Fix word count issue for Chinese and Japanese

javatests/org/chromium/distiller/StringUtilTest.java

 // Copyright 2014 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 package org.chromium.distiller;
 
 import com.google.gwt.regexp.shared.RegExp;
 
 public class StringUtilTest extends JsTestCase {
     public void testCountWords() {
         assertEquals(0, StringUtil.countWords(""));
         assertEquals(0, StringUtil.countWords("  -@# ';]"));
         assertEquals(1, StringUtil.countWords("word"));
         assertEquals(1, StringUtil.countWords("b'fore"));
         assertEquals(1, StringUtil.countWords(" _word.under_score_ "));
         assertEquals(2, StringUtil.countWords(" \ttwo\nwords"));
         assertEquals(2, StringUtil.countWords(" \ttwo @^@^&(@#$([][;;\nwords"));
         assertEquals(5, StringUtil.countWords("dør når på svært dårlig"));
         assertEquals(5, StringUtil.countWords("svært få dør av blåbærsyltetøy"));
+
+        // One Chinese sentence, or a series of Japanese glyphs should not be treated
+        // as a single word.
+        assertTrue(StringUtil.countWords("一個中文句子不應該當成一個字") > 1); // zh-Hant
+        assertTrue(StringUtil.countWords("中国和马来西亚使用简体字") > 1); // zh-Hans
+        assertTrue(StringUtil.countWords("ファイナルファンタジー") > 1); // Katakana
+        assertTrue(StringUtil.countWords("いってらっしゃい") > 1); // Hiragana
+        assertTrue(StringUtil.countWords("仏仮駅辺") > 1); // Kanji
+        // However, treating each Chinese/Japanese glyph as a word is also wrong.
+        assertTrue(StringUtil.countWords("一個中文句子不應該當成一個字") < 14);
+        assertTrue(StringUtil.countWords("中国和马来西亚使用简体字") < 12);
+        assertTrue(StringUtil.countWords("ファイナルファンタジー") < 11);
+        assertTrue(StringUtil.countWords("いってらっしゃい") < 8);
+        assertTrue(StringUtil.countWords("仏仮駅辺") < 4);
+        // Even if they are separated by spaces.
+        assertTrue(StringUtil.countWords("一 個 中 文 句 子 不 應 該 當 成 一 個 字") < 14);
+        assertTrue(StringUtil.countWords("中 国 和 马 来 西 亚 使 用 简 体 字") < 12);
+        assertTrue(StringUtil.countWords("フ ァ イ ナ ル フ ァ ン タ ジ ー") < 11);
+        assertTrue(StringUtil.countWords("い っ て ら っ し ゃ い") < 8);
+        assertTrue(StringUtil.countWords("仏 仮 駅 辺") < 4);
+
+        assertEquals(1, StringUtil.countWords("字"));
+        assertEquals(1, StringUtil.countWords("が"));
+
+        // Mixing ASCII words and Chinese/Japanese glyphs
+        assertEquals(2, StringUtil.countWords("word字"));
+        assertEquals(2, StringUtil.countWords("word 字"));
+
+        // Hangul uses space as word delimiter like English.
+        assertEquals(1, StringUtil.countWords("어"));
+        assertEquals(2, StringUtil.countWords("한국어 단어"));
+        assertEquals(5, StringUtil.countWords("한 국 어 단 어"));
+        assertEquals(8, StringUtil.countWords(
+                "예비군 훈련장 총기 난사범 최모씨의 군복에서 발견된 유서."));
+    }
+
+    public void testCountWordsFast() {
+        StringUtil.selectCountWordsFunc("");

[cjhopman, 2015/05/29 19:38:57]

This should probably use the inner classes directly rather than relying on the
implementation of selectCountWordsFunc

Though, we should probably have a test of selectCountWordsFunc. I think it'd be
better to sort of expose the selection itself rather than determining what was
selected based on how it counts different words, though.

[wychen, 2015/05/31 09:04:03]

Done.

+        assertEquals(0, StringUtil.countWords("어"));
+        StringUtil.selectCountWordsFunc("어");
+        assertEquals(1, StringUtil.countWords("어"));
+        assertEquals(0, StringUtil.countWords("字"));
+        StringUtil.selectCountWordsFunc("字");
+        assertEquals(1, StringUtil.countWords("字"));
     }
 
     public void testIsWhitespace() {
         assertTrue(StringUtil.isWhitespace(' '));
         assertTrue(StringUtil.isWhitespace('\t'));
         assertTrue(StringUtil.isWhitespace('\n'));
         assertTrue(StringUtil.isWhitespace('\u00a0'));
         assertFalse(StringUtil.isWhitespace('a'));
         assertFalse(StringUtil.isWhitespace('$'));
         assertFalse(StringUtil.isWhitespace('_'));
@@ skipping 77 matching lines @@
         assertEquals(-1, StringUtil.toNumber("'8_"));
         assertEquals(-1, StringUtil.toNumber(""));
         assertEquals(-1, StringUtil.toNumber(" "));
         assertEquals(-1, StringUtil.toNumber("\u00a0\u0460"));
         assertEquals(-1, StringUtil.toNumber("abc"));
         assertEquals(-1, StringUtil.toNumber("$"));
         assertEquals(-1, StringUtil.toNumber("_"));
     }
 
 }