jars/icu4j-52_1/main/classes/translit/src/com/ibm/icu/text/SourceTargetUtility.java

   1 /*
   2  *******************************************************************************
   3  * Copyright (C) 2010-2011, Google, International Business Machines            *
   4  * Corporation and others. All Rights Reserved.                                *
   5  *******************************************************************************
   6  */
   7 package com.ibm.icu.text;
   8
   9 import java.util.HashSet;
  10 import java.util.Set;
  11
  12 import com.ibm.icu.lang.CharSequences;
  13
  14 /**
  15  * Simple internal utility class for helping with getSource/TargetSet
  16  */
  17 class SourceTargetUtility {
  18     final Transform<String, String> transform;
  19     final UnicodeSet sourceCache;
  20     final Set<String> sourceStrings;
  21     static final UnicodeSet NON_STARTERS = new UnicodeSet("[:^ccc=0:]").freeze();
  22     static Normalizer2 NFC = Normalizer2.getNFCInstance();
  23     //static final UnicodeSet TRAILING_COMBINING = new UnicodeSet();
  24
  25     public SourceTargetUtility(Transform<String, String> transform) {
  26         this(transform, null);
  27     }
  28
  29     public SourceTargetUtility(Transform<String, String> transform, Normalizer2 normalizer) {
  30         this.transform = transform;
  31         if (normalizer != null) {
  32 //            synchronized (SourceTargetUtility.class) {
  33 //                if (NFC == null) {
  34 //                    NFC = Normalizer2.getInstance(null, "nfc", Mode.COMPOSE);
  35 //                    for (int i = 0; i <= 0x10FFFF; ++i) {
  36 //                        String d = NFC.getDecomposition(i);
  37 //                        if (d == null) {
  38 //                            continue;
  39 //                        }
  40 //                        String s = NFC.normalize(d);
  41 //                        if (!CharSequences.equals(i, s)) {
  42 //                            continue;
  43 //                        }
  44 //                        // composes
  45 //                        boolean first = false;
  46 //                        for (int trailing : CharSequences.codePoints(d)) {
  47 //                            if (first) {
  48 //                                first = false;
  49 //                            } else {
  50 //                                TRAILING_COMBINING.add(trailing);
  51 //                            }
  52 //                        }
  53 //                    }
  54 //                }
  55 //            }
  56             sourceCache = new UnicodeSet("[:^ccc=0:]");
  57         } else {
  58             sourceCache = new UnicodeSet();
  59         }
  60         sourceStrings = new HashSet<String>();
  61         for (int i = 0; i <= 0x10FFFF; ++i) {
  62             String s = transform.transform(UTF16.valueOf(i));
  63             boolean added = false;
  64             if (!CharSequences.equals(i, s)) {
  65                 sourceCache.add(i);
  66                 added = true;
  67             }
  68             if (normalizer == null) {
  69                 continue;
  70             }
  71             String d = NFC.getDecomposition(i);
  72             if (d == null) {
  73                 continue;
  74             }
  75             s = transform.transform(d);
  76             if (!d.equals(s)) {
  77                 sourceStrings.add(d);
  78             }
  79             if (added) {
  80                 continue;
  81             }
  82             if (!normalizer.isInert(i)) {
  83                 sourceCache.add(i);
  84                 continue;
  85             }
  86             // see if any of the non-starters change s; if so, add i
  87 //            for (String ns : TRAILING_COMBINING) {
  88 //                String s2 = transform.transform(s + ns);
  89 //                if (!s2.startsWith(s)) {
  90 //                    sourceCache.add(i);
  91 //                    break;
  92 //                }
  93 //            }
  94
  95             // int endOfFirst = CharSequences.onCharacterBoundary(d, 1) ? 1 : 2;
  96             // if (endOfFirst >= d.length()) {
  97             // continue;
  98             // }
  99             // // now add all initial substrings
 100             // for (int j = 1; j < d.length(); ++j) {
 101             // if (!CharSequences.onCharacterBoundary(d, j)) {
 102             // continue;
 103             // }
 104             // String dd = d.substring(0,j);
 105             // s = transform.transform(dd);
 106             // if (!dd.equals(s)) {
 107             // sourceStrings.add(dd);
 108             // }
 109             // }
 110         }
 111         sourceCache.freeze();
 112     }
 113
 114     public void addSourceTargetSet(Transliterator transliterator, UnicodeSet inputFilter, UnicodeSet sourceSet,
 115             UnicodeSet targetSet) {
 116         UnicodeSet myFilter = transliterator.getFilterAsUnicodeSet(inputFilter);
 117         UnicodeSet affectedCharacters = new UnicodeSet(sourceCache).retainAll(myFilter);
 118         sourceSet.addAll(affectedCharacters);
 119         for (String s : affectedCharacters) {
 120             targetSet.addAll(transform.transform(s));
 121         }
 122         for (String s : sourceStrings) {
 123             if (myFilter.containsAll(s)) {
 124                 String t = transform.transform(s);
 125                 if (!s.equals(t)) {
 126                     targetSet.addAll(t);
 127                     sourceSet.addAll(s);
 128                 }
 129             }
 130         }
 131     }
 132 }