🚀 go-pugleaf

Somebody requested this yesterday on sci.crypt.  This code is ANSI-C,
and will run on DOS or UNIX provided you have an ANSI-C compiler
(sorry, old K&R compilers won't work).  Counts letter usage and percent,
doubled letter usage and percent, leading letter usage and percent,
trailing letter usage and percent, and two letter combinations (digraphs).

#	This is a shell archive.
#	Remove everything above and including the cut line.
#	Then run the rest of the file through sh.
#----cut here-----cut here-----cut here-----cut here----#
#!/bin/sh
# shar:    Shell Archiver
#	Run the following text with /bin/sh to create:
#	letcount.c
# This archive created: Wed Aug 03 07:47:40 1994
sed 's/^X//' << \SHAR_EOF > letcount.c
X#include <stdio.h>
X#include <stdlib.h>
X
Xtypedef unsigned int  word;
Xtypedef unsigned char byte;
Xtypedef unsigned long lword;
X
X#define APOSTROPHE 27
X#define BLANK      28
X#define DASH       29
X
X#define ALPHASIZE  32
X
Xbyte sym2alpha[] = "#ABCDEFGHIJKLMNOPQRSTUVWXYZ' -##";
X
X
Xbyte alpha2sym( byte alpha )
X{
X   if ( alpha >  0x80 ) return 0;
X
X   if ( alpha >= 'a'  ) alpha -= ' ';
X
X   if ( alpha == '-'  ) return DASH;
X
X   if ( alpha == ' '  ) return BLANK;
X   if ( alpha == '\n' ) return BLANK;
X   if ( alpha == '\t' ) return BLANK;
X   if ( alpha == '.'  ) return BLANK;
X   if ( alpha == ','  ) return BLANK;
X   if ( alpha == '!'  ) return BLANK;
X   if ( alpha == '?'  ) return BLANK;
X   if ( alpha == ';'  ) return BLANK;
X   if ( alpha == ':'  ) return BLANK;
X   if ( alpha == '"'  ) return BLANK;
X   if ( alpha == '('  ) return BLANK;
X   if ( alpha == ')'  ) return BLANK;
X
X   if ( alpha == '\'' ) return APOSTROPHE;
X
X   if ( alpha < 'A'   ) return 0;
X
X   if ( alpha > 'Z'   ) return 0;
X
X   return 1 + alpha - 'A';
X}
X
X
Xlword lettercount[ALPHASIZE];
Xlword digraphcount[ALPHASIZE][ALPHASIZE];
Xlword doublelet[ALPHASIZE];
X
X
Xint main( int nargs, char *cargs[] )
X{
X   FILE *infile, *outfile;
X   byte newblivit, oldblivit, spareblivit, anotherblivit;
X   lword totalcount=0, initcount=0, endcount=0, doublecount=0, cmpcount, count;
X
X   if ( nargs < 2 )
X   {
X      printf( "letcount infile [outfile]\n"
X              "\n"
X              "   if no oufile given, output to screen\n" );
X
X      exit(0);
X   }
X
X   infile = fopen( cargs[1], "r" );
X   if ( infile == 0 )
X   {
X      printf( "%s not found\n", cargs[1] );
X      exit(-1);
X   }
X
X   if ( nargs < 3 )
X   {
X      outfile = stdout;
X   }
X   else
X   {
X      outfile = fopen( cargs[2], "r" );
X      if ( outfile != 0 )
X      {
X         printf ( "%s already exists\n", cargs[2] );
X         exit(-2);
X      }
X
X      fclose( outfile );
X
X      outfile = fopen( cargs[2], "w" );
X      if ( outfile == 0 )
X      {
X         printf( "couldn't open %s for output\n", cargs[2] );
X      }
X   }
X
X   for ( newblivit = 0; newblivit < ALPHASIZE; newblivit++ )
X   {
X      lettercount[newblivit] = 0;
X      doublelet[newblivit] = 0;
X
X      for ( oldblivit = 0; oldblivit < ALPHASIZE; oldblivit++ )
X      {
X         digraphcount[newblivit][oldblivit] = 0;
X      }
X   }
X
X   newblivit = BLANK;
X
X   while ( 1 )
X   {
X      spareblivit = oldblivit;
X
X      oldblivit = newblivit;
X
X      newblivit = fgetc( infile );
X
X      if ( feof( infile ) ) break;
X
X      newblivit = alpha2sym( newblivit );
X
X      if ( newblivit != 0 )
X      {
X         if ( newblivit != BLANK ) totalcount++;
X         lettercount[newblivit]++;
X
X         if ( (newblivit == oldblivit) && (newblivit != BLANK) )
X         {
X            doublecount++;
X            doublelet[newblivit]++;
X         }
X
X         if ( ( oldblivit == BLANK ) && ( newblivit != BLANK ) &&
X              ( newblivit != 0 ) )
X         {
X            initcount++;
X         }
X
X         if ( ( oldblivit != BLANK ) && ( newblivit == BLANK ) &&
X              ( oldblivit != 0 ) )
X         {
X            endcount++;
X         }
X
X         if ( oldblivit != 0 )
X         {
X            digraphcount[newblivit][oldblivit]++;
X         }
X      }
X      else
X      {
X         oldblivit = spareblivit;
X      }
X   }
X
X   fprintf( outfile, "Total letter count = %7ld\n\n", totalcount );
X
X   fprintf( outfile, "Letter use frequencies:\n" );
X
X   for ( newblivit = 0; newblivit < ALPHASIZE-3; newblivit++ )
X   {
X      for ( oldblivit = 0, spareblivit = 0, cmpcount = 0;
X            oldblivit < ALPHASIZE;
X            oldblivit++ )
X      {
X         if ( lettercount[oldblivit] > cmpcount )
X         {
X            cmpcount = lettercount[oldblivit];
X            spareblivit = oldblivit;
