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

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

my $infile = $ARGV[0];

# put in comments what is not relevant
# for modified files: currently known cases
# my @regexes = (qr/–§/, qr/–Ü/, qr/ߧ/);
# for original files: check for entities separated by space
my @regexes = (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 $!;

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\n" while $xml =~ /$regex/g;;   
 }
 
close $in;
close $out;