jars/icu4j-4_4_2-src/main/classes/translit/src/com/ibm/icu/text/NormalizationTransliterator.java

   1 /*\r
   2 **********************************************************************\r
   3 *   Copyright (C) 2001-2010, International Business Machines\r
   4 *   Corporation and others.  All Rights Reserved.\r
   5 **********************************************************************\r
   6 *   Date        Name        Description\r
   7 *   06/08/01    aliu        Creation.\r
   8 **********************************************************************\r
   9 */\r
  10 \r
  11 package com.ibm.icu.text;\r
  12 import com.ibm.icu.impl.Norm2AllModes;\r
  13 import com.ibm.icu.impl.Normalizer2Impl;\r
  14 \r
  15 /**\r
  16  * @author Alan Liu, Markus Scherer\r
  17  */\r
  18 final class NormalizationTransliterator extends Transliterator {\r
  19     private final Normalizer2 norm2;\r
  20 \r
  21     /**\r
  22      * System registration hook.\r
  23      */\r
  24     static void register() {\r
  25         Transliterator.registerFactory("Any-NFC", new Transliterator.Factory() {\r
  26             public Transliterator getInstance(String ID) {\r
  27                 return new NormalizationTransliterator(\r
  28                         "NFC", Norm2AllModes.getNFCInstance().comp);\r
  29             }\r
  30         });\r
  31         Transliterator.registerFactory("Any-NFD", new Transliterator.Factory() {\r
  32             public Transliterator getInstance(String ID) {\r
  33                 return new NormalizationTransliterator(\r
  34                         "NFD", Norm2AllModes.getNFCInstance().decomp);\r
  35             }\r
  36         });\r
  37         Transliterator.registerFactory("Any-NFKC", new Transliterator.Factory() {\r
  38             public Transliterator getInstance(String ID) {\r
  39                 return new NormalizationTransliterator(\r
  40                         "NFKC", Norm2AllModes.getNFKCInstance().comp);\r
  41             }\r
  42         });\r
  43         Transliterator.registerFactory("Any-NFKD", new Transliterator.Factory() {\r
  44             public Transliterator getInstance(String ID) {\r
  45                 return new NormalizationTransliterator(\r
  46                         "NFKD", Norm2AllModes.getNFKCInstance().decomp);\r
  47             }\r
  48         });\r
  49         Transliterator.registerFactory("Any-FCD", new Transliterator.Factory() {\r
  50             public Transliterator getInstance(String ID) {\r
  51                 return new NormalizationTransliterator(\r
  52                         "FCD", Norm2AllModes.getFCDNormalizer2());\r
  53             }\r
  54         });\r
  55         Transliterator.registerFactory("Any-FCC", new Transliterator.Factory() {\r
  56             public Transliterator getInstance(String ID) {\r
  57                 return new NormalizationTransliterator(\r
  58                         "FCC", Norm2AllModes.getNFCInstance().fcc);\r
  59             }\r
  60         });\r
  61         Transliterator.registerSpecialInverse("NFC", "NFD", true);\r
  62         Transliterator.registerSpecialInverse("NFKC", "NFKD", true);\r
  63         Transliterator.registerSpecialInverse("FCC", "NFD", false);\r
  64         Transliterator.registerSpecialInverse("FCD", "FCD", false);\r
  65     }\r
  66 \r
  67     /**\r
  68      * Constructs a transliterator.\r
  69      */\r
  70     private NormalizationTransliterator(String id, Normalizer2 n2) {\r
  71         super(id, null);\r
  72         norm2 = n2;\r
  73     }\r
  74 \r
  75     /**\r
  76      * Implements {@link Transliterator#handleTransliterate}.\r
  77      */\r
  78     protected void handleTransliterate(Replaceable text,\r
  79                                        Position offsets, boolean isIncremental) {\r
  80         // start and limit of the input range\r
  81         int start = offsets.start;\r
  82         int limit = offsets.limit;\r
  83         if(start >= limit) {\r
  84             return;\r
  85         }\r
  86 \r
  87         /*\r
  88          * Normalize as short chunks at a time as possible even in\r
  89          * bulk mode, so that styled text is minimally disrupted.\r
  90          * In incremental mode, a chunk that ends with offsets.limit\r
  91          * must not be normalized.\r
  92          *\r
  93          * If it was known that the input text is not styled, then\r
  94          * a bulk mode normalization could be used.\r
  95          * (For details, see the comment in the C++ version.)\r
  96          */\r
  97         StringBuilder segment = new StringBuilder();\r
  98         StringBuilder normalized = new StringBuilder();\r
  99         int c = text.char32At(start);\r
 100         do {\r
 101             int prev = start;\r
 102             // Skip at least one character so we make progress.\r
 103             // c holds the character at start.\r
 104             segment.setLength(0);\r
 105             do {\r
 106                 segment.appendCodePoint(c);\r
 107                 start += Character.charCount(c);\r
 108             } while(start < limit && !norm2.hasBoundaryBefore(c = text.char32At(start)));\r
 109             if(start == limit && isIncremental && !norm2.hasBoundaryAfter(c)) {\r
 110                 // stop in incremental mode when we reach the input limit\r
 111                 // in case there are additional characters that could change the\r
 112                 // normalization result\r
 113                 start=prev;\r
 114                 break;\r
 115             }\r
 116             norm2.normalize(segment, normalized);\r
 117             if(!Normalizer2Impl.UTF16Plus.equal(segment, normalized)) {\r
 118                 // replace the input chunk with its normalized form\r
 119                 text.replace(prev, start, normalized.toString());\r
 120 \r
 121                 // update all necessary indexes accordingly\r
 122                 int delta = normalized.length() - (start - prev);\r
 123                 start += delta;\r
 124                 limit += delta;\r
 125             }\r
 126         } while(start < limit);\r
 127 \r
 128         offsets.start = start;\r
 129         offsets.contextLimit += limit - offsets.limit;\r
 130         offsets.limit = limit;\r
 131     }\r
 132 }\r