics(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -905,7 +905,7 @@ func Japanese(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -929,7 +929,7 @@ func Chinese(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}
@@ -952,7 +952,7 @@ func Special(input string) (bool, string) {
 		if strings.Contains(input, item) {
 			return true, item
 		}
-		if Levenshtein(input, item) <= 4 {
+		if Levenshtein(input, item) <= 3 {
 			return true, item
 		}
 	}

diff --git a/leven.go b/leven.go
@@ -1,8 +1,36 @@
 package korcen
 
+import (
+	"unicode"
+)
+
+func SplitHangul(s string) []rune {
+	var result []rune
+	for _, r := range s {
+		if unicode.Is(unicode.Hangul, r) {
+			code := r - 0xAC00
+			chun := code / (21 * 28)
+			jung := (code % (21 * 28)) / 28
+			jong := code % 28
+			if chun != 0 {
+				result = append(result, rune(chun+0x1100))
+			}
+			if jung != 0 {
+				result = append(result, rune(jung+0x1161))
+			}
+			if jong != 0 {
+				result = append(result, rune(jong+0x11A7))
+			}
+		} else {
+			result = append(result, r)
+		}
+	}
+	return result
+}
+
 func Levenshtein(a, b string) int {
-	m := len(a)
-	n := len(b)
+	m := len(SplitHangul(a))
+	n := len(SplitHangul(b))
 	dp := make([][]int, m+1)
 	for i := range dp {
 		dp[i] = make([]int, n+1)

diff --git a/test/speed_test.go b/test/speed_test.go
@@ -1,6 +1,7 @@
 package test
 
 import (
+	"fmt"
 	"testing"
 
 	"github.com/fluffy-melli/korcen-go"
@@ -12,3 +13,7 @@ func Benchmark0(b *testing.B) {
 		korcen.Check("")
 	}
 }
+
+func Test0(t *testing.T) {
+	fmt.Println(korcen.Levenshtein("안녕하세요", "안녕안해요"))
+}