i18npool/source/textconversion/textconversion_zh.cxx

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
   2 /*
   3  * This file is part of the LibreOffice project.
   4  *
   5  * This Source Code Form is subject to the terms of the Mozilla Public
   6  * License, v. 2.0. If a copy of the MPL was not distributed with this
   7  * file, You can obtain one at http://mozilla.org/MPL/2.0/.
   8  *
   9  * This file incorporates work covered by the following license notice:
  10  *
  11  *   Licensed to the Apache Software Foundation (ASF) under one or more
  12  *   contributor license agreements. See the NOTICE file distributed
  13  *   with this work for additional information regarding copyright
  14  *   ownership. The ASF licenses this file to you under the Apache
  15  *   License, Version 2.0 (the "License"); you may not use this file
  16  *   except in compliance with the License. You may obtain a copy of
  17  *   the License at http://www.apache.org/licenses/LICENSE-2.0 .
  18  */
  19
  20
  21 #include <textconversion.hxx>
  22 #include <com/sun/star/i18n/TextConversionType.hpp>
  23 #include <com/sun/star/i18n/TextConversionOption.hpp>
  24 #include <com/sun/star/lang/NoSupportException.hpp>
  25 #include <com/sun/star/linguistic2/ConversionDirection.hpp>
  26 #include <com/sun/star/linguistic2/ConversionDictionaryType.hpp>
  27 #include <com/sun/star/linguistic2/ConversionDictionaryList.hpp>
  28 #include <memory>
  29
  30 using namespace com::sun::star::lang;
  31 using namespace com::sun::star::i18n;
  32 using namespace com::sun::star::linguistic2;
  33 using namespace com::sun::star::uno;
  34
  35
  36 namespace i18npool {
  37
  38 TextConversion_zh::TextConversion_zh( const Reference < XComponentContext >& xContext )
  39     : TextConversionService("com.sun.star.i18n.TextConversion_zh")
  40 {
  41     xCDL = ConversionDictionaryList::create(xContext);
  42 }
  43
  44 static sal_Unicode getOneCharConversion(sal_Unicode ch, const sal_Unicode* Data, const sal_uInt16* Index)
  45 {
  46     if (Data && Index) {
  47         sal_Unicode address = Index[ch>>8];
  48         if (address != 0xFFFF)
  49             address = Data[address + (ch & 0xFF)];
  50         return (address != 0xFFFF) ? address : ch;
  51     } else {
  52         return ch;
  53     }
  54 }
  55
  56 #ifdef DISABLE_DYNLOADING
  57
  58 extern "C" {
  59
  60 const sal_Unicode* getSTC_CharData_T2S();
  61 const sal_uInt16* getSTC_CharIndex_T2S();
  62 const sal_Unicode* getSTC_CharData_S2V();
  63 const sal_uInt16* getSTC_CharIndex_S2V();
  64 const sal_Unicode* getSTC_CharData_S2T();
  65 const sal_uInt16* getSTC_CharIndex_S2T();
  66
  67 const sal_Unicode *getSTC_WordData(sal_Int32&);
  68
  69 const sal_uInt16 *getSTC_WordIndex_T2S(sal_Int32&);
  70 const sal_uInt16 *getSTC_WordEntry_T2S();
  71 const sal_uInt16 *getSTC_WordIndex_S2T(sal_Int32&);
  72 const sal_uInt16 *getSTC_WordEntry_S2T();
  73
  74 }
  75
  76 #endif
  77
  78 OUString
  79 TextConversion_zh::getCharConversion(const OUString& aText, sal_Int32 nStartPos, sal_Int32 nLength, bool toSChinese, sal_Int32 nConversionOptions)
  80 {
  81     const sal_Unicode *Data;
  82     const sal_uInt16 *Index;
  83
  84 #ifndef DISABLE_DYNLOADING
  85     if (toSChinese) {
  86         Data = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_T2S"))();
  87         Index = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_T2S"))();
  88     } else if (nConversionOptions & TextConversionOption::USE_CHARACTER_VARIANTS) {
  89         Data = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_S2V"))();
  90         Index = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_S2V"))();
