use strict;

my $data_suffix = q[.dat];
my $data_dir_name = q[data/];

sub normalize ($) {
  my $s = shift;
  $s =~ s/\s+/ /g;
  $s =~ s/^ //;
  $s =~ s/ $//g;
  return $s;
} # normalize

sub create_pattern1 ($) {
  my $s = quotemeta shift;
  $s =~ s/\\\*/(.+)/g;
  return $s;
} # create_pattern1

sub replace_pattern2 ($@) {
  my $s = shift;
  my @arg = @_;
  $s =~ s/\$(\d+)/$arg[$1 - 1]/g;
  return $s;
} # replace_pattern2

sub load_data_file ($$$) {
  my ($data_file_name, $exact_data, $pattern_data) = @_;

  open my $data_file, '<:utf8', $data_file_name
      or die "$0: $data_file_name: $!";
  local $/ = undef;
  my $data = <$data_file>;
  $data =~ s/\x0D?\x0A/\n/g;
  for (split /\n\n+(?=#)/, $data) {
    my ($en, $ja) = split /\n#ja\n/, $_;
    if ($en =~ s/^#en\n//) {
      $exact_data->{normalize ($en)} = $ja;
    } elsif ($en =~ s/^#pattern\n//) {
      $pattern_data->{create_pattern1 (normalize ($en))} = $ja;
    }
  }
} # load_data_file

sub for_each_data_file ($) {
  my ($code) = @_;

  opendir my $data_dir, $data_dir_name or die "$0: $data_dir_name: $!";
  for (sort {$a cmp $b} readdir $data_dir) {
    next if /^\./; # hidden files
    next if /^_/; # editable but not-used files
    my $data_file_name = qq[$data_dir_name$_];
    next unless $data_file_name =~ /\Q$data_suffix\E$/;
    $code->($data_file_name);
  }
} # for_each_data_file

1;