icuSources/tools/genctd/genctd.cpp

   1 /*
   2 **********************************************************************
   3 *   Copyright (C) 2002-2006, International Business Machines
   4 *   Corporation and others.  All Rights Reserved.
   5 **********************************************************************
   6 *
   7 * File genctd.c
   8 */
   9
  10 //--------------------------------------------------------------------
  11 //
  12 //   Tool for generating CompactTrieDictionary data files (.ctd files).
  13 //
  14 //   Usage:  genctd [options] -o output-file.ctd input-file
  15 //
  16 //       options:   -v         verbose
  17 //                  -? or -h   help
  18 //
  19 //   The input  file is a plain text file containing words, one per line.
  20 //    Words end at the first whitespace; lines beginning with whitespace
  21 //    are ignored.
  22 //    The file can be encoded as utf-8, or utf-16 (either endian), or
  23 //    in the default code page (platform dependent.).  utf encoded
  24 //    files must include a BOM.
  25 //
  26 //--------------------------------------------------------------------
  27
  28 #include "unicode/utypes.h"
  29 #include "unicode/uchar.h"
  30 #include "unicode/ucnv.h"
  31 #include "unicode/uniset.h"
  32 #include "unicode/unistr.h"
  33 #include "unicode/uclean.h"
  34 #include "unicode/udata.h"
  35 #include "unicode/putil.h"
  36
  37 #include "uoptions.h"
  38 #include "unewdata.h"
  39 #include "ucmndata.h"
  40 #include "rbbidata.h"
  41 #include "triedict.h"
  42 #include "cmemory.h"
  43
  44 #include <stdio.h>
  45 #include <stdlib.h>
  46 #include <string.h>
  47
  48 static char *progName;
  49 static UOption options[]={
  50     UOPTION_HELP_H,             /* 0 */
  51     UOPTION_HELP_QUESTION_MARK, /* 1 */
  52     UOPTION_VERBOSE,            /* 2 */
  53     { "out",   NULL, NULL, NULL, 'o', UOPT_REQUIRES_ARG, 0 },   /* 3 */
  54     UOPTION_ICUDATADIR,         /* 4 */
  55     UOPTION_DESTDIR,            /* 5 */
  56     UOPTION_COPYRIGHT,          /* 6 */
  57 };
  58
  59 void usageAndDie(int retCode) {
  60         printf("Usage: %s [-v] [-options] -o output-file dictionary-file\n", progName);
  61         printf("\tRead in word list and write out compact trie dictionary\n"
  62             "options:\n"
  63             "\t-h or -? or --help  this usage text\n"
  64             "\t-V or --version     show a version message\n"
  65             "\t-c or --copyright   include a copyright notice\n"
  66             "\t-v or --verbose     turn on verbose output\n"
  67             "\t-i or --icudatadir  directory for locating any needed intermediate data files,\n"
  68             "\t                    followed by path, defaults to %s\n"
  69             "\t-d or --destdir     destination directory, followed by the path\n",
  70             u_getDataDirectory());
  71         exit (retCode);
  72 }
  73
  74
  75 #if UCONFIG_NO_BREAK_ITERATION
  76
  77 /* dummy UDataInfo cf. udata.h */
  78 static UDataInfo dummyDataInfo = {
  79     sizeof(UDataInfo),
  80     0,
  81
  82     U_IS_BIG_ENDIAN,
  83     U_CHARSET_FAMILY,
  84     U_SIZEOF_UCHAR,
  85     0,
  86
  87     { 0, 0, 0, 0 },                 /* dummy dataFormat */
  88     { 0, 0, 0, 0 },                 /* dummy formatVersion */
  89     { 0, 0, 0, 0 }                  /* dummy dataVersion */
  90 };
  91
  92 #else
  93
  94 //
  95 //  Set up the ICU data header, defined in ucmndata.h
  96 //
  97 DataHeader dh ={
  98     {sizeof(DataHeader),           // Struct MappedData
  99         0xda,
 100         0x27},
 101
 102     {                               // struct UDataInfo
 103         sizeof(UDataInfo),          //     size
 104         0,                          //     reserved
 105         U_IS_BIG_ENDIAN,
 106         U_CHARSET_FAMILY,
 107         U_SIZEOF_UCHAR,
 108         0,                          //     reserved
 109
 110     { 0x54, 0x72, 0x44, 0x63 },     // "TrDc" Trie Dictionary
 111     { 1, 0, 0, 0 },                 // 1.0.0.0
 112     { 0, 0, 0, 0 },                 // Irrelevant for this data type
 113     }};
 114
 115 #endif
 116
 117 //----------------------------------------------------------------------------
 118 //
 119 //  main      for genctd
 120 //
 121 //----------------------------------------------------------------------------
 122 int  main(int argc, char **argv) {
 123     UErrorCode  status = U_ZERO_ERROR;
 124     const char *wordFileName;
 125     const char *outFileName;
 126     const char *outDir = NULL;
 127     const char *copyright = NULL;
 128
 129     //
 130     // Pick up and check the command line arguments,
 131     //    using the standard ICU tool utils option handling.
