Handle unicode characters in autotag (#2336)

pkg/sqlite/performer.go

@@ -21,6 +21,8 @@ WHERE performers_tags.tag_id = ?
 GROUP BY performers_tags.performer_id
 `
 
+const singleFirstCharacterRegex = `^[\w\p{L}][.\-_ ]`
+
 type performerQueryBuilder struct {
 	repository
 }
@@ -184,7 +186,7 @@ func (qb *performerQueryBuilder) QueryForAutoTag(words []string) ([]*models.Perf
 	var args []interface{}
 
 	whereClauses = append(whereClauses, "name regexp ?")
-	args = append(args, "^[\\w][.\\-_ ]")
+	args = append(args, singleFirstCharacterRegex)
 
 	for _, w := range words {
 		whereClauses = append(whereClauses, "name like ?")

pkg/sqlite/repository.go

@@ -171,6 +171,8 @@ func (r *repository) runSumQuery(query string, args []interface{}) (float64, err
 }
 
 func (r *repository) queryFunc(query string, args []interface{}, single bool, f func(rows *sqlx.Rows) error) error {
+	logger.Tracef("SQL: %s, args: %v", query, args)
+
 	rows, err := r.tx.Queryx(query, args...)
 
 	if err != nil && !errors.Is(err, sql.ErrNoRows) {

pkg/sqlite/studio.go

@@ -145,7 +145,6 @@ func (qb *studioQueryBuilder) QueryForAutoTag(words []string) ([]*models.Studio,
 	var args []interface{}
 
 	// always include names that begin with a single character
-	singleFirstCharacterRegex := "^[\\w][.\\-_ ]"
 	whereClauses = append(whereClauses, "studios.name regexp ? OR COALESCE(studio_aliases.alias, '') regexp ?")
 	args = append(args, singleFirstCharacterRegex, singleFirstCharacterRegex)
 

pkg/sqlite/tag.go

@@ -236,7 +236,6 @@ func (qb *tagQueryBuilder) QueryForAutoTag(words []string) ([]*models.Tag, error
 	var args []interface{}
 
 	// always include names that begin with a single character
-	singleFirstCharacterRegex := "^[\\w][.\\-_ ]"
 	whereClauses = append(whereClauses, "tags.name regexp ? OR COALESCE(tag_aliases.alias, '') regexp ?")
 	args = append(args, singleFirstCharacterRegex, singleFirstCharacterRegex)