src/bin/SVMTagger.cc

   1 /*
   2  * Copyright (C) 2004 Jesus Gimenez, Lluis Marquez and Senen Moya
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * This library is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with this library; if not, write to the Free Software
  16  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  17  */
  18
  19 #include <stdio.h>
  20 #include <stdlib.h>
  21 #include <string.h>
  22 #include "hash.h"
  23 #include "list.h"
  24 #include "dict.h"
  25 #include "weight.h"
  26 #include "swindow.h"
  27 #include "tagger.h"
  28 #include "er.h"
  29
  30 #define ERROR1  "\nERROR: Value incorrect in -T option.\n"
  31 #define ERROR2  "\nERROR: Value incorrect in -S option.\n"
  32 #define ERROR3  "\nERROR: Value incorrect in -I option.\n"
  33 #define ERROR4  "\nERROR: Value incorrect in -L option.\n"
  34 #define ERROR5  "\nERROR: Value incorrect in -I or -L option. Window length have to be greater than Interest Point.\n"
  35 #define ERROR6  "\nERROR: You can not change Interes Point (-I) if you don't change the length of the window (-L).\n"
  36 #define ERROR8  "\nERROR: Incorrect Options\n"
  37 #define ERRORK  "\nERROR: Value incorrect in -K option.\n"
  38 #define ERRORU  "\nERROR: Value incorrect in -U option.\n"
  39
  40 extern int verbose_svmtool;
  41
  42 void printHelp()
  43 {
  44         fprintf(stderr,"\nSVMTool++ v 1.1.2 -- SVMTagger\n\n");
  45         fprintf(stderr,"Usage : svmt [options] <model>  < stdin > stdout\n\n");
  46         fprintf(stderr,"options:\n");
  47         fprintf(stderr,"\t-L or -l <Window Lenght>\n");
  48         fprintf(stderr,"\t\t<Window Lenght> have to be greater than 2\n");
  49         fprintf(stderr,"\t\t7 (default)\n");
  50         fprintf(stderr,"\t-I or -i <Interest Point>\n");
  51         fprintf(stderr,"\t\t<Interest Point> have to be greater than 0\n");
  52         fprintf(stderr,"\t\t4 (default)\n");
  53         fprintf(stderr,"\t-S or -s <direction>\n");
  54         fprintf(stderr,"\t\tLR\tleft-to-right\t(default)\n");
  55         fprintf(stderr,"\t\tRL\tright-to-left\n");
  56         fprintf(stderr,"\t\tLRL\tboth left-to-right and right-to-left\n");
  57         fprintf(stderr,"\t-T or -t <strategy>\n");
  58         fprintf(stderr,"\t\t0\tone-pass   [default - requires model 0]\n");
  59         fprintf(stderr,"\t\t1\ttwo-passes [revisiting results and relabeling - requires model 2 and model 1]\n");
  60         fprintf(stderr,"\t\t2\tone-pass   [robust against unknown words - requires model 0 and model 2]\n");
  61         fprintf(stderr,"\t\t3\tone-pass   [unsupervised learning models - requires model 3]\n");
  62         fprintf(stderr,"\t\t4\tone-pass   [very robust against unknown words - requires model 4]\n");
  63         fprintf(stderr,"\t\t5\tone-pass   [sentence-level likelihood - requires model 0] Not implemented!!\n");
  64         fprintf(stderr,"\t\t6\tone-pass   [robust sentence-level likelihood - requires model 4] Not implemented!!\n");
  65         fprintf(stderr,"\t-B or -b <backup_lexicon>\n");
  66         fprintf(stderr,"\t-K <n> weight filtering threshold for known words (default is 0)\n");
  67         fprintf(stderr,"\t-U <n> weight filtering threshold for unknown words (default is 0)\n");
  68         fprintf(stderr,"\t-V or -v verbose\n");
  69         fprintf(stderr,"\nmodel: model location (path/name)\n");
  70         fprintf(stderr,"\nUsage : SVMTagger -V -S LRL -T 0 /home/users/me/SVMTool/models/eng/WSJTP < WSJTP.TEST > WSJTP.TEST.OUT\n\n");
  71 }
  72
  73
  74 /*
  75 -1 -t error
  76 -2 -s error
  77 -3 -i error
  78 -4 -l error
  79 -5 length<=ip
  80 */
  81
  82 int options(int argc,char *argv[])
  83 {
  84         int isIP=0,ip=4,isLength=0,length=7;
  85
  86         for (int i=1;i<argc-1;i++)
  87         {
  88                 if (strcmp(argv[i],"-t")==0 || strcmp(argv[i],"-T")==0)
  89                 {
  90                         i++;
  91                         //if (atoi(argv[i])==0) return -1;
  92                 }
  93                 else if (strcmp(argv[i],"-s")==0 || strcmp(argv[i],"-S")==0)
  94                 {
  95                         i++;
  96                         if (!((strcmp(argv[i],"LR")==0) ||
  97                                 (strcmp(argv[i],"RL")==0)   ||
  98                                 (strcmp(argv[i],"LRL")==0)) ) return -2;
  99                 }
 100                 else if (strcmp(argv[i],"-i")==0 || strcmp(argv[i],"-I")==0)
 101                 {
 102                         i++;
 103                         if (atoi(argv[i])<1) return -3;
 104                         else ip = atoi(argv[i]);
 105                         isIP = 1;
 106                 }
 107                 else if (strcmp(argv[i],"-l")==0 || strcmp(argv[i],"-L")==0)
