#!/bin/bash
#
#  Look up words using a string of letters and wildcard characters
#
#  14Sep2004: Incorporate a new class of wildcard characters to facilitate
#   searching for repeat letter patterns in matched words. These characters,
#   '+', '@', and '=' each match a single letter, but with the added wrinkle
#   that each occurance of '+' (for example) in the search string matches
#   _the same_ character. For example, the search string "+@==@+" finds
#   "pullup", "redder" and several less common words, and ".+@==@+" finds
#   "dresser", "grammar", and a few others.
#
PG=wlu
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 "$PG: Find words using a string of letters and these wildcards:"
   echo "$PG:    '.' matches a single letter"
   echo "$PG:    '@' like '.', but all occurances of '@' match the same letter"
   echo "$PG:    '=' like '.', but all occurances of '=' match the same letter"
   echo "$PG:    '+' like '.', but all occurances of '+' match the same letter"
   echo "$PG:    '-' matches 0 or more successive letters"  
   echo "$PG:    '1' matches 0 or 1 letters"
   echo "$PG:    '2' matches 0, 1, or 2 letters, ... and so forth for 3,4, ...9"
   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: Examples: '-cats-', 'f.r.o.g', '2ee2', 'q...', '.+@==@+'"
}
if [ -z "$1" ] ; then
   ONESHOT="0"
   display_help
else
   ONESHOT="1"
fi
while [ -n "$ONESHOT" ] ; do
  if [ "$ONESHOT" == "0" ] ; then
     read -e -p "wlu>" WORD1 WORD2
     if [ -z "$WORD1" ] ; then
        exit 0
     fi
  elif [ "$ONESHOT" == "1" ] ; then
     WORD1="$1"
     WORD2="$2"
     unset ONESHOT
  fi
  if [ -n "$WORD2" ] ; then
     display_help
  else
# Convert user's argument into a regular expression for 'grep'. The script
# uses \xHH escape sequences for characters that the shell messes with,
# namely '\', '<', '>', '{', and '}'. 
TARGET=` gawk -v arg=$WORD1 ' \
BEGIN {
   # The next few statements set up for the wildcard characters that
   #  stand for unique but unspecified letters. First we specify the
   #  number of such characters, then we initialize an array that we
   #  use to remember when we have encountered the first occurances
   #  of each of the wildcards in this class. The variable 'sbx_cnt'
   #  holds the running count of parenthesized sub-expressions in the
   #  'grep' argument that we are building, each sub-expressions being
   #  used for the _first_ occurance of a particular such wildcard.
   ucw_cnt = 3;
   ucw_idx = 11;
   sbx_cnt = 0;
   for (i=0;i<ucw_cnt;i++) sbx_num[i] = 0;
   s = "\x5C\x3C";
   m = length(arg);
   dot_idx = ucw_idx + ucw_cnt;
   for (i=1;i<=m;i++) {
      c = tolower(substr(arg,i,1));
      n = index("123456789-@=+.abcdefghijklmnopqrstuvwxyz",c);
      if      (n  > dot_idx) s = s c;        # 'a' through 'z'
      else if (n == dot_idx) s = s "[a-z]";  # '.'
      else if (n >= ucw_idx) {
         # find multiple occurances of _the same_ letter
         j = n - ucw_idx;          # 0, 1, ...
         if (sbx_num[j] == 0) {
            # this is the 1st occurance of this wildcard character
            sbx_cnt++;
            s = s "\x5C([a-z]\x5C)";
            sbx_num[j] = sbx_cnt;
         }
         else {
            # this is the 2nd, 3rd, ... occurance of this wildcard character
            s = s "\x5C" sbx_num[j];
         }
      }
      else if (n == 10) s = s "[a-z]*";                         # '-'
      else if (n  >  0) s = s "[a-z]\x5C\x7B" "0," c "\x5C\x7D" # '1' - '9'
      else             exit 1;
   }
   print s "\x5C\x3E";
   exit 0;
} ' `
if  [ $? -ne 0 ] ; then
   display_help
else
#  echo "${TARGET}"
   grep "${TARGET}" $WL >$TMPFILE1
   RESULT=$?
   if   [ $RESULT -gt 1 ] ; then
      echo "$PG: CAUTION: an error was encountered running 'grep'."
   elif [ $RESULT -eq 1 ] ; then
      echo "$PG: No matches were 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
fi
done