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

print "Find pair of entities without/with separating space\n";

# read input file and param ('mod' = for modified files)
my ($infile, $param) = @ARGV;

my @regexes = $param ? (qr/–§/,
                        qr/–Ü/,
                        qr/ߧ/)
                     : (qr/&[^;]+;\s&[^;]+;/);

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

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

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

#output statistics
print {$out} "Find pair of entities without/with separating space\n\ninput file: ";
print {$out} "$infile";
print {$out} "\n========================================================================\n\n";

for my $i (0 .. $#regexes) {
    my $regex = $regexes[$i];
    $regex =~ s/^\(\?\^://;
    $regex =~ s/\)$//;
    print {$out} "$regex: $1\n" while $xml =~ /$regex/g;
 }
 
close $in;
close $out;