 132     //
 133     U_MAIN_INIT_ARGS(argc, argv);
 134     progName = argv[0];
 135     argc=u_parseArgs(argc, argv, sizeof(options)/sizeof(options[0]), options);
 136     if(argc<0) {
 137         // Unrecognized option
 138         fprintf(stderr, "error in command line argument \"%s\"\n", argv[-argc]);
 139         usageAndDie(U_ILLEGAL_ARGUMENT_ERROR);
 140     }
 141
 142     if(options[0].doesOccur || options[1].doesOccur) {
 143         //  -? or -h for help.
 144         usageAndDie(0);
 145     }
 146
 147     if (!options[3].doesOccur || argc < 2) {
 148         fprintf(stderr, "input and output file must both be specified.\n");
 149         usageAndDie(U_ILLEGAL_ARGUMENT_ERROR);
 150     }
 151     outFileName  = options[3].value;
 152     wordFileName = argv[1];
 153
 154     if (options[4].doesOccur) {
 155         u_setDataDirectory(options[4].value);
 156     }
 157
 158     /* Initialize ICU */
 159     u_init(&status);
 160     if (U_FAILURE(status)) {
 161         fprintf(stderr, "%s: can not initialize ICU.  status = %s\n",
 162             argv[0], u_errorName(status));
 163         exit(1);
 164     }
 165     status = U_ZERO_ERROR;
 166
 167     /* Combine the directory with the file name */
 168     if(options[5].doesOccur) {
 169         outDir = options[5].value;
 170     }
 171     if (options[6].doesOccur) {
 172         copyright = U_COPYRIGHT_STRING;
 173     }
 174
 175 #if UCONFIG_NO_BREAK_ITERATION
 176
 177     UNewDataMemory *pData;
 178     char msg[1024];
 179
 180     /* write message with just the name */
 181     sprintf(msg, "genctd writes dummy %s because of UCONFIG_NO_BREAK_ITERATION, see uconfig.h", outFileName);
 182     fprintf(stderr, "%s\n", msg);
 183
 184     /* write the dummy data file */
 185     pData = udata_create(outDir, NULL, outFileName, &dummyDataInfo, NULL, &status);
 186     udata_writeBlock(pData, msg, strlen(msg));
 187     udata_finish(pData, &status);
 188     return (int)status;
 189
 190 #else
 191
 192     //
 193     //  Read in the dictionary source file
 194     //
 195     long        result;
 196     long        wordFileSize;
 197     FILE        *file;
 198     char        *wordBufferC;
 199
 200     file = fopen(wordFileName, "rb");
 201     if( file == 0 ) {
 202         fprintf(stderr, "Could not open file \"%s\"\n", wordFileName);
 203         exit(-1);
 204     }
 205     fseek(file, 0, SEEK_END);
 206     wordFileSize = ftell(file);
 207     fseek(file, 0, SEEK_SET);
 208     wordBufferC = new char[wordFileSize+10];
 209
 210     result = (long)fread(wordBufferC, 1, wordFileSize, file);
 211     if (result != wordFileSize)  {
 212         fprintf(stderr, "Error reading file \"%s\"\n", wordFileName);
 213         exit (-1);
 214     }
 215     wordBufferC[wordFileSize]=0;
 216     fclose(file);
 217
 218     //
 219     // Look for a Unicode Signature (BOM) on the word file
 220     //
 221     int32_t        signatureLength;
 222     const char *   wordSourceC = wordBufferC;
 223     const char*    encoding = ucnv_detectUnicodeSignature(
 224                            wordSourceC, wordFileSize, &signatureLength, &status);
 225     if (U_FAILURE(status)) {
 226         exit(status);
 227     }
 228     if(encoding!=NULL ){
 229         wordSourceC  += signatureLength;
 230         wordFileSize -= signatureLength;
 231     }
 232
 233     //
 234     // Open a converter to take the rule file to UTF-16
 235     //
 236     UConverter* conv;
 237     conv = ucnv_open(encoding, &status);
 238     if (U_FAILURE(status)) {
 239         fprintf(stderr, "ucnv_open: ICU Error \"%s\"\n", u_errorName(status));
 240         exit(status);
 241     }
 242
 243     //
 244     // Convert the words to UChar.
 245     //  Preflight first to determine required buffer size.
