src/contrib/hrefify.c

   1 /*
   2  *
   3  *                This source code is part of
   4  *
   5  *                 G   R   O   M   A   C   S
   6  *
   7  *          GROningen MAchine for Chemical Simulations
   8  *
   9  *                        VERSION 3.3.99_development_20071104
  10  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  11  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  12  * Copyright (c) 2001-2006, The GROMACS development team,
  13  * check out http://www.gromacs.org for more information.
  14
  15  * This program is free software; you can redistribute it and/or
  16  * modify it under the terms of the GNU General Public License
  17  * as published by the Free Software Foundation; either version 2
  18  * of the License, or (at your option) any later version.
  19  *
  20  * If you want to redistribute modifications, please consider that
  21  * scientific software is very special. Version control is crucial -
  22  * bugs must be traceable. We will be happy to consider code for
  23  * inclusion in the official distribution, but derived work must not
  24  * be called official GROMACS. Details are found in the README & COPYING
  25  * files - if they are missing, get the official version at www.gromacs.org.
  26  *
  27  * To help us fund GROMACS development, we humbly ask that you cite
  28  * the papers on the package - you can find them in the top README file.
  29  *
  30  * For more info, check our website at http://www.gromacs.org
  31  *
  32  * And Hey:
  33  * Groningen Machine for Chemical Simulation
  34  */
  35 #ifdef HAVE_CONFIG_H
  36 #include <config.h>
  37 #endif
  38
  39 #include <math.h>
  40 #include <string.h>
  41 #include <ctype.h>
  42 #include "gromacs/utility/smalloc.h"
  43 #include "typedefs.h"
  44 #include "macros.h"
  45 #include "gromacs/utility/cstringutil.h"
  46 #include "gromacs/fileio/confio.h"
  47 #include "gromacs/math/vec.h"
  48 #include "gromacs/commandline/pargs.h"
  49 #include "copyrite.h"
  50 #include "gromacs/fileio/pdbio.h"
  51 #include "gromacs/utility/strdb.h"
  52
  53 gmx_bool isword(char c)
  54 {
  55   return (isalnum(c) || (c=='-') || (c=='_'));
  56 }
  57
  58 char *strncasestr(char *line,char *str)
  59 {
  60   char *dum;
  61
  62   dum=line;
  63   if (dum) {
  64     while (strlen(dum) && strncasecmp(dum,str,strlen(str)))
  65       dum++;
  66     if (strlen(dum)==0)
  67       dum=NULL;
  68   }
  69
  70   return dum;
  71 }
  72
  73 char *strstr_href(char *line,gmx_bool *bInHREF,int *i_dat,int n_dat,char **dat)
  74 {
  75   char *start,*found,*href=NULL;
  76   gmx_bool bIn;
  77   int i;
  78
  79   found=NULL;
  80   *i_dat=-1;
  81   bIn=*bInHREF;
  82   start=line;
  83   do {
  84     if (bIn) {
  85       while (strlen(start) && (strncasecmp(start,"</a",3) != 0))
  86         start++;
  87       if (strlen(start)>0) {
  88         start+=3;
  89         bIn=FALSE;
  90       }
  91     }
  92     else {
  93       href=strncasestr(start,"<a href");
  94       if (href)
  95         bIn=TRUE;
  96       i=0;
  97       while((i<n_dat) && !found) {
  98         found=strncasestr(start,dat[i]);
  99         if (found) {
 100           if (href && (found>href))
 101             found=NULL;
 102           else {
 103             if (((found!=start) && isword(found[-1])) ||
 104                 isword(found[strlen(dat[i])]))
 105               found=NULL;
 106             else
 107               *i_dat=i;
 108           }
 109           i++;
 110         }
 111       }
 112     }
 113   } while (strlen(start) && !found && href);
 114   *bInHREF=bIn;
 115
 116   return found;
 117 }
 118
 119 int main(int argc, char *argv[])
 120 {
 121   static char *desc[] = {
 122     "[TT]hrefify[tt] adds href's for all the words in the input file which are not",
 123     "already hyperlinked and which appear in the file specified with the",
 124     "option [TT]-l[tt].[PAR]",
 125     "If the href's should call a script, text can be added",
 126     "with the [TT]-t[tt] option."
