#!/bin/bash
#
# Look up words with a given pattern of repeating characters
#
PG=plu
WL=/usr/dict/word.list
LIM1=4   # If this many or fewer matches, list eash on it's own line.
LIM2=79  # Maximum number of characters on a line. 
LIM3=23  # If more than this many lines, output to 'less'.
TMPFILE1=/tmp/${PG}-tmp1_.tmp
TMPFILE2=/tmp/${PG}-tmp2_.tmp
function display_help()
{
   echo "$PG: Find words that match the number of letters and the repeating letter"
   echo "$PG:   pattern in the user-supplied argument; this is useful in finding"
   echo "$PG:   individual candidate words in the solution a simple substitution"
   echo "$PG:   crypogram."
   echo "$PG: Invoke with no argument to begin and exit interactive mode."
   echo "$PG: Long lists of matches are displayed using 'less'; 'q' to exit."
   echo "$PG: Example: 'zyxyz' matches  'kaiak', 'rotor', and several others."
}
if [ -z "$1" ] ; then
   ONESHOT="0"
   display_help
else
   ONESHOT="1"
fi
while [ -n "$ONESHOT" ] ; do
   if [ "$ONESHOT" == "0" ] ; then
      read -e -p "$PG>" WORD1 WORD2
      if [ -z "$WORD1" ] ; then
         exit 0
      fi
   elif [ "$ONESHOT" == "1" ] ; then
      WORD1="$1"
      WORD2="$2"
      unset ONESHOT
   fi
   if [ -n "$WORD2" ] ; then
      display_help
   else
   # Check the user's argument and output a list of pattern matches.
      gawk -v arg=$WORD1 ' \
      BEGIN {
         ok = 0;
         arg_len = length(arg);
         if (arg_len < 2) exit 2;
         s = "";
         for (i=1;i<=arg_len;i++) {
            c = tolower(substr(arg,i,1));
            n = index("abcdefghijklmnopqrstuvwxyz",c);
            if (n > 0) s = s c;
            else       exit 2;
         }
         ok = 1;
         num = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";
         num_cnt = length(num);
         if (arg_len > num_cnt) exit 3;
         # Here we generate the pattern of our argument. The result is in the 
         # form of a character string with one character per character of the
         # argument; the pattern character for the first argument character is
         # always '0'; for other characters in the argument,
         #     - the pattern character is '0' if the argument character is not
         #       the same  as any other character BEFORE it in the argument;
         #     - otherwise the pattern character is a character from 'num' that
         #       represents an index (1,2, ) to the first character in the argument
         #       that it matches.
         #
         p0 = "0";
         for (i=2;i<=arg_len;i++) {
            p0 = p0 substr(num,index(substr(arg,1,i-1),substr(arg,i,1))+1,1);
         }
         cnt = 0;
      }
      {
         if (length($1) != arg_len) next;
         p1 = "0";
         for (i=2;i<=arg_len;i++) {
            p1 = p1 substr(num,index(substr($1,1,i-1),substr($1,i,1))+1,1);
         }
         if (p1 != p0) next;
         print $1
         cnt++;
      }
      END {
        if (ok != 1) exit 2;
        if (cnt < 1) exit 1;
        exit 0;
      } ' $WL >$TMPFILE1
      RTN=$?
      if  [ $RTN -eq 2 ] ; then
         display_help
      elif [ $RTN -eq 3 ] ; then
         echo "$PG: Too many characters in pattern."
      elif [ $RTN -eq 1 ] ; then
         echo "No pattern matches found."
      else
         # Count matches and possibly consolidate onto fewer lines
         gawk -v lim1=$LIM1 -v lim2=$LIM2 -v lim3=$LIM3 ' \
         {
            if (FNR == 1) {
               filecnt++;
               if (filecnt == 2) {
                  # Determine words per line for the output pass
                  reccnt = recnumb;
                  if (reccnt <= lim1) {
                     words_per_line = 1;
                  } 
                  else {
                     words_per_line = 1 + int((lim2-maxlength)/(maxlength+1));
                     if (words_per_line < 1) words_per_line = 1;
                  }
                  # Calculate the format string
                  fmt = "%-" maxlength "s";
               }
               words_so_far = 0;
               lines_out = 0;
               recnumb = 0;
            }
            recnumb++;
            if (filecnt == 1) {
               n = length($1);
               if (n > maxlength) maxlength = n;
            } else {
               if (words_so_far == words_per_line) {
                  printf "\n";
                  words_so_far = 0;
               }
               if (words_so_far == 0) {
                  printf fmt,$1;
                  lines_out++;
               }
               else printf " " fmt,$1
               words_so_far++;
            }
         }
         END {
            if (words_so_far > 0) {
               printf "\n";
               lines_out++;
            }
            if (lines_out > lim3) exit 1;
            else                  exit 0;
         }; ' $TMPFILE1 $TMPFILE1 >$TMPFILE2
         if [ $? -eq 0 ] ; then
            cat $TMPFILE2
         else
            cat $TMPFILE2 | less
         fi
      fi
   fi
done