OLD | NEW |
---|---|
1 // Copyright 2014 The Chromium Authors. All rights reserved. | 1 // Copyright 2014 The Chromium Authors. All rights reserved. |
2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
4 | 4 |
5 package org.chromium.distiller; | 5 package org.chromium.distiller; |
6 | 6 |
7 import com.google.gwt.regexp.shared.RegExp; | 7 import com.google.gwt.regexp.shared.RegExp; |
8 | 8 |
9 public class StringUtil { | 9 public class StringUtil { |
10 // For the whitespace-related functions below, Java's and Javascript's versi ons of '\s' and '\S' | 10 // For the whitespace-related functions below, Java's and Javascript's versi ons of '\s' and '\S' |
(...skipping 24 matching lines...) Expand all Loading... | |
35 } | 35 } |
36 | 36 |
37 public static boolean match(String input, String regex) { | 37 public static boolean match(String input, String regex) { |
38 return RegExp.compile(regex, "i").test(input); | 38 return RegExp.compile(regex, "i").test(input); |
39 } | 39 } |
40 | 40 |
41 public static String findAndReplace(String input, String regex, String repla ce) { | 41 public static String findAndReplace(String input, String regex, String repla ce) { |
42 return RegExp.compile(regex, "gi").replace(input, replace); | 42 return RegExp.compile(regex, "gi").replace(input, replace); |
43 } | 43 } |
44 | 44 |
45 public static native boolean containsWordCharacter(String s) /*-{ | 45 /** |
46 return /[\w\u00C0-\u1FFF\u2C00-\uD7FF]/.test(s); | 46 * For some languages, counting the number of words relies on non-trivial wo rd |
47 * segmentation algorithms, or even huge look-up tables. This function needs to | |
48 * be reasonably fast, so the word count for some languages would only be an | |
49 * approximation. | |
50 * Read https://crbug.com/484750 for more info. | |
51 */ | |
52 public static native void selectCountWordsFunc(String text) /*-{ | |
53 if (/[\u3040-\uA4CF]/.test(text)) { | |
54 $wnd._countWords = function(s) { | |
cjhopman
2015/05/21 19:32:02
Don't make changes to $wnd
wychen
2015/05/21 23:07:03
Done.
| |
55 // The following range includes broader alphabetical letters and Hangul Syllables. | |
56 var m = s.match(/(\S*[\w\u00C0-\u1FFF\uAC00-\uD7AF]\S*)/g); | |
57 var c = (m ? m.length : 0); | |
58 // The following range includes Hiragana, Katakana, and CJK Unif ied Ideographs. | |
59 // Hangul Syllables are not included. | |
60 m = s.match(/([\u3040-\uA4CF])/g); | |
61 c += Math.ceil((m ? m.length : 0) * 0.55); | |
62 return c; | |
63 }; | |
64 return; | |
65 } | |
66 if (/[\AC00-\uD7AF]/.test(text)) { | |
67 $wnd._countWords = function(s) { | |
68 // The following range includes broader alphabetical letters and Hangul Syllables. | |
69 var m = s.match(/(\S*[\w\u00C0-\u1FFF\uAC00-\uD7AF]\S*)/g); | |
70 return (m ? m.length : 0); | |
71 }; | |
72 return; | |
73 } | |
74 $wnd._countWords = function(s) { | |
75 // The following range includes broader alphabetical letters. | |
76 var m = s.match(/(\S*[\w\u00C0-\u1FFF]\S*)/g); | |
77 return (m ? m.length : 0); | |
78 }; | |
47 }-*/; | 79 }-*/; |
48 | 80 |
81 private static final int __dummy_setCountWords = initCountWords(); | |
82 private static int initCountWords() { | |
83 selectCountWordsFunc("字"); | |
84 return 0; | |
85 } | |
86 | |
49 public static native int countWords(String s) /*-{ | 87 public static native int countWords(String s) /*-{ |
50 var m = s.match(/(\S*[\w\u00C0-\u1FFF\u2C00-\uD7FF]\S*)/g); | 88 return $wnd._countWords(s); |
51 return m ? m.length : 0; | |
52 }-*/; | 89 }-*/; |
53 | 90 |
54 public static native String regexEscape(String s) /*-{ | 91 public static native String regexEscape(String s) /*-{ |
55 return s.replace(/[\-\[\]{}()*+?.,\\\^$|#\s]/g, "\\$&"); | 92 return s.replace(/[\-\[\]{}()*+?.,\\\^$|#\s]/g, "\\$&"); |
56 }-*/; | 93 }-*/; |
57 | 94 |
58 /* | 95 /* |
59 * Returns true if character is a digit. | 96 * Returns true if character is a digit. |
60 */ | 97 */ |
61 public static native boolean isDigit(Character c) /*-{ | 98 public static native boolean isDigit(Character c) /*-{ |
(...skipping 17 matching lines...) Expand all Loading... | |
79 /** | 116 /** |
80 * Returns the plain number if given string can be converted to one >= 0. | 117 * Returns the plain number if given string can be converted to one >= 0. |
81 * Returns -1 if string is empty or not all digits. | 118 * Returns -1 if string is empty or not all digits. |
82 */ | 119 */ |
83 public static int toNumber(String s) { | 120 public static int toNumber(String s) { |
84 if (s.isEmpty() || !StringUtil.isStringAllDigits(s)) return -1; | 121 if (s.isEmpty() || !StringUtil.isStringAllDigits(s)) return -1; |
85 return JavaScript.parseInt(s, 10); | 122 return JavaScript.parseInt(s, 10); |
86 } | 123 } |
87 | 124 |
88 } | 125 } |
OLD | NEW |