Subversion Repositories wimsdev

/*    Copyright (C) 1998-2003 XIAO, Gang of Universite de Nice - Sophia Antipolis

 *

 *  This program is free software; you can redistribute it and/or modify

 *  it under the terms of the GNU General Public License as published by

 *  the Free Software Foundation; either version 2 of the License, or

 *  (at your option) any later version.

 *

 *  This program is distributed in the hope that it will be useful,

 *  but WITHOUT ANY WARRANTY; without even the implied warranty of

 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

 *  GNU General Public License for more details.

 *

 *  You should have received a copy of the GNU General Public License

 *  along with this program; if not, write to the Free Software

 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

 */

#include "libwims.h"

char t_buf[4][MAX_LINELEN+1];

char maskbuf[MAX_LINELEN+1];

/* internal routine. */

void _text_cut(char *p, char *w)

{

  char *p1, *p2;

  p1=wordchr(p,w); if(p1==NULL) error("syntax_error");

  *p1=0; p2=find_word_start(p1+strlen(w));

  ovlstrcpy(t_buf[0],p); ovlstrcpy(t_buf[1],p2);

  strip_trailing_spaces(t_buf[0]);

  substitute(t_buf[0]); substitute(t_buf[1]);

}

/* Extract characters in buf[0] which are identical to

 * corresponding characters in buf[1]. */

void text_common(char *p)

{

  int i,j,n1,n2;

  _text_cut(p,"and");

  n1=strlen(t_buf[0]);n2=strlen(t_buf[1]);

  if(n2<n1) n1=n2;

  for(i=j=0;i<n1;i++) {

    if(t_buf[0][i]==t_buf[1][i] && maskbuf[i]!='0') p[j++]=t_buf[0][i];

  }

  p[j]=0;

}

/* Returns a mask string composed of '0's and '1's, where

 * '0' means corresponding positions of buf[0] and buf[1] are

 * equal.

 */

void text_compare(char *p)

{

  int min,max, i;

  _text_cut(p,"and");

  min=strlen(t_buf[0]); max=strlen(t_buf[1]);

  if(min>max) {

    i=min; min=max; max=i;

  }

  for(i=0; i<min; i++) {

    if(t_buf[0][i]==t_buf[1][i]) p[i]='0'; else p[i]='1';

  }

  for(; i<max; i++) p[i]='1';

  p[max]=0;

}

/* copy text according to mask. */

void text_copy(char *p)

{

  int i, j, n;

  snprintf(t_buf[0],MAX_LINELEN,"%s",p);

  strip_trailing_spaces(t_buf[0]); substitute(t_buf[0]);

  n=strlen(t_buf[0]);

  for(i=j=0;i<n;i++) {

    if(maskbuf[i]!='0') p[j++]=t_buf[0][i];

  }

  p[j]=0;

}

/* returns count of characters in buf[1] which appear in buf[0]. */

void text_count(char *p)

{

  int i, n, c;

  _text_cut(p,"in");

  n=strlen(t_buf[1]);

  for(i=c=0;i<n;i++) {

    if(strchr(t_buf[0],t_buf[1][i])!=NULL && maskbuf[i]!='0') c++;

  }

  snprintf(p,MAX_LINELEN,"%d",c);

}

/* Extract characters in buf[0] which are different than

 * corresponding characters in buf[1].

 */

void text_diff(char *p)

{

  int i,j,n1,n2;

  _text_cut(p,"from");

  n1=strlen(t_buf[0]);n2=strlen(t_buf[1]);

  if(n2<n1) n1=n2;

  for(i=j=0;i<n1;i++) {

    if(t_buf[0][i]!=t_buf[1][i] && maskbuf[i]!='0') p[j++]=t_buf[0][i];

  }

  p[j]=0;

}

/* put chars in buf[0] in a new string, into positions

 * corresponding to '1's in the mask buf[1].

 * Positions corresponding to '0's are filled by space.

