#!/usr/bin/perl 
### NOTE: Fixed the VOLATILE spot in line 17 to stabilize from one semester to another
### First we do routine parse of the  arguments passed with the classparse through the command line ###
if   ($#ARGV == 0) { $input = $ARGV[0]; print STDERR "Extracting records from ", $input, "\n"}
else {print STDERR "Need one  argument.\n", "Usage:\n  classparse input_html_file\n  
                    (TAB delimited output goes to the standard output device, typically the screen)\n";
             exit 0; };                                       # little help 
### End of command line parse and the real stuff begins
open(DATAFILE,"$input") || die("cannot open the html data file: $!");     # opens the input html file for reading 
$name = "";                        # initialize the name to empty (meaning no name at hand initially)
$counter = 0;                    # initialize the counter of records extracted
while (<DATAFILE>)               # big loop over the lines of the file      
{  # begin{big loop}
  chomp;                         # strip  the new line character \n of the end of the line (just cosmetics)
  if ($name eq "")                 # test if we have no name at hand (generally found in the previous line unles this one is first) 
    {  # begin{no name on previous line}           we did not find name on the previous line so we look for it now:
#     @name = split /<TD.+430">/;  # split at  substrings of the form:  <TD anything 430"> and store in an array VOLATILE
 @name = split /<TD.+term.[0-9]+">/;  # split at  substrings of the form:  <TD anyword term anycharacter somedigits"> and store in an array (MOD in Jan 2004)
     if ($#name == 1)  # if the max index  of non-zero elements of the array @name (indices start with zero) is one 
       { # begin{name found}
 # TEST   print "name tokens: "; foreach $token (@name) { print ($token, "*\n"); }; print "\n\n";
         $name = $name[1];    # Last, First, Middle Names with some html trailing it. (Note $name[0] is typically empty after the above split.) 
         $name =~ s/<.+>//;     # strips <whatever>,  which is the form of the  html trailing the name
 # TEST  print("Got name =", $name,"\n"); 
     }  #  end{name found}
    next ;                                # name found so break off the loop and go to the next line to look for ssn 
  }   # end{no name on previous line}
if ($name ne "")                 # test if we have name at hand found in the previous parsed line
  {   # begin{name number found on previous line}
       @ssn = split /<TD.+fieldmediumtext>-/;  # @ssn is typically an empty string followed by ssn trailed by some html junk
 #TEST print "ssn tokens: "; foreach $token (@ssn) { print ($token, "*\n"); }; print "\n\n";     # print tokens separated by *
      if ($#ssn > 1) { print STDERR "possible ssn parsing error: too many tokens \n";};    # if more than two tokens (first empty)
      if ($#ssn < 1) { print STDERR "parsing warning: no ssn matches Name =", $name, " (do not worry)\n";};  # if no name after ssn then this is not ssn so just issue a warning 
      if ($#ssn == 1)                                                                # if two tokens then we got the ssn  
      {   # begin{found ssn} 
          $ssn = $ssn[1];   
          $ssn =~ s/<.+>//;       # substitutes for "< anything >" emptyspace 

 ###############  OUTPUT PANEL BELOW (customize if you need comma delimited etc...) ###################

 #### Uncomment the one line below that does what you want: 
   #### screen output:	
         print ($ssn, "\t", $name, "\n");           # print to screen name followed by TAB followed by ssn followed by NEW LINE 
         # print ($name, "\t", $ssn, "\n");    # print  name followed by TAB followed by ssn followed by NEW LINE 
         # print ($name, ",", $ssn, "\n");     # print   ssn followed by TAB followed by name followed by NEW LINE 
         # print ($name, ",", $ssn, "\n");     # print    ssn followed by COMMA followed by name followed by NEW LINE 

####################### END OUTPUT PANEL #######################################
          ++$counter
       }  # end{found name} 
       $name = "";                # reset ssn #MOD 3.1 so that it does not hold $ssn for more than one line 
       }   #end{name found on the previous line parsed}
}  # end{big loop}
print STDERR "Done!  Extracted ", $counter, " records\n ";