 127   };
 128
 129   int n;
 130
 131   char **text,**str,line[1024],*ptr,*ref,
 132     start[STRLEN],word[STRLEN],end[STRLEN];
 133   int n_text,n_str,i_str;
 134   gmx_bool bInHREF,bIn;
 135
 136   FILE    *fp;
 137   char    title[STRLEN];
 138   int     i,l,n_repl;
 139   t_filenm fnm[] = {
 140     { efDAT, "-l", "links", ffLIBRD },
 141   };
 142 #define NFILE asize(fnm)
 143   static char *in=NULL,*out=NULL,*excl=NULL,*link_text=NULL;
 144   static gmx_bool peratom=FALSE;
 145   t_pargs pa[] = {
 146     { "-f", FALSE, etSTR, { &in } , "HTML input" },
 147     { "-o", FALSE, etSTR, { &out } , "HTML output" },
 148     { "-e", FALSE, etSTR, { &excl } , "Exclude a string from HREF's, "
 149       "when this option is not set, the filename without path and extension "
 150       "will be excluded from HREF's"},
 151     { "-t", FALSE, etSTR, { &link_text } , "Insert a string in front of the "
 152       "href file name, useful for scripts" }
 153   };
 154
 155   CopyRight(stderr,argv[0]);
 156   parse_common_args(&argc,argv,0,NFILE,fnm,asize(pa),pa,
 157                     asize(desc),desc,0,NULL);
 158
 159   if (!in || !out)
 160     gmx_fatal(FARGS,"Input or output filename is not set");
 161
 162   n_text = get_file(in, &text);
 163   fprintf(stderr,"Read %d lines from %s\n",n_text,in);
 164
 165   n_str=get_file(ftp2fn(efDAT,NFILE,fnm),&str);
 166   fprintf(stderr,"Read %d strings %s\n",n_str,ftp2fn(efDAT,NFILE,fnm));
 167   if (!excl) {
 168     for (i=strlen(in)-1; i>0 && in[i-1]!='/'; i--);
 169     excl=strdup(in+i);
 170     for(i=strlen(excl)-1; i>0 && (excl[i]!='.'); i--);
 171     if (excl[i]=='.')
 172       excl[i]='\0';
 173   }
 174   fprintf(stderr,"Excluding '%s' from references\n",excl);
 175   for(l=0; l<n_str && strcasecmp(str[l],excl); l++);
 176   if (l<n_str) {
 177     for(i=l+1; i<n_str; i++)
 178       str[i-1]=str[i];
 179     n_str--;
 180   }
 181
 182   if (!link_text)
 183     link_text=strdup("\0");
 184   else
 185     fprintf(stderr,"Adding '%s' to href's\n",link_text);
 186
 187   fp=gmx_ffopen(out,"w");
 188
 189   n_repl=0;
 190   i_str=-1;
 191   bInHREF=FALSE;
 192   for(l=0; l<n_text; l++) {
 193     strcpy(line,text[l]);
 194     do {
 195       bIn=bInHREF;
 196       ptr=strstr_href(line,&bIn,&i_str,n_str,str);
 197       if (ptr) {
 198         ref=ptr;
 199         if ((ref!=line) && (ref[-1]=='.')) {
 200           ref--;
 201           while((ref>line) && isword(ref[-1]))
 202             ref--;
 203         }
 204         strcpy(start,line);
 205         start[ref-line]='\0';
 206         strcpy(word,ref);
 207         word[ptr-ref+strlen(str[i_str])]='\0';
 208         strcpy(end,ptr+strlen(str[i_str]));
 209         sprintf(line,"%s<a href=\"%s%s.html\">%s</a>%s",
 210                 start,link_text,str[i_str],word,end);
 211         fprintf(stderr,"line %d: %s\n",l+1,str[i_str]);
 212         n_repl++;
 213       }
 214     } while (ptr);
 215     bInHREF=bIn;
 216     fprintf(fp,"%s\n",line);
 217   }
 218
 219   gmx_ffclose(fp);
 220
 221   fprintf(stderr,"Added %d HTML references\n",n_repl);
 222
 223   return 0;
 224 }
 225