 * Fill stops at the end of buf[0]. If buf[1] is

 * too short, it is reused from the start.

 * FIXME: using 0 as a mask gives a bug

 */

/*

   @@ !text expand abcdef mask 01

   @@ a b c d e f

   @@ !text expand abcdef mask 011

   @@ ab cd e

   @@ !text expand abcdefg using 101110

   @@ a bcd e fg

 */

void text_expand(char *p)

{

  int i,j,k,n1,n2;

  _text_cut(p,"using");

  n1=strlen(t_buf[0]);n2=strlen(t_buf[1]);

  if(n2==0) {p[0]=0; return;}

  for(i=j=k=0;i<n1 && j<MAX_LINELEN;j++,k=j%n2) {

    if(t_buf[1][k]=='0') p[j]=' ';

    else p[j]=t_buf[0][i++];

  }

  p[j]=0;

}

/* character by character replacement of buf[1] by buf[0],

 * replacing only mask-effective chars.

 * The resulting string is as long as buf[1], and the replacement

 * stops when chars buf[0] has run out.

 */

/*

  @@ !text insert abcefg into hijkl mask 10100

  @@ aibkl

 */

void text_insert(char *p)

{

  int i,j,n1,n2;

  _text_cut(p,"into");

  n1=strlen(t_buf[0]);n2=strlen(t_buf[1]);

  for(i=j=0; i<n2 && j<n1; i++) {

    if(maskbuf[i]!='0') t_buf[1][i]=t_buf[0][j++];

  }

  snprintf(p,MAX_LINELEN,"%s",t_buf[1]);

}

#define MAX_TLEN 96

/* interact of two strings according to rules

 * defined a table.

 */

void text_interact(char *p)

{

  char *table, *dline, *tline[MAX_TLEN];

  char *p1, *p2;

  int i,j1,j2,k,l,l2,n;

  table=wordchr(p,"table");

  if(table==NULL) error("syntax_error");

  *table=0; strip_trailing_spaces(p);

  table=find_word_start(table+strlen("table"));

  snprintf(t_buf[2],MAX_LINELEN,"%s",table);

  _text_cut(p,"and");

  strip_trailing_spaces(t_buf[2]); substitute(t_buf[2]);

  n=linenum(t_buf[2])-1;

  if(n>=MAX_TLEN) error("text_bad_table");

  p2=strchr(t_buf[2],'\n'); if(p2!=NULL) *p2++=0;

  if(strlen(t_buf[2])!=n) error("text_bad_table");

  dline=t_buf[2];

  for(i=0,p1=p2;i<n;i++,p1=p2) {

    if(p1==NULL) error("text_bad_table");

    p2=strchr(p1,'\n');

    if(p2!=NULL) *p2++=0;

    if(strlen(p1)!=n) error("text_bad_table");

    tline[i]=p1;

  }

  l=strlen(t_buf[0]); l2=strlen(t_buf[1]); if(l2<l) l=l2;

  for(i=k=0;i<l;i++) {

    if(maskbuf[i]!='0') {

      p1=strchr(dline,t_buf[0][i]);

      p2=strchr(dline,t_buf[1][i]);

      if(p1==NULL || p2==NULL) continue;

      j1=p1-dline; j2=p2-dline;

      if(j1>=n || j2>=n) continue; /* should not occur */

      p[k++]=tline[j1][j2];

    }

  }

  p[k]=0;

}

/* returns a mask string composed of '0's and '1's, where

 * '0' means corresponding char in buf[1] appears in buf[0].

 */

/*

   @@ !text mark a in abcaefa

   @@ 001001

 */

void text_mark(char *p)

{

  int i, n;

  _text_cut(p,"in");

  n=strlen(t_buf[1]);

  for(i=0;i<n;i++) {

    if(strchr(t_buf[0],t_buf[1][i])!=NULL) p[i]='1';

    else p[i]='0';

  }

  p[i]=0;

}

/* Returns a string whose characters are the maximum

 * of the two corresponding chars in buf[0] and buf[1].

