#!/usr/local/bin/perl
#Usage: perl dictionary.pl article.txt manywords.txt

%dict;


sub open_files {
  if ($#ARGV != 1) {
      print "Wrong number of arguments\n";
      print "Usage: perl dictionary.pl article.txt manywords.txt\n";
      exit (1); 
  }
  else {
      open (IN1,"<$ARGV[0]") or die "Cannot open $ARGV[0]: $!\n";
      open (IN2,"<$ARGV[1]") or die "Cannot open $ARGV[1]: $!\n";
  }
}


sub dictionary_to_hash {
  my ($line);
  $line=<IN2>; #read header
  while ($line=<IN2>) {
       ($w,$f)=split(/\t/,$line);
        $dict{$w}=$f;
    }
}

sub line_to_words {
   my ($line)=$_[0];
   my (@words);
   $line =~ tr/[A-Z]/[a-z]/;
   $line =~ s/[!,\.;':\"\)\(\?\-\$]/ /g;
   @words = split(/\s+/,$line);
   return @words;
}


sub insert {
   my (@words)=@_;
   my ($w);
   my ($f);
   foreach $w (@words) {
     if ($w !=~ m/\s+/){
       if (exists($dict{$w})) {
       $f=$dict{$w};
       $dict{$w}=1+$f;      
       }
       else {
        $dict{$w}=1; 
       }
     }
   }
   my (@temp)=keys(%dict); 
}


sub process_new_file {
   my ($line);
   my (@words);
   while ($line=<IN1>) {
       @words=line_to_words($line);
       insert(@words); 
     }
}



sub update_db {
   open (IN3,">$ARGV[1]") or die "Cannot open $ARGV[1]: $!\n";
   my (@sw)=sort keys(%dict);
   my ($i);
   print IN3 "Word", "\t", "Frequency\n";
   for $i (0..$#sw) {
      print IN3 "$sw[$i]", "\t", "$dict{$sw[$i]}\n";  
   }
}

sub close_files {
   close(IN1); close(IN2); close(IN3);
}


open_files; 
dictionary_to_hash;
process_new_file;
update_db;
close_files;