X         }
X
X      }
X
X      lettercount[spareblivit] = 0;
X
X      if ( ( spareblivit != BLANK ) && (spareblivit <= DASH) &&
X           ( spareblivit != 0 ) && (cmpcount != 0) )
X      {
X         fprintf( outfile, "%c:  %7ld   %3ld.%1ld\%\n",
X                  sym2alpha[spareblivit],
X                  cmpcount,
X                  cmpcount*100 / totalcount,
X                  (cmpcount*1000 / totalcount) % 10 );
X      }
X   }
X
X   fprintf( outfile, "\nTotal doubled letter count = %7ld\n\n", doublecount );
X
X   fprintf( outfile, "Doubled letter frequencies:\n" );
X
X   for ( newblivit = 0; newblivit < ALPHASIZE-3; newblivit++ )
X   {
X      for ( oldblivit = 0, spareblivit = 0, cmpcount = 0;
X            oldblivit < ALPHASIZE;
X            oldblivit++ )
X      {
X         if ( doublelet[oldblivit] > cmpcount )
X         {
X            cmpcount = doublelet[oldblivit];
X            spareblivit = oldblivit;
X         }
X
X      }
X
X      doublelet[spareblivit] = 0;
X
X      if ( ( spareblivit != BLANK ) && (spareblivit <= DASH) &&
X           ( spareblivit != 0 ) && (cmpcount != 0) )
X      {
X         fprintf( outfile, "%c%c:  %7ld   %3ld.%1ld\%\n",
X                  sym2alpha[spareblivit],
X                  sym2alpha[spareblivit],
X                  cmpcount,
X                  cmpcount*100 / doublecount,
X                  (cmpcount*1000 / doublecount) % 10 );
X      }
X   }
X
X   fprintf( outfile, "\nTotal initial letters = %7ld\n", initcount );
X   fprintf( outfile, "Initial letter frequencies:\n" );
X
X   for ( newblivit = 0; newblivit < ALPHASIZE-3; newblivit++ )
X   {
X      for ( oldblivit = 0, spareblivit = 0, cmpcount = 0;
X            oldblivit < ALPHASIZE;
X            oldblivit++ )
X      {
X         if ( digraphcount[oldblivit][BLANK] > cmpcount )
X         {
X            cmpcount = digraphcount[oldblivit][BLANK];
X            spareblivit = oldblivit;
X         }
X
X      }
X
X      digraphcount[spareblivit][BLANK] = 0;
X
X      if ( ( spareblivit != BLANK ) && (spareblivit <= DASH) &&
X           ( spareblivit != 0 ) )
X      {
X         fprintf( outfile, "%c:  %7ld   %3ld.%1ld\%\n",
X                  sym2alpha[spareblivit],
X                  cmpcount,
X                  cmpcount*100 / initcount,
X                  (cmpcount*1000 / initcount) % 10 );
X      }
X   }
X
X
X
X
X   fprintf( outfile, "\nTotal ending letters = %7ld\n", endcount );
X   fprintf( outfile, "Ending letter frequencies:\n" );
X
X   for ( newblivit = 0; newblivit < ALPHASIZE-3; newblivit++ )
X   {
X      for ( oldblivit = 0, spareblivit = 0, cmpcount = 0;
X            oldblivit < ALPHASIZE;
X            oldblivit++ )
X      {
X         if ( digraphcount[BLANK][oldblivit] > cmpcount )
X         {
X            cmpcount = digraphcount[BLANK][oldblivit];
X            spareblivit = oldblivit;
X         }
X
X      }
X
X      digraphcount[BLANK][spareblivit] = 0;
X
X      if ( ( spareblivit != BLANK ) && (spareblivit <= DASH) &&
X           ( spareblivit != 0 ) && (cmpcount != 0) )
X      {
X         fprintf( outfile, "%c:  %7ld   %3ld.%1ld\%\n",
X                  sym2alpha[spareblivit],
X                  cmpcount,
X                  cmpcount*100 / initcount,
X                  (cmpcount*1000 / initcount) % 10 );
X      }
X   }
X
X   fprintf( outfile, "\nTop 250 digraphs:\n" );
X
X   for ( count = 0; count < 250; count++ )
X   {
X      for (newblivit = 0, spareblivit = 0, anotherblivit = 0, cmpcount = 0;
X           newblivit < ALPHASIZE;
X           newblivit++ )
X      {
X         for ( oldblivit = 0; oldblivit < ALPHASIZE; oldblivit++ )
X         {
X            if ( digraphcount[newblivit][oldblivit] > cmpcount )
X            {
X               cmpcount = digraphcount[newblivit][oldblivit];
X               spareblivit = newblivit;
X               anotherblivit = oldblivit;
X            }
X         }
X      }
X
X      digraphcount[spareblivit][anotherblivit] = 0;
X
X      if ( cmpcount != 0 )
X      {
X         fprintf ( outfile, "%c%c:   %ld\n",
X                   sym2alpha[anotherblivit],
X                   sym2alpha[spareblivit],
X                   cmpcount );
X      }
X   }
X
X}
SHAR_EOF
#	End of shell archive
exit 0



--
|         (V)              |  "Tiger gotta hunt.  Bird gotta fly.
|   (^    (`>              |   Man gotta sit and wonder why, why, why.
|  ((\\__/ )               |   Tiger gotta sleep.  Bird gotta land.
|  (\\<   )   der Nethahn  |   Man gotta tell himself he understand."
|    \<  )                 |  
|     ( /                  |                Kurt Vonnegut Jr.
|      |                   |  
|      ^           hahn@lds.loral.com