 * Length of the string is the longuest one.

 */

void text_max(char *p)

{

  int min,max, i, j, k;

  _text_cut(p,"and");

  min=strlen(t_buf[0]); max=strlen(t_buf[1]);

  if(min>max) {

    i=min; min=max; max=i; j=0;

  }

  else j=1;

  for(i=k=0; i<min; i++) {

    if(maskbuf[i]=='0') continue;

    if((unsigned char)t_buf[0][i]>(unsigned char)t_buf[1][i])

      p[k++]=t_buf[0][i];

    else p[k++]=t_buf[1][i];

  }

  for(;i<max;i++) {

    if(maskbuf[i]!='0') p[k++]=t_buf[j][i];

  }

  p[k]=0;

}

/* Returns a string whose characters are the minimum

 * of the two corresponding chars in buf[0] and buf[1].

 * Length of the string is the shortest one.

 */

void text_min(char *p)

{

  int min,max, i,k;

  _text_cut(p,"and");

  min=strlen(t_buf[0]); max=strlen(t_buf[1]);

  if(min>max) {

    i=min; min=max; max=i;

  }

  for(i=k=0; i<min; i++) {

    if(maskbuf[i]=='0') continue;

    if((unsigned char)t_buf[0][i]< (unsigned char)t_buf[1][i])

      p[k++]=t_buf[0][i];

    else p[k++]=t_buf[1][i];

  }

  p[k]=0;

}

/* extract chars in buf[0] which occur in buf[1]. */

void text_occur(char *p)

{

  int i,j,n;

  char buf[MAX_LINELEN+1];

  memset(buf,0,sizeof(buf));

  _text_cut(p,"in");

  n=strlen(t_buf[1]);

  for(i=0;i<n;i++) {

    char *pp;

    if(maskbuf[i]=='0') continue;

    pp=strchr(t_buf[0],t_buf[1][i]);

    if(pp!=NULL) buf[pp - t_buf[0]]=1;

  }

  n=strlen(t_buf[0]);

  for(i=j=0;i<n;i++) {

    if(buf[i]) p[j++]=t_buf[0][i];

  }

  p[j]=0;

}

/* remove characters of buf[1] in buf[0]. */

void text_remove(char *p)

{

  int i, j, n;

  _text_cut(p,"in");

  n=strlen(t_buf[1]);

  for(i=j=0;i<n;i++) {

    if(strchr(t_buf[0],t_buf[1][i])==NULL

       && maskbuf[i]!='0') p[j++]=t_buf[1][i];

  }

  p[j]=0;

}

/* Cyclic reordering of text. */

void text_reorder(char *p)

{

  int i,j,k,l,n,t;

  int list[10240];

  char buf[MAX_LINELEN+1];

  _text_cut(p,"by"); *p=0;

  n=itemnum(t_buf[1]); if(n<=0 || n>=10240) return;

  for(i=0;i<n;i++) {

    buf[0]=0; fnd_item(t_buf[1],i+1,buf);

    j=atoi(buf); if(j<=0 || j>n) return;

    list[i]=j;

  }

  t=strlen(t_buf[0]);

  for(i=l=0;l<t && i<t+n;i++) {

    j=i/n; k=j*n+list[i%n];

    if(k>t || k<=0) continue;

    p[l++]=t_buf[0][k-1];

  }

  p[l]=0;

}

/* repeat a string to a given length. */

void text_repeat(char *p)

{

  int n,i,k;

  _text_cut(p,"to");

  n=strevalue(t_buf[1]); if(n>MAX_LINELEN) n=MAX_LINELEN;

  if(n<0) n=0;

  k=strlen(t_buf[0]); if(k<=0) {*p=0; return;}

  for(i=0;i<n;i++) {

    p[i]=t_buf[0][i%k];

