Merge pull request #16 from snapp-incubator/fix/check-minchar-length

fix min char length
snapp-incubator · Feb 9, 2025 · f0c464f · f0c464f
2 parents 31ab7b5 + b1d4656
commit f0c464f
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 3 deletions.
diff --git a/internal/lookup_compound.go b/internal/lookup_compound.go
@@ -46,7 +46,7 @@ func (s *SymSpell) LookupCompound(phrase string, maxEditDistance int) *items.Sug
 		// Combine adjacent terms
 		if i > 0 && !cp.isLastCombi {
 			cp.terms2 = terms1[i-1]
-			suggestionsCombi, _ := s.Lookup(cp.terms2+cp.terms1, verbositypkg.Top, maxEditDistance)
+			suggestionsCombi, _ := s.Lookup(fmt.Sprintf("%s %s", cp.terms2, cp.terms1), verbositypkg.Top, maxEditDistance)
 			if len(suggestionsCombi) > 0 {
 				best1 := cp.suggestionParts[len(cp.suggestionParts)-1]
 				best2 := s.getBestSuggestion2(cp, maxEditDistance)
@@ -114,7 +114,7 @@ func (s *SymSpell) LookupCompound(phrase string, maxEditDistance int) *items.Sug
 }
 
 func (s *SymSpell) getSuggestion(cp *compoundProcessor, maxEditDistance int) {
-	if len(cp.terms1) > s.MinimumCharToChange {
+	if len([]rune(cp.terms1)) > s.MinimumCharToChange {
 		cp.suggestions, _ = s.Lookup(cp.terms1, verbositypkg.Top, maxEditDistance)
 	} else {
 		cp.suggestions = []items.SuggestItem{{

diff --git a/symspell_test.go b/symspell_test.go
@@ -193,6 +193,22 @@ func TestSymspellLookupCompoundUnigram(t *testing.T) {
 			},
 			want: "بیمارستان ا",
 		},
+		{
+			name: "Test Min Character 2",
+			args: args{
+				a:               "بیمارستان اما",
+				maxEditDistance: 3,
+			},
+			want: "بیمارستان اما",
+		},
+		{
+			name: "Test Min Character 3",
+			args: args{
+				a:               "ا ب ث",
+				maxEditDistance: 3,
+			},
+			want: "ا ب ث",
+		},
 		{
 			name: "Split Number",
 			args: args{
@@ -245,7 +261,7 @@ func TestSymspellLookupCompoundUnigram(t *testing.T) {
 		options.WithMaxDictionaryEditDistance(3),
 		options.WithSplitItemThreshold(100),
 		options.WithSplitWordBySpace(),
-		options.WithMinimumCharacterToChange(2),
+		options.WithMinimumCharacterToChange(3),
 		options.WithSplitWordAndNumbers(),
 	)
 	for _, tt := range tests {