  91     } else {
  92         Data = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_S2T"))();
  93         Index = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_S2T"))();
  94     }
  95 #else
  96     if (toSChinese) {
  97         Data = getSTC_CharData_T2S();
  98         Index = getSTC_CharIndex_T2S();
  99     } else if (nConversionOptions & TextConversionOption::USE_CHARACTER_VARIANTS) {
 100         Data = getSTC_CharData_S2V();
 101         Index = getSTC_CharIndex_S2V();
 102     } else {
 103         Data = getSTC_CharData_S2T();
 104         Index = getSTC_CharIndex_S2T();
 105     }
 106 #endif
 107
 108     rtl_uString * newStr = rtl_uString_alloc(nLength);
 109     for (sal_Int32 i = 0; i < nLength; i++)
 110         newStr->buffer[i] =
 111             getOneCharConversion(aText[nStartPos+i], Data, Index);
 112     return OUString(newStr, SAL_NO_ACQUIRE); //take ownership
 113 }
 114
 115 OUString
 116 TextConversion_zh::getWordConversion(const OUString& aText, sal_Int32 nStartPos, sal_Int32 nLength, bool toSChinese, sal_Int32 nConversionOptions, Sequence<sal_Int32>& offset)
 117 {
 118     sal_Int32 dictLen = 0;
 119     sal_Int32 maxLen = 0;
 120     const sal_uInt16 *index;
 121     const sal_uInt16 *entry;
 122     const sal_Unicode *charData;
 123     const sal_uInt16 *charIndex;
 124     bool one2one=true;
 125
 126 #ifndef DISABLE_DYNLOADING
 127     const sal_Unicode *wordData = reinterpret_cast<const sal_Unicode* (*)(sal_Int32&)>(getFunctionBySymbol("getSTC_WordData"))(dictLen);
 128     if (toSChinese) {
 129         index = reinterpret_cast<const sal_uInt16* (*)(sal_Int32&)>(getFunctionBySymbol("getSTC_WordIndex_T2S"))(maxLen);
 130         entry = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_WordEntry_T2S"))();
 131         charData = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_T2S"))();
 132         charIndex = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_T2S"))();
 133     } else {
 134         index = reinterpret_cast<const sal_uInt16* (*)(sal_Int32&)>(getFunctionBySymbol("getSTC_WordIndex_S2T"))(maxLen);
 135         entry = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_WordEntry_S2T"))();
 136         if (nConversionOptions & TextConversionOption::USE_CHARACTER_VARIANTS) {
 137             charData = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_S2V"))();
 138             charIndex = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_S2V"))();
 139         } else {
 140             charData = reinterpret_cast<const sal_Unicode* (*)()>(getFunctionBySymbol("getSTC_CharData_S2T"))();
 141             charIndex = reinterpret_cast<const sal_uInt16* (*)()>(getFunctionBySymbol("getSTC_CharIndex_S2T"))();
 142         }
 143     }
 144 #else
 145     const sal_Unicode *wordData = getSTC_WordData(dictLen);
 146     if (toSChinese) {
 147         index = getSTC_WordIndex_T2S(maxLen);
 148         entry = getSTC_WordEntry_T2S();
 149         charData = getSTC_CharData_T2S();
 150         charIndex = getSTC_CharIndex_T2S();
 151     } else {
 152         index = getSTC_WordIndex_S2T(maxLen);
 153         entry = getSTC_WordEntry_S2T();
 154         if (nConversionOptions & TextConversionOption::USE_CHARACTER_VARIANTS) {
 155             charData = getSTC_CharData_S2V();
 156             charIndex = getSTC_CharIndex_S2V();
 157         } else {
 158             charData = getSTC_CharData_S2T();
 159             charIndex = getSTC_CharIndex_S2T();
 160         }
 161     }
 162 #endif
 163
 164     if ((!wordData || !index || !entry) && !xCDL.is()) // no word mapping defined, do char2char conversion.