  }

  p[i]=0;

}

/* reverse a string */

void text_reverse(char *p)

{

  int i,n;

  char buf[MAX_LINELEN+1];

  snprintf(t_buf[0],sizeof(t_buf[0]),"%s",p);

  substitute(t_buf[0]);

  n=strlen(t_buf[0]); if(n>MAX_LINELEN) n=MAX_LINELEN;

  for(i=0;i<n;i++) buf[i]=t_buf[0][n-1-i];

  buf[n]=0;

  ovlstrcpy(p,buf);

}

/* remove characters of buf[1] not in buf[0]. */

void text_select(char *p)

{

  int i, j, n;

  _text_cut(p,"in");

  n=strlen(t_buf[1]);

  for(i=j=0;i<n;i++) {

    if(strchr(t_buf[0],t_buf[1][i])!=NULL

       && maskbuf[i]!='0') p[j++]=t_buf[1][i];

  }

  p[j]=0;

}

/* tag: bit 0 is mask. */

struct {

  char *name;

  int tag;

  void (*routine) (char *p);

} text_proc[]={

  {"appear",    1,  text_occur},

  {"common",    1,  text_common},

  {"compare",   0,  text_compare},

  {"copy",      1,  text_copy},

  {"count",     1,  text_count},

  {"delete",    1,  text_remove},

  {"diff",      1,  text_diff},

  {"differ",    1,  text_diff},

  {"drop",      1,  text_remove},

  {"expand",    0,  text_expand},

  {"extract",   1,  text_select},

  {"insert",    1,  text_insert},

  {"interact",  1,  text_interact},

  {"mark",      0,  text_mark},

  {"max",       1,  text_max},

  {"min",       1,  text_min},

  {"occur",     1,  text_occur},

  {"occurrence",1,  text_occur},

  {"pick",      1,  text_select},

  {"pickup",    1,  text_select},

  {"remove",    1,  text_remove},

  {"reorder",   0,  text_reorder},

  {"repeat",    0,  text_repeat},

  {"reverse",   0,  text_reverse},

  {"select",    1,  text_select}

};

#define TEXT_PROC_NO (sizeof(text_proc)/sizeof(text_proc[0]))

int textab_verify(void)

{

  return verify_order(text_proc,TEXT_PROC_NO,sizeof(text_proc[0]));

}

/* main entry point for text routines */

void text(char *p)

{

  int i,j,n;

  char *p1, *p2;

  char c,cc;

  char buf[MAX_LINELEN+1];

  p1=find_word_start(p); p2=find_word_end(p1);

  if(p2<=p1 || *p2==0) error("syntax_error");

  *p2=0;

  i=search_list(text_proc,TEXT_PROC_NO,sizeof(text_proc[0]),p1);

  if(i<0) error("syntax_error");

  snprintf(buf,sizeof(buf),"%s",find_word_start(p2+1));

  if((text_proc[i].tag&1)!=0 && (p1=wordchr(buf,"mask"))!=NULL) {

    *p1=0; strip_trailing_spaces(buf);

    p2=find_word_start(p1+strlen("mask"));

    strip_trailing_spaces(p2);

    snprintf(maskbuf,sizeof(maskbuf),"%s",p2);

    substitute(maskbuf);

    n=strlen(maskbuf); if(n==0) goto zeromask;

    c=maskbuf[n-1]; cc=0;

    if(c=='+') cc='1';

    if(c=='-') cc='0';

    if(cc!=0) memset(maskbuf+n-1,cc,sizeof(maskbuf)-n);

    else for(j=n;j<MAX_LINELEN;j++) maskbuf[j]=maskbuf[j%n];

    maskbuf[sizeof(maskbuf)-1]=0;

  }

  else zeromask: memset(maskbuf,0,sizeof(maskbuf));

  text_proc[i].routine(buf);

  buf[MAX_LINELEN]=0;ovlstrcpy(p,buf);

}