#!/usr/local/bin/perl
# fil 1.0, 30 Mar 1999
#
#   Purpose:
#      extract items from a test bank of LaTeX formatted questions,
#      i.e. it's an enumerated list item filter, hence the name fil
#
#   Usage example (make the file executable):
#      fil test2j
#
#   Instructions:
#    command line argument is the name of the version to produce, e.g. test2j
#    Any filename extension is ignored there, e.g. test2j.num
#
#    Currently the program assumes that the name of the test bank is in the
#    first five characters of the name provided, plus ".tex"
#    e.g. If the test bank is test2.tex and the numbers selected are in
#    test2j.num then the test produced will be test2j.tex
#
#    A file with numbers designating problems to select, one per line,
#    must have the .num extension, and before each section of numbers
#    a comment should be inserted which may include text to indicate
#    which section the following numbers are for.
#
#    Comments are indicated with a pound sign at the beginning of a line.
#    Multiple comment lines may be grouped together, but none should appear
#    after the last list of numbers.
#       Example num file for a two-section test:
#                 # Short Answer
#                 1
#                 4
#                 23
#                 # Extra Credit
#                 3
#
#    If you do not wish to choose problems from a particular section, then
#    put a zero as the only problem number in a section.  This might be
#    the case if you wanted to use Short Answer #23 as an Extra Credit problem.
#    You can have #23 chosen from the Short Answer section and then edit
#    the version that is produced and move that problem to the Extra Credit
#    section.
#
#   Warnings:
#      assumed that enumerated lists are the only types of lists in the bank
#
#      assumed that there are no enumerated lists outside of the lists of
#      problems
#
#      \clearpage, \item, \begin{enumerate}, and \end{enumerate}
#      must appear on separate lines.  If such a LaTeX element is included
#      or excluded from a test version then so is everything else on the same
#      line
#
#      don't use the Enter key on the numeric keypad to enter numbers or
#      this program may not recognize the line breaks and won't produce a
#      complete test version and may need to be interrupted
# 
#   Assurances:
#      can handle a hierarchy of enumerated item lists within problems
#
#      it is expected that \clearpage may have been used to force the printing
#      of figures near their problems in the test bank.
#      In a test version these \clearpage commands won't be needed in
#      the same places, so this script will ommit them
#
#      before each problem written to the test version is a LaTeX comment
#      that for a problem numbered 23 would look like this:
#                      % P#23
#      this is so that problems may be found quickly in post-editing, such
#      as modifying problems or eliminating problem subitems or eliminating
#      extra figures when not all are needed
#

$out = $ARGV[0];
$out =~ s/\..*//g;
$in = substr($out,0,5);
$testin = $in . '.tex';
$testout = $out . '.tex';
$numfile = $out . '.num';

open (BANK, "$testin") || die "Can't open $testin\n";
open (NUM, $numfile) || die "Can't open $numfile\n";
open (OUT, ">$testout") || die "Can't open $testout\n";;

while ($numctr = <NUM>) {
   &processItems;
}

&copyPast("\\\\end{document}");

close(NUM);
close(BANK);
close(OUT);

exit;


sub copyPast {
   # sends lines out until match
   # terminates with line in $_ following match
   $target = $_[0];
   while (($_ = <BANK>) && ! /$target/) {
      if (/\\clearpage/) {
         # do nothing with input line
      } else {
         print OUT;
      }
   }
   print OUT;
   $_ = <BANK>;
}


sub rCopyPast {
   $target = $_[0];
   while (($_ = <BANK>) &&! /$target/) {
      if (/\\begin{enumerate}/) {
         print OUT;
         &rCopyPast("\\\\end{enumerate}");
      }
      if (/\\clearpage/) {
         # do nothing with input line
      } else {
         print OUT;
      }
   }
   print OUT;
   $_ = <BANK>;
}

sub rSkipPast {
   $target = $_[0];
   while (($_ = <BANK>) &&! /$target/) {
      if (/\\begin{enumerate}/) {
         # do nothing with input line
         &rSkipPast("\\\\end{enumerate}");
      } else {
         # do nothing with input line
      }
   }
   $_ = <BANK>;
}

sub processItems {

   &copyPast("\\\\begin{enumerate}");

   while (index($numctr, "#") >= 0) {
      $numctr=<NUM>;
   }

   $itemctr = 0;
   $copystate = 0;
   $status = 1;

   while ($status) {
      SWITCH: {

         if (/\\item/) {
            $itemctr++;
            if ($itemctr == $numctr) {
               $copystate = 1;
               print OUT "% P#$numctr";
               print OUT;
               $numctr = <NUM>;
            } else {
               $copystate = 0;
               #do nothing with input line
            }
            $_ = <BANK>;
            last SWITCH;
         }

         if (/\\end{enumerate}/) {
            print OUT;
            $_ = <BANK>;
            $status = 0;
            last SWITCH;
         }

         if (/\\begin{enumerate}/) {
               if ($copystate) {
                  print OUT;
                  &rCopyPast("\\\\end{enumerate}");
               } else {
                  # do nothing with input line
                  &rSkipPast("\\\\end{enumerate}");
               }
         last SWITCH;
         }

         if (/\\clearpage/) {
            # do nothing with input line
            $_ = <BANK>;
            last SWITCH;
         }

         if ($copystate) {
            # made it to here because input line didn't match any of above,
            # so we must have an extra line that's part of an item
            print OUT;
         } else {
            #do nothing with input line
         }

         # end with an input line ready to process
         $_ = <BANK>;
      }
   }
}