 165         return getCharConversion(aText, nStartPos, nLength, toSChinese, nConversionOptions);
 166
 167     std::unique_ptr<sal_Unicode[]> newStr(new sal_Unicode[nLength * 2 + 1]);
 168     sal_Int32 currPos = 0, count = 0;
 169     auto offsetRange = asNonConstRange(offset);
 170     while (currPos < nLength) {
 171         sal_Int32 len = nLength - currPos;
 172         bool found = false;
 173         if (len > maxLen)
 174             len = maxLen;
 175         for (; len > 0 && ! found; len--) {
 176             OUString word = aText.copy(nStartPos + currPos, len);
 177             sal_Int32 current = 0;
 178             // user dictionary
 179             if (xCDL.is()) {
 180                 Sequence < OUString > conversions;
 181                 try {
 182                     conversions = xCDL->queryConversions(word, 0, len,
 183                             aLocale, ConversionDictionaryType::SCHINESE_TCHINESE,
 184                             /*toSChinese ?*/ ConversionDirection_FROM_LEFT /*: ConversionDirection_FROM_RIGHT*/,
 185                             nConversionOptions);
 186                 }
 187                 catch ( NoSupportException & ) {
 188                     // clear reference (when there is no user dictionary) in order
 189                     // to not always have to catch this exception again
 190                     // in further calls. (save time)
 191                     xCDL = nullptr;
 192                 }
 193                 catch (...) {
 194                     // catch all other exceptions to allow
 195                     // querying the system dictionary in the next line
 196                 }
 197                 if (conversions.hasElements()) {
 198                     if (offset.hasElements()) {
 199                         if (word.getLength() != conversions[0].getLength())
 200                             one2one=false;
 201                         while (current < conversions[0].getLength()) {
 202                             offsetRange[count] = nStartPos + currPos + (current *
 203                                     word.getLength() / conversions[0].getLength());
 204                             newStr[count++] = conversions[0][current++];
 205                         }
 206                         // offset[count-1] = nStartPos + currPos + word.getLength() - 1;
 207                     } else {
 208                         while (current < conversions[0].getLength())
 209                             newStr[count++] = conversions[0][current++];
 210                     }
 211                     currPos += word.getLength();
 212                     found = true;
 213                 }
 214             }
 215
 216             if (wordData && !found && index[len+1] - index[len] > 0) {
 217                 sal_Int32 bottom = static_cast<sal_Int32>(index[len]);
 218                 sal_Int32 top = static_cast<sal_Int32>(index[len+1]) - 1;
 219
 220                 while (bottom <= top && !found) {
 221                     current = (top + bottom) / 2;
 222                     const sal_Int32 result = rtl_ustr_compare(
 223                         word.getStr(), wordData + entry[current]);
 224                     if (result < 0)
 225                         top = current - 1;
 226                     else if (result > 0)
 227                         bottom = current + 1;
 228                     else {
 229                         if (toSChinese)   // Traditionary/Simplified conversion,
 230                             for (current = entry[current]-1; current > 0 && wordData[current-1]; current--) ;
 231                         else  // Simplified/Traditionary conversion, forwards search for next word
 232                             current = entry[current] + word.getLength() + 1;
 233                         sal_Int32 start=current;
 234                         if (offset.hasElements()) {
 235                             if (word.getLength() != static_cast<sal_Int32>(std::u16string_view(&wordData[current]).size()))
 236                                 one2one=false;
 237                             sal_Int32 convertedLength=std::u16string_view(&wordData[current]).size();
 238                             while (wordData[current]) {