 246     //
 247     uint32_t destCap = ucnv_toUChars(conv,
 248                        NULL,           //  dest,
 249                        0,              //  destCapacity,
 250                        wordSourceC,
 251                        wordFileSize,
 252                        &status);
 253     if (status != U_BUFFER_OVERFLOW_ERROR) {
 254         fprintf(stderr, "ucnv_toUChars: ICU Error \"%s\"\n", u_errorName(status));
 255         exit(status);
 256     };
 257
 258     status = U_ZERO_ERROR;
 259     UChar *wordSourceU = new UChar[destCap+1];
 260     ucnv_toUChars(conv,
 261                   wordSourceU,     //  dest,
 262                   destCap+1,
 263                   wordSourceC,
 264                   wordFileSize,
 265                   &status);
 266     if (U_FAILURE(status)) {
 267         fprintf(stderr, "ucnv_toUChars: ICU Error \"%s\"\n", u_errorName(status));
 268         exit(status);
 269     };
 270     ucnv_close(conv);
 271
 272     // Get rid of the original file buffer
 273     delete[] wordBufferC;
 274
 275     // Create a MutableTrieDictionary, and loop through all the lines, inserting
 276     // words.
 277
 278     // First, pick a median character.
 279     UChar *current = wordSourceU + (destCap/2);
 280     UChar uc = *current++;
 281     UnicodeSet breaks;
 282     breaks.add(0x000A);     // Line Feed
 283     breaks.add(0x000D);     // Carriage Return
 284     breaks.add(0x2028);     // Line Separator
 285     breaks.add(0x2029);     // Paragraph Separator
 286
 287     do {
 288         // Look for line break
 289         while (uc && !breaks.contains(uc)) {
 290             uc = *current++;
 291         }
 292         // Now skip to first non-line-break
 293         while (uc && breaks.contains(uc)) {
 294             uc = *current++;
 295         }
 296     }
 297     while (uc && (breaks.contains(uc) || u_isspace(uc)));
 298
 299     MutableTrieDictionary *mtd = new MutableTrieDictionary(uc, status);
 300
 301     if (U_FAILURE(status)) {
 302         fprintf(stderr, "new MutableTrieDictionary: ICU Error \"%s\"\n", u_errorName(status));
 303         exit(status);
 304     }
 305
 306     // Now add the words. Words are non-space characters at the beginning of
 307     // lines, and must be at least one UChar.
 308     current = wordSourceU;
 309     UChar *candidate = current;
 310     uc = *current++;
 311     int32_t length = 0;
 312
 313     while (uc) {
 314         while (uc && !u_isspace(uc)) {
 315             ++length;
 316             uc = *current++;
 317         }
 318         if (length > 0) {
 319             mtd->addWord(candidate, length, status);
 320             if (U_FAILURE(status)) {
 321                 fprintf(stderr, "MutableTrieDictionary::addWord: ICU Error \"%s\"\n",
 322                         u_errorName(status));
 323                 exit(status);
 324             }
 325         }
 326         // Find beginning of next line
 327         while (uc && !breaks.contains(uc)) {
 328             uc = *current++;
 329         }
 330         while (uc && breaks.contains(uc)) {
 331             uc = *current++;
 332         }
 333         candidate = current-1;
 334         length = 0;
 335     }
 336
 337     // Get rid of the Unicode text buffer
 338     delete[] wordSourceU;
 339
 340     // Now, create a CompactTrieDictionary from the mutable dictionary
 341     CompactTrieDictionary *ctd = new CompactTrieDictionary(*mtd, status);
 342     if (U_FAILURE(status)) {
 343         fprintf(stderr, "new CompactTrieDictionary: ICU Error \"%s\"\n", u_errorName(status));
 344         exit(status);
 345     }
 346
 347     // Get rid of the MutableTrieDictionary
 348     delete mtd;
 349
 350     //
 351     //  Get the binary data from the dictionary.
 352     //
 353     uint32_t        outDataSize = ctd->dataSize();
 354     const uint8_t  *outData = (const uint8_t *)ctd->data();
 355
 356     //
 357     //  Create the output file
 358     //
 359     size_t bytesWritten;
 360     UNewDataMemory *pData;
 361     pData = udata_create(outDir, NULL, outFileName, &(dh.info), copyright, &status);
 362     if(U_FAILURE(status)) {
 363         fprintf(stderr, "genctd: Could not open output file \"%s\", \"%s\"\n",
 364                          outFileName, u_errorName(status));
 365         exit(status);
 366     }
 367
 368
 369     //  Write the data itself.
 370     udata_writeBlock(pData, outData, outDataSize);
 371     // finish up
 372     bytesWritten = udata_finish(pData, &status);
 373     if(U_FAILURE(status)) {
 374         fprintf(stderr, "genctd: error \"%s\" writing the output file\n", u_errorName(status));
 375         exit(status);
 376     }
 377
 378     if (bytesWritten != outDataSize) {
 379         fprintf(stderr, "Error writing to output file \"%s\"\n", outFileName);
 380         exit(-1);
 381     }
 382
 383     // Get rid of the CompactTrieDictionary
 384     delete ctd;
 385
 386     u_cleanup();
 387
 388     printf("genctd: tool completed successfully.\n");
 389     return 0;
 390
 391 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */
 392 }
 393