 108                 {
 109                         i++;
 110                         if (atoi(argv[i])<0) return -4;
 111                         else length = atoi(argv[i]);
 112                         isLength = 1;
 113                 }
 114                 else if (strcmp(argv[i],"-b")==0 || strcmp(argv[i],"-B")==0)
 115                 {
 116                         if (strcmp(argv[i+1],"-v")!=0 && strcmp(argv[i+1],"-V")!=0
 117                                 && strcmp(argv[i+1],"-b")!=0 && strcmp(argv[i+1],"-B")!=0
 118                                 && strcmp(argv[i+1],"-l")!=0 && strcmp(argv[i+1],"-L")!=0
 119                                 && strcmp(argv[i+1],"-s")!=0 && strcmp(argv[i+1],"-S")!=0
 120                                 && strcmp(argv[i+1],"-t")!=0 && strcmp(argv[i+1],"-T")!=0
 121                                 && strcmp(argv[i+1],"-i")!=0 && strcmp(argv[i+1],"-I")!=0) i++;
 122
 123                 }
 124                 else if (strcmp(argv[i],"-k")==0 || strcmp(argv[i],"-K")==0)
 125                 {
 126                         float k;
 127                         if ( sscanf(argv[i],"%.5f",&k) == EOF) return 'k';
 128                 }
 129                 else if (strcmp(argv[i],"-u")==0 || strcmp(argv[i],"-U")==0)
 130                 {
 131                         float u;
 132                         if (sscanf(argv[i],"%.5f",&u) == EOF) return 'u';
 133                 }
 134                 else if (strcmp(argv[i],"-v")!=0 && strcmp(argv[i],"-V")!=0) return -8;
 135         }
 136
 137         if (isIP && !isLength) return -6;
 138         if (isIP && isLength && length<=ip) return -5;
 139
 140         return 0;
 141 }
 142
 143
 144 int main(int argc, char *argv[])
 145 {
 146         int ret=0;
 147
 148         verbose_svmtool = FALSE;
 149
 150         erCompRegExp();
 151
 152         if ((argc<=1) || ((ret=options(argc,argv))<0))
 153         {
 154                 switch(ret)
 155                 {
 156                         case -1: fprintf(stderr,ERROR1); break;
 157                         case -2: fprintf(stderr,ERROR2); break;
 158                         case -3: fprintf(stderr,ERROR3); break;
 159                         case -4: fprintf(stderr,ERROR4); break;
 160                         case -5: fprintf(stderr,ERROR5); break;
 161                         case -6: fprintf(stderr,ERROR6); break;
 162                         case -8: fprintf(stderr,ERROR8); break;
 163                         case 'k': fprintf(stderr,ERRORK); break;
 164                         case 'u': fprintf(stderr,ERRORU); break;
 165                 }
 166
 167                 printHelp();
 168                 exit(0);
 169         }
 170
 171         tagger t(argv[argc-1]);
 172
 173         for (int i=1;i<argc-1;i++)
 174         {
 175                 if (strcmp(argv[i],"-k")==0 || strcmp(argv[i],"-K")==0)
 176                         { t.taggerPutKWeightFilter(atof(argv[i+1])); i++;}
 177                         else if (strcmp(argv[i],"-u")==0 || strcmp(argv[i],"-U")==0)
 178                                 { t.taggerPutUWeightFilter(atof(argv[i+1])); i++;}
 179                                 else if (strcmp(argv[i],"-t")==0 || strcmp(argv[i],"-T")==0)
 180                                 {
 181                         //modstat t.taggerPutStrategy(atoi(argv[i+1])-1);
 182                                         t.taggerPutStrategy(atoi(argv[i+1]));
 183                                         i++;
 184                                 }
 185                                 else if (strcmp(argv[i],"-s")==0 || strcmp(argv[i],"-S")==0)
 186                                         { t.taggerPutFlow(argv[i+1]); i++;}
 187                                         else if (strcmp(argv[i],"-v")==0 || strcmp(argv[i],"-V")==0) verbose_svmtool = TRUE;
 188                                 else if (strcmp(argv[i],"-l")==0 || strcmp(argv[i],"-L")==0)
 189                                         { t.taggerPutWinLength(atoi(argv[i+1])); i++;}
 190                                         else if (strcmp(argv[i],"-i")==0 || strcmp(argv[i],"-I")==0)
 191                                                 { t.taggerPutWinIndex(atoi(argv[i+1])); i++;}
 192                                                 else if (strcmp(argv[i],"-b")==0 || strcmp(argv[i],"-B")==0)
 193                                                         { t.taggerPutBackupDictionary(argv[i+1]); i++;}
 194         }
 195
 196         if (verbose_svmtool) fprintf(stderr,"\nSVMTool++ v 1.1.2 -- SVMTagger\n\n");
 197
 198         t.taggerLoadModelsForTagging();
 199
 200         t.taggerInit();
 201         t.taggerRun();
 202
 203         erFreeRegExp();
 204         return 0;
 205 }