i18npool/source/characterclassification/scripttypedetector.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2000, 2010 Oracle and/or its affiliates.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * This file is part of OpenOffice.org.
  10  *
  11  * OpenOffice.org is free software: you can redistribute it and/or modify
  12  * it under the terms of the GNU Lesser General Public License version 3
  13  * only, as published by the Free Software Foundation.
  14  *
  15  * OpenOffice.org is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU Lesser General Public License version 3 for more details
  19  * (a copy is included in the LICENSE file that accompanied this code).
  20  *
  21  * You should have received a copy of the GNU Lesser General Public License
  22  * version 3 along with OpenOffice.org.  If not, see
  23  * <http://www.openoffice.org/license.html>
  24  * for a copy of the LGPLv3 License.
  25  *
  26  ************************************************************************/
  27
  28 // MARKER(update_precomp.py): autogen include statement, do not remove
  29 #include "precompiled_i18npool.hxx"
  30
  31 #include <com/sun/star/i18n/CTLScriptType.hpp>
  32 #include <com/sun/star/i18n/ScriptDirection.hpp>
  33 #include <com/sun/star/i18n/UnicodeScript.hpp>
  34 #include <scripttypedetector.hxx>
  35 #include <i18nutil/unicode.hxx>
  36
  37 //      ----------------------------------------------------
  38 //      class ScriptTypeDetector
  39 //      ----------------------------------------------------;
  40
  41 using namespace com::sun::star::i18n;
  42
  43 ScriptTypeDetector::ScriptTypeDetector()
  44 {
  45 }
  46
  47 ScriptTypeDetector::~ScriptTypeDetector()
  48 {
  49 }
  50
  51 static sal_Int16 scriptDirection[] = {
  52     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT = 0,
  53     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT = 1,
  54     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER = 2,
  55     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER_SEPARATOR = 3,
  56     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_EUROPEAN_NUMBER_TERMINATOR = 4,
  57     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_ARABIC_NUMBER = 5,
  58     ScriptDirection::NEUTRAL,           // DirectionProperty_COMMON_NUMBER_SEPARATOR = 6,
  59     ScriptDirection::NEUTRAL,           // DirectionProperty_BLOCK_SEPARATOR = 7,
  60     ScriptDirection::NEUTRAL,           // DirectionProperty_SEGMENT_SEPARATOR = 8,
  61     ScriptDirection::NEUTRAL,           // DirectionProperty_WHITE_SPACE_NEUTRAL = 9,
  62     ScriptDirection::NEUTRAL,           // DirectionProperty_OTHER_NEUTRAL = 10,
  63     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT_EMBEDDING = 11,
  64     ScriptDirection::LEFT_TO_RIGHT,     // DirectionProperty_LEFT_TO_RIGHT_OVERRIDE = 12,
  65     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_ARABIC = 13,
  66     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_EMBEDDING = 14,
  67     ScriptDirection::RIGHT_TO_LEFT,     // DirectionProperty_RIGHT_TO_LEFT_OVERRIDE = 15,
  68     ScriptDirection::NEUTRAL,           // DirectionProperty_POP_DIRECTIONAL_FORMAT = 16,
  69     ScriptDirection::NEUTRAL,           // DirectionProperty_DIR_NON_SPACING_MARK = 17,
  70     ScriptDirection::NEUTRAL,           // DirectionProperty_BOUNDARY_NEUTRAL = 18,
  71 };
  72
  73 sal_Int16 SAL_CALL
  74 ScriptTypeDetector::getScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 defaultScriptDirection ) throw (::com::sun::star::uno::RuntimeException)
  75 {
  76     sal_Int16 dir = scriptDirection[unicode::getUnicodeDirection(Text[nPos])];
  77     return (dir == ScriptDirection::NEUTRAL) ? defaultScriptDirection : dir;
  78 }
  79
  80 // return value '-1' means either the direction on nPos is not same as scriptDirection or nPos is out of range.
