use v5.14;
use Sort::Key::Top 'nkeytopsort';

sub sort_letters {
	my @letters = split '', $_[0];
	join '', sort @letters;
}

sub can_make {
	my ($letters, $word) = @_;
	$word = join '.*', sort split '', $word;
	$letters =~ qr{$word};
}

say "Reading words from dictionary...";
open my $dict, '<', '/usr/share/dict/words';
chomp( my @WORDS = map lc, grep length($_) <= 9, <$dict> );

say "Scoring microcombinations...";
my %SCORE; $SCORE{ sort_letters($_) }++ for @WORDS;
my @GOOD = nkeytopsort { $SCORE{$_} } -40 => keys %SCORE;

say "Generating candidate answers...";
my %CANDIDATES;
for my $x (@GOOD)
{
	for my $y (@GOOD)
	{
		for my $z (@GOOD)
		{
			my $letters = sort_letters substr("$x$y$z", 0, 8);
			$CANDIDATES{$letters}++;
		}
	}
}
my @VERYGOOD = nkeytopsort { $SCORE{$_} } -40 => keys %CANDIDATES;

say "Finding best candidate answer...";
my %RESULTS = map {
	my $letters  = $_;
	my @can_make = grep can_make($letters, $_), @WORDS;
	say "   $letters - ", scalar(@can_make);
	$letters => \@can_make;
} @VERYGOOD;
my ($BEST) = nkeytopsort { scalar @{$RESULTS{$_}} } -1 => @VERYGOOD;

say "BEST: $BEST";
say for sort @{$RESULTS{$BEST}};