Use RegExp singletons to improve performance (#27)

lovell · Sep 8, 2023 · 7d5a58c · 7d5a58c
1 parent c39b4e5
commit 7d5a58c
Showing 1 changed file with 24 additions and 13 deletions.
diff --git a/lib/hepburn.js b/lib/hepburn.js
@@ -180,6 +180,10 @@ var nihonShiki = {
     "DYO": "JO"
 };
 
+function keysToRegex(obj) {
+  return new RegExp(Object.keys(obj).join('|'), "g");
+}
+
 // For use with toHiragana
 var hiraganaMap = {};
 
@@ -200,6 +204,8 @@ Object.keys(hiraganaDigraphs).forEach(function(key) {
 var hiraganaRegex = new RegExp(Object.keys(hiraganaMap).sort(function(a, b) {
   return b.length - a.length;
 }).join("|"), "g");
+var hiraganaMonographsRegex = keysToRegex(hiraganaMonographs);
+var hiraganaDigraphsRegex = keysToRegex(hiraganaDigraphs);
 
 // For use with toKatakana
 var katakanaMap = {};
@@ -228,17 +234,23 @@ Object.keys(katakanaTrigraphs).forEach(function(key) {
 var katakanaRegex = new RegExp(Object.keys(katakanaMap).sort(function(a, b) {
   return b.length - a.length;
 }).join("|"), "g");
+var katakanaMonographsRegex = keysToRegex(katakanaMonographs);
+var katakanaDigraphsRegex = keysToRegex(katakanaDigraphs);
+var katakanaHalfwidthsCombinedRegex = keysToRegex(katakanaHalfwidthsCombined);
+var katakanaHalfwidthsRegex = keysToRegex(katakanaHalfwidths);
+
+var nihonShikiRegex = keysToRegex(nihonShiki);
 
 // API
 
 exports.fromKana = function(str) {
   // Initial transliteration
-  str = bulkReplace(str, katakanaHalfwidthsCombined);
-  str = bulkReplace(str, katakanaHalfwidths);
-  str = bulkReplace(str, hiraganaDigraphs);
-  str = bulkReplace(str, katakanaDigraphs);
-  str = bulkReplace(str, hiraganaMonographs);
-  str = bulkReplace(str, katakanaMonographs);
+  str = bulkReplace(str, katakanaHalfwidthsCombinedRegex, katakanaHalfwidthsCombined);
+  str = bulkReplace(str, katakanaHalfwidthsRegex, katakanaHalfwidths);
+  str = bulkReplace(str, hiraganaDigraphsRegex, hiraganaDigraphs);
+  str = bulkReplace(str, katakanaDigraphsRegex, katakanaDigraphs);
+  str = bulkReplace(str, hiraganaMonographsRegex, hiraganaMonographs);
+  str = bulkReplace(str, katakanaMonographsRegex, katakanaMonographs);
 
   // Correct use of sokuon
   str = str.replace(/[っッ]C/g, "TC").replace(/[っッ](.)/g, "$1$1");
@@ -312,7 +324,7 @@ exports.cleanRomaji = function(str) {
   str = str.replace(/OH([^AIEO]|$)/g, "OO$1");
 
   // Replace old Nihon-shiki usage with modern Hepburn form
-  str = bulkReplace(str, nihonShiki);
+  str = bulkReplace(str, nihonShikiRegex, nihonShiki);
 
   return str;
 };
@@ -369,20 +381,19 @@ exports.splitRomaji = function(str) {
 }
 
 exports.containsHiragana = function(str) {
-  return new RegExp(Object.keys(hiraganaMonographs).join('|')).test(str);
+  hiraganaMonographsRegex.lastIndex = 0;
+  return hiraganaMonographsRegex.test(str);
 };
 
 exports.containsKatakana = function(str) {
-  return (
-    new RegExp(Object.keys(katakanaMonographs).join('|')).test(str) ||
-    new RegExp(Object.keys(katakanaHalfwidths).join('|')).test(str)
-  );
+  katakanaMonographsRegex.lastIndex = katakanaHalfwidthsRegex.lastIndex = 0;
+  return katakanaMonographsRegex.test(str) || katakanaHalfwidthsRegex.test(str);
 };
 
 exports.containsKana = function(str){
   return (exports.containsHiragana(str) || exports.containsKatakana(str));
 };
 
 exports.containsKanji = function(str){
-  return new RegExp(/[\u4e00-\u9fcf\uf900-\ufaff\u3400-\u4dbf]/).test(str);
+  return /[\u4e00-\u9fcf\uf900-\ufaff\u3400-\u4dbf]/.test(str);
 }