  81 sal_Int32 SAL_CALL
  82 ScriptTypeDetector::beginOfScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 direction ) throw (::com::sun::star::uno::RuntimeException)
  83 {
  84         sal_Int32 cPos = nPos;
  85
  86         if (cPos < Text.getLength()) {
  87             for (; cPos >= 0; cPos--) {
  88                 if (direction != getScriptDirection(Text, cPos, direction))
  89                     break;
  90             }
  91         }
  92         return cPos == nPos ? -1 : cPos + 1;
  93 }
  94
  95 sal_Int32 SAL_CALL
  96 ScriptTypeDetector::endOfScriptDirection( const ::rtl::OUString& Text, sal_Int32 nPos, sal_Int16 direction ) throw (::com::sun::star::uno::RuntimeException)
  97 {
  98         sal_Int32 cPos = nPos;
  99         sal_Int32 len = Text.getLength();
 100
 101         if (cPos >=0) {
 102             for (; cPos < len; cPos++) {
 103                 if (direction != getScriptDirection(Text, cPos, direction))
 104                     break;
 105             }
 106         }
 107         return cPos == nPos ? -1 : cPos;
 108 }
 109
 110 sal_Int16 SAL_CALL
 111 ScriptTypeDetector::getCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 112 {
 113     static ScriptTypeList typeList[] = {
 114         { UnicodeScript_kHebrew, UnicodeScript_kHebrew, CTLScriptType::CTL_HEBREW },    // 10
 115         { UnicodeScript_kArabic, UnicodeScript_kArabic, CTLScriptType::CTL_ARABIC },    // 11
 116         { UnicodeScript_kDevanagari, UnicodeScript_kDevanagari, CTLScriptType::CTL_INDIC },     // 14
 117         { UnicodeScript_kThai, UnicodeScript_kThai, CTLScriptType::CTL_THAI },      // 24
 118         { UnicodeScript_kScriptCount, UnicodeScript_kScriptCount,   CTLScriptType::CTL_UNKNOWN }    // 88
 119     };
 120
 121     return unicode::getUnicodeScriptType(Text[nPos], typeList, CTLScriptType::CTL_UNKNOWN);
 122 }
 123
 124 // Begin of Script Type is inclusive.
 125 sal_Int32 SAL_CALL
 126 ScriptTypeDetector::beginOfCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 127 {
 128     if (nPos < 0)
 129         return 0;
 130     else if (nPos >= Text.getLength())
 131         return Text.getLength();
 132     else {
 133         sal_Int16 cType = getCTLScriptType(Text, nPos);
 134         for (nPos--; nPos >= 0; nPos--) {
 135             if (cType != getCTLScriptType(Text, nPos))
 136                 break;
 137         }
 138         return nPos + 1;
 139     }
 140 }
 141
 142 // End of the Script Type is exclusive, the return value pointing to the begin of next script type
 143 sal_Int32 SAL_CALL
 144 ScriptTypeDetector::endOfCTLScriptType( const ::rtl::OUString& Text, sal_Int32 nPos ) throw (::com::sun::star::uno::RuntimeException)
 145 {
 146     if (nPos < 0)
 147         return 0;
 148     else if (nPos >= Text.getLength())
 149         return Text.getLength();
 150     else {
 151         sal_Int16 cType = getCTLScriptType(Text, nPos);
 152         sal_Int32 len = Text.getLength();
 153         for (nPos++; nPos < len; nPos++) {
 154             if (cType != getCTLScriptType(Text, nPos))
 155                 break;
 156         }
 157         return nPos;
 158     }
 159 }
 160
 161 const sal_Char sDetector[] = "draft.com.sun.star.i18n.ScriptTypeDetector";
 162
 163 rtl::OUString SAL_CALL
 164 ScriptTypeDetector::getImplementationName() throw( ::com::sun::star::uno::RuntimeException )
 165 {
 166     return ::rtl::OUString::createFromAscii(sDetector);
 167 }
 168
 169 sal_Bool SAL_CALL
 170 ScriptTypeDetector::supportsService(const rtl::OUString& ServiceName) throw( ::com::sun::star::uno::RuntimeException )
 171 {
 172     return !ServiceName.compareToAscii(sDetector);
 173 }
 174
 175 ::com::sun::star::uno::Sequence< rtl::OUString > SAL_CALL
 176 ScriptTypeDetector::getSupportedServiceNames() throw( ::com::sun::star::uno::RuntimeException )
 177 {
 178     ::com::sun::star::uno::Sequence< ::rtl::OUString > aRet(1);
 179     aRet[0] = ::rtl::OUString::createFromAscii(sDetector);
 180     return aRet;
 181 }
 182