#!/usr/bin/perl
use warnings;
use strict;
use diagnostics;


#my personal data left out!

print "Generate statistics: Whitespace in context\n";

my $infile = $ARGV[0];

#define regexes as search target (in the array @regexes)
my @regexes = (qr/§\s*[0-9]/, qr/Art\.\s*[0-9IVX]/, qr/Artikel\s*[0-9IVX]/, qr/Artikels\s*[0-9IVX]/, qr/Artikeln\s*[0-9IVX]/);

open my $in, '<', $infile or die "Cannot open $infile for reading: $!";

#read input file in variable $xml
my $xml;
{
  local $/ = undef;
  $xml = <$in>;
}

#define array for frequency values
my @tally;

#count routine for each regex
for my $i (0 .. $#regexes) {
	my $regex = $regexes[$i];
	++$tally[$i] while $xml =~ /$regex/g;
}	

#define output file
open my $out, '>', 'stats.txt' or die $!;


#output statistics
print {$out} "Statistics: Whitespace in context\n\ninput file: ";
print {$out} "$infile";
print {$out} "\n========================================================================\n\n";

for my $i (0 .. $#regexes) {
    my $regex = $regexes[$i];
    $regex =~ s/^\(\?\^://;
    $regex =~ s/\)$//;
    print {$out} "$regex:\t\t$tally[$i]\n";
    
 }
 
close $in;
close $out;