Subversion Repositories wimsdev

Rev

Rev 3808 | Rev 8161 | Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | RSS feed

  1. /*    Copyright (C) 1998-2003 XIAO, Gang of Universite de Nice - Sophia Antipolis
  2.  *
  3.  *  This program is free software; you can redistribute it and/or modify
  4.  *  it under the terms of the GNU General Public License as published by
  5.  *  the Free Software Foundation; either version 2 of the License, or
  6.  *  (at your option) any later version.
  7.  *
  8.  *  This program is distributed in the hope that it will be useful,
  9.  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  10.  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  11.  *  GNU General Public License for more details.
  12.  *
  13.  *  You should have received a copy of the GNU General Public License
  14.  *  along with this program; if not, write to the Free Software
  15.  *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  16.  */
  17.  
  18. #define suflim    256
  19. #define sufbuflim 102400
  20.  
  21. int suffixcnt;
  22. struct {
  23.     unsigned char *original;
  24.     int olen;
  25.     unsigned char *replace;
  26. }
  27. suf[suflim];
  28. char *sufbuf;
  29. int sufwordlen, sufminlen;
  30.  
  31. /* Suffix translation, to be used within translator. */
  32.  
  33. int sufcomp(int t, const unsigned char *s2)
  34. {
  35.     int k;
  36.  
  37.     for(k=0;k<suf[t].olen && k<sufwordlen
  38.       && suf[t].original[k]==s2[sufwordlen-k-1];k++);
  39.     if(k>=suf[t].olen) {
  40.       if(sufwordlen>k) return -1; else return 0;
  41.     }
  42.     else return suf[t].original[k]-s2[sufwordlen-k-1];
  43. }
  44.  
  45. /* searches a list. Returns index if found, -1 if nomatch.
  46.  * This routine is faster than naive one by one comparisons,
  47.  * and is especially suited for large lists.
  48.  */
  49. int suffix_list(void *list, int items, size_t item_size, const unsigned char *str)
  50. {
  51.     int i1,i2,j,k,t,v;
  52.     unsigned char c,d;
  53.  
  54.     if(items<=0) return -1;
  55.     k=sufcomp(0,str);
  56.     if(k==0) return 0; if(k>0) return -1;
  57.     j=items-1; k=sufcomp(j,str);
  58.     if(k==0) return j;
  59.     if(k>0) for(i1=0,i2=j;i2>i1+1;) {
  60.       j=i1+(i2-i1)/2; k=sufcomp(j,str);
  61.       if(k==0) return j;
  62.       if(k>0) {i2=j; continue;}
  63.       if(k<0) {i1=j; continue;}
  64.     }
  65.     if(k>0 && j>0) j--;
  66.     backcheck:
  67.     v=j;for(t=0;t<suf[j].olen && t<sufwordlen
  68.       && suf[j].original[t]==str[sufwordlen-t-1];t++);
  69.     if(t<sufminlen) return -1; if(t>=suf[j].olen) return j;
  70.     for(j--,c=str[sufwordlen-1],d=str[sufwordlen-t];
  71.       j>=0 && suf[j].original[0]==c && suf[j].olen>t
  72.       && suf[j].original[t-1]==d;j--);
  73.     if(j>=0 && suf[j].original[0]==c &&
  74.        strncmp((char*)suf[j].original,(char*)suf[v].original,suf[j].olen)==0)
  75.       return j;
  76.     else goto backcheck;
  77. }
  78.  
  79. /* Prepare dictionary.  */
  80. void suffix_dic(char *sdicname)
  81. {
  82.     int i,k,l;
  83.     FILE *suff;
  84.     char *p1, *p2, *pp;
  85.     long int flen;
  86.  
  87.     suffixcnt=0; sufminlen=100000;
  88.     suff=fopen(sdicname,"r"); if(suff==NULL) return;
  89.     fseek(suff,0,SEEK_END);flen=ftell(suff); fseek(suff,0,SEEK_SET);
  90.     if(flen>sufbuflim) return;
  91.     sufbuf=xmalloc(flen+16);flen=fread(sufbuf,1,flen,suff);
  92.     fclose(suff);
  93.     if(flen>0 && flen<sufbuflim) sufbuf[flen]=0;
  94.     else return;
  95.     for(i=0,p1=sufbuf;p1!=NULL && *p1!=0 && i<suflim;p1=p2) {
  96.       p2=strchr(p1+1,'\n'); if(p2>p1) *p2++=0;
  97.       pp=strchr(p1,':'); if(pp==NULL) continue;
  98.       *pp++=0;
  99.       strip_trailing_spaces(p1); strip_trailing_spaces(pp);
  100.       p1=find_word_start(p1); pp=find_word_start(pp);
  101.       if(*p1==0) continue;
  102.       if(i>0) {
  103.           k=strcmp((char*)suf[i-1].original,p1);
  104.           if(k>0) {
  105.             pp=strrchr(sdicname,'/'); if(pp==NULL) pp=sdicname; else pp++;
  106.             error("unsorted_dictionary %s: %s > %s.\n",
  107.                   pp,suf[i-1].original,p1);
  108.           }
  109.           if(k==0) {
  110.             pp=strrchr(sdicname,'/'); if(pp==NULL) pp=sdicname; else pp++;
  111.             error("duplication_in_dictionary %s: %s.\n",pp,p1);
  112.           }
  113.       }
  114.       suf[i].original=(unsigned char*)p1; suf[i].olen=l=strlen(p1);
  115.       if(l<sufminlen) sufminlen=l;
  116.       suf[i].replace=(unsigned char*)pp; i++;
  117.     }
  118.     suffixcnt=i;
  119. }
  120.  
  121. /* Suffix translation. */
  122. void suffix_translate(char *p)
  123. {
  124.     char *p1, *p2;
  125.     int t;
  126.  
  127.     for(p1=find_word_start(p);
  128.       p1!=NULL && p1-p<MAX_LINELEN && *p1!=0;
  129.       p1=p2) {
  130.        if(!isalpha(*p1)) {p2=p1+1; continue;}
  131.        for(p2=p1;isalpha(*p2);p2++);
  132.        if(*p2!=0 && strchr(" ,.?!'\"\n`:;()[]{}<>",*p2)==NULL) continue;
  133.        sufwordlen=p2-p1;
  134.        t=suffix_list(suf,suffixcnt,sizeof(suf[0]),(unsigned char*)p1);
  135.        if(t<0) continue;
  136.        string_modify(p,p2-suf[t].olen,p2,(char*)suf[t].replace);
  137.        p2=p2-suf[t].olen+strlen((char*)suf[t].replace);
  138.      }
  139.      p[MAX_LINELEN]=0;
  140. }
  141.  
  142. void suffix(char *p, char *sdicname)
  143. {
  144.     suffix_dic(sdicname); if(suffixcnt>0) suffix_translate(p);
  145. }
  146.  
  147.