 239                                 offsetRange[count]=nStartPos + currPos + ((current-start) *
 240                                     word.getLength() / convertedLength);
 241                                 newStr[count++] = wordData[current++];
 242                             }
 243                             // offset[count-1]=nStartPos + currPos + word.getLength() - 1;
 244                         } else {
 245                             while (wordData[current])
 246                                 newStr[count++] = wordData[current++];
 247                         }
 248                         currPos += word.getLength();
 249                         found = true;
 250                     }
 251                 }
 252             }
 253         }
 254         if (!found) {
 255             if (offset.hasElements())
 256                 offsetRange[count]=nStartPos+currPos;
 257             newStr[count++] =
 258                 getOneCharConversion(aText[nStartPos+currPos], charData, charIndex);
 259             currPos++;
 260         }
 261     }
 262     if (offset.hasElements())
 263         offset.realloc(one2one ? 0 : count);
 264     OUString aRet(newStr.get(), count);
 265     return aRet;
 266 }
 267
 268 TextConversionResult SAL_CALL
 269 TextConversion_zh::getConversions( const OUString& aText, sal_Int32 nStartPos, sal_Int32 nLength,
 270     const Locale& rLocale, sal_Int16 nConversionType, sal_Int32 nConversionOptions)
 271 {
 272     TextConversionResult result;
 273
 274     result.Candidates =
 275         { getConversion( aText, nStartPos, nLength, rLocale, nConversionType, nConversionOptions) };
 276     result.Boundary.startPos = nStartPos;
 277     result.Boundary.endPos = nStartPos + nLength;
 278
 279     return result;
 280 }
 281
 282 OUString SAL_CALL
 283 TextConversion_zh::getConversion( const OUString& aText, sal_Int32 nStartPos, sal_Int32 nLength,
 284     const Locale& rLocale, sal_Int16 nConversionType, sal_Int32 nConversionOptions)
 285 {
 286     if (rLocale.Language != "zh" || ( nConversionType != TextConversionType::TO_SCHINESE && nConversionType != TextConversionType::TO_TCHINESE) )
 287         throw NoSupportException(); // Conversion type is not supported in this service.
 288
 289     aLocale=rLocale;
 290     bool toSChinese = nConversionType == TextConversionType::TO_SCHINESE;
 291
 292     if (nConversionOptions & TextConversionOption::CHARACTER_BY_CHARACTER)
 293         // char to char dictionary
 294         return getCharConversion(aText, nStartPos, nLength, toSChinese, nConversionOptions);
 295     else {
 296         Sequence <sal_Int32> offset;
 297         // word to word dictionary
 298         return  getWordConversion(aText, nStartPos, nLength, toSChinese, nConversionOptions, offset);
 299     }
 300 }
 301
 302 OUString SAL_CALL
 303 TextConversion_zh::getConversionWithOffset( const OUString& aText, sal_Int32 nStartPos, sal_Int32 nLength,
 304     const Locale& rLocale, sal_Int16 nConversionType, sal_Int32 nConversionOptions, Sequence<sal_Int32>& offset)
 305 {
 306     if (rLocale.Language != "zh" || ( nConversionType != TextConversionType::TO_SCHINESE && nConversionType != TextConversionType::TO_TCHINESE) )
 307         throw NoSupportException(); // Conversion type is not supported in this service.
 308
 309     aLocale=rLocale;
 310     bool toSChinese = nConversionType == TextConversionType::TO_SCHINESE;
 311
 312     if (nConversionOptions & TextConversionOption::CHARACTER_BY_CHARACTER) {
 313         offset.realloc(0);
 314         // char to char dictionary
 315         return getCharConversion(aText, nStartPos, nLength, toSChinese, nConversionOptions);
 316     } else {
 317         if (offset.getLength() < 2*nLength)
 318             offset.realloc(2*nLength);
 319         // word to word dictionary
 320         return  getWordConversion(aText, nStartPos, nLength, toSChinese, nConversionOptions, offset);
 321     }
 322 }
 323
 324 sal_Bool SAL_CALL
 325 TextConversion_zh::interactiveConversion( const Locale& /*rLocale*/, sal_Int16 /*nTextConversionType*/, sal_Int32 /*nTextConversionOptions*/ )
 326 {
 327     return false;
 328 }
 329
 330 }
 331
 332 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */