#!/usr/bin/perl
# Copyright (c) 2012-2018 Steve McIntyre <93sam@debian.org>
# This code is hereby licensed for public consumption under either the
# GNU GPL v2 or greater, or Larry Wall's Artistic license - your choice.
#
# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
#
# abcde-musicbrainz-tool
#
# Helper script for abcde to work with the MusicBrainz WS API (v2)

use strict;
use utf8;
use POSIX qw(ceil);
use Digest::SHA;
use MusicBrainz::DiscID;
use WebService::MusicBrainz 1.0.4;
use Getopt::Long;
use Pod::Usage;

my $FRAMES_PER_S = 75;

my ($device, $command, $discid, @discinfo, $workdir, $help, $man, $start);
Getopt::Long::Configure ('no_ignore_case');
Getopt::Long::Configure ('no_auto_abbrev');
GetOptions ("device=s"       => \$device,
            "command=s"      => \$command,
            "discid=s"       => \$discid,
            "discinfo=i{5,}" => \@discinfo,
            "workdir=s"      => \$workdir,
            "start=s"        => \$start,
            "help|h"         => \$help,
            "man"            => \$man) or pod2usage(-verbose => 0, -exitcode => 2);
if (@ARGV) {
    print STDERR "Extraneous arguments given.\n";
    pod2usage(-verbose => 0, -exitcode => 2);
}
pod2usage(-verbose => 1, -exitcode => 0) if $help;
pod2usage(-verbose => 2, -exitcode => 0) if $man;

# defaults
if (!defined($device)) {
    $device = "/dev/cdrom";
}
if (!defined($command)) {
    $command = "id";
}
if (!defined($workdir)) {
    $workdir = "/tmp";
}
if (!defined($start)) {
    $start = "0";
}

sub calc_sha1($) {
    my $filename = shift;
    my $s = Digest::SHA->new(1);
    $s->addfile($filename);
    return $s->hexdigest;
}

if ($command =~ m/^id/) {
    my $disc = new MusicBrainz::DiscID($device);

    # read the disc in the default disc drive */
    if ( $disc->read() == 0 ) {
        printf STDERR "Error: %s\n", $disc->error_msg();
        exit(1);
    }

    printf("%s ", $disc->id());
    printf("%d ", $disc->last_track_num() + 1 - $disc->first_track_num());

    for ( my $i = $disc->first_track_num;
          $i <= $disc->last_track_num; $i++ ) {
        printf("%d ", $disc->track_offset($i));
    }
    printf("%d\n", $disc->sectors() / $FRAMES_PER_S);
    undef $disc;
} elsif ($command =~ m/data/) {
    if (!defined $discid or !$discid) {
	print STDERR "Discid undefined.\n";
	exit(1);
    }
    my $ws = WebService::MusicBrainz->new();
    my $response = $ws->search(discid => {
	discid => $discid,
	inc => ['artists', 'artist-credits', 'recordings']
        });

    if ($response->{'error'}) {
	print STDERR "MusicBrainz lookup returned an error \"$response->{'error'}\"\n";
	exit(0);
    }

    my $releasenum = $start;
    my @sums;

    if ($response->{'releases'}) {
	my @releases = @{ $response->{'releases'} };
	foreach my $release (@releases) {
	    my $a_artist = "";
	    my $number_artists = @{ $release->{'artist-credit'}};
	    if ($number_artists > 0) {
		for (my $i = 0; $i < $number_artists; $i++) {
		    if ($i > 0) {
			$a_artist = $a_artist . @{ $release->{'artist-credit'} }[$i-1]->{'joinphrase'};
		    }
		    $a_artist = $a_artist . @{ $release->{'artist-credit'} }[$i]->{'name'};
		}
	    }
	    my $va = 0;
	    my $rel_year = "";
	    if ($a_artist =~ /Various Artists/) {
		$va = 1;
	    }

	    if ($release->{'release-events'}) {
		my @release_events = @{ $release->{'release-events'} };
		if (@release_events > 0) {
		    $rel_year =  substr(@release_events[0]->{'date'},0,4);
		}
	    }
	    $releasenum++;
	    open (OUT, "> $workdir/cddbread.$releasenum");
	    binmode OUT, ":utf8";
	    print OUT "# xmcd style database file\n";
	    print OUT "#\n";
	    print OUT "# Track frame offsets:\n";

	    my @offsets = @{ $response->{'offsets'}};
	    foreach my $offset (@offsets) {
		printf OUT "#       %d\n", $offset;
	    }

	    # Locate the media that contains a disc with the discid we requested
	    # initially. The API may return multiple media associated with the
	    # release, including media with different discids
	    my @mediums = grep {
		my @disks = @{ $_->{'discs'} };
		grep { $_->{'id'} eq $discid } @disks;
	    } @{ $release->{'media'} };

	    if (not @mediums) {
		# This release doesn't have a media with our requested dicsid
		# Shouldn't happen (?), skip it
		next;
	    }

	    # Only consider the first medium
	    my $medium = @mediums[0];
	    my @tracks = @{ $medium->{'tracks'} };

	    my $total_len = 0;
	    for (my $i = 0; $i < scalar(@tracks); $i++) {
		my $track = $tracks[$i];
		$total_len += $track->{'length'};
	    }

	    print OUT "#\n";
	    printf OUT "# Disc length: %d seconds\n", $total_len / 1000.0;
	    print OUT "#\n";
	    print OUT "# Submitted via: XXXXXX\n";
	    print OUT "#\n";
	    print OUT "#blues,classical,country,data,folk,jazz,newage,reggae,rock,soundtrack,misc\n";
	    print OUT "#CATEGORY=none\n";
	    print OUT "DISCID=" . $discid . "\n";
	    print OUT "DTITLE=" . $a_artist. " / " . $release->{'title'} . "\n";
	    print OUT "DYEAR=" . $rel_year . "\n";
	    print OUT "DGENRE=\n";        

	    for (my $i = 0; $i < scalar(@tracks); $i++) {
		my $track = $tracks[$i];
		my $t_name = $track->{'title'};
		my $number_artists = @{$track->{'recording'}->{'artist-credit'}};
		if ($va and $number_artists > 0) {
		    my $t_artist = "";
		    for (my $j = 0; $j < $number_artists; $j++) {
			if ($j > 0) {
			    $t_artist = $t_artist . @{$track->{'recording'}->{'artist-credit'}}[$j-1]->{'joinphrase'};
			}
			$t_artist = $t_artist . @{$track->{'recording'}->{'artist-credit'}}[$j]->{'name'};
		    }
		    printf OUT "TTITLE%d=%s / %s\n", $i, $t_artist, $t_name;
		} else {
		    printf OUT "TTITLE%d=%s\n", $i, $t_name;
		}
	    }

	    print OUT "EXTD=\n";
	    for (my $i = 0; $i < scalar(@tracks); $i++) {
		printf OUT "EXTT%d=\n", $i;
	    }
	    print OUT "PLAYORDER=\n";
	    print OUT ".\n";
	    close OUT;

	    # save release mbid
	    open (OUT, "> $workdir/mbid.$releasenum");
	    print OUT $release->{'id'};
	    close OUT;

	    # save release asin
	    open (OUT, "> $workdir/asin.$releasenum");
	    print OUT $release->{'asin'};
	    close OUT;

	    # Check to see that this entry is unique; generate a checksum
	    # and compare to any previous checksums
	    my $checksum = calc_sha1("$workdir/cddbread.$releasenum");
	    foreach my $sum (@sums) {
		if ($checksum eq $sum) {
		    unlink("$workdir/cddbread.$releasenum");
		    $releasenum--;
		    last;
		}
	    }
	    push (@sums, $checksum);
	}
    } else {
	# No release events found - looks like we have a stub
	# entry. We can parse it, but it's going to be very
	# different.
	print STDERR "MusicBrainz lookup only returned a stub, trying to cope\n";
	
	my $va = 0;
	my $a_artist = "";
	my $a_title = "";
	my $rel_year = "";

	if ($response->{'artist'}) {
	    $a_artist = $response->{'artist'};
	}
	if ($response->{'title'}) {
	    $a_title = $response->{'title'};
	}
	if ($a_artist =~ /Various Artists/) {
	    $va = 1;
	}

	$releasenum++;
	open (OUT, "> $workdir/cddbread.$releasenum");
	binmode OUT, ":utf8";
	print OUT "# xmcd style database file\n";
	print OUT "#\n";
	print OUT "# Musicbrainz stub entry - check carefully!\n";
	print OUT "#\n";

	my @tracks = @{ $response->{'tracks'} };

	my $total_len = 0;
	for (my $i = 0; $i < scalar(@tracks); $i++) {
	    my $track = $tracks[$i];
	    $total_len += $track->{'length'};
	}

	printf OUT "# Disc length: %d seconds\n", $total_len / 1000.0;
	print OUT "#\n";
	print OUT "# Submitted via: XXXXXX\n";
	print OUT "#\n";
	print OUT "#blues,classical,country,data,folk,jazz,newage,reggae,rock,soundtrack,misc\n";
	print OUT "#CATEGORY=none\n";
	print OUT "DISCID=" . $discid . "\n";
	print OUT "DTITLE=" . $a_artist. " / " . $a_title . "\n";
	print OUT "DYEAR=" . $rel_year . "\n";
	print OUT "DGENRE=\n";        

	for (my $i = 0; $i < scalar(@tracks); $i++) {
	    my $track = $tracks[$i];
	    my $t_name = $track->{'title'};
	    if ($va) {
		my $t_artist = $track->{'artist'};
		printf OUT "TTITLE%d=%s / %s\n", $i, $t_artist, $t_name;
	    } else {
		printf OUT "TTITLE%d=%s\n", $i, $t_name;
	    }
	}

	print OUT "EXTD=\n";
	for (my $i = 0; $i < scalar(@tracks); $i++) {
	    printf OUT "EXTT%d=\n", $i;
	}
	print OUT "PLAYORDER=\n";
	print OUT ".\n";
	close OUT;

	# Check to see that this entry is unique; generate a checksum
	# and compare to any previous checksums
	my $checksum = calc_sha1("$workdir/cddbread.$releasenum");
	foreach my $sum (@sums) {
	    if ($checksum eq $sum) {
		unlink("$workdir/cddbread.$releasenum");
		$releasenum--;
		last;
	    }
	}
	push (@sums, $checksum);
    }
} elsif ($command =~ m/calcid/) {
    # Calculate MusicBrainz ID from disc offsets; see
    # https://musicbrainz.org/doc/DiscIDCalculation

    if ($#discinfo < 5) {
	print STDERR "Insufficient or missing discinfo data.\n";
	exit(1);
    }
    my ($first, $last, $leadin, $leadout, @offsets) = @discinfo;

    my $s = Digest::SHA->new(1);
    $s->add(sprintf "%02X", int($first));
    $s->add(sprintf "%02X", int($last));

    my @a;
    for (my $i = 0; $i < 100; $i++) {
        $a[$i] = 0;
    }
    my $i = 0;
    foreach my $o ($leadout, @offsets) {
       $a[$i++] = int($o) + int($leadin);
    }
    for (my $i = 0; $i < 100; $i++) {
       $s->add(sprintf "%08X", $a[$i]);
    }

    my $id = $s->b64digest;
    # CPAN Digest modules do not pad their Base64 output, so we have to do it.
    while (length($id) % 4) {
        $id .= '=';
    }

    $id =~ tr#+#.#;
    $id =~ tr#/#_#;
    $id =~ tr#=#-#;

    print $id;
    if (-t STDOUT) { print "\n"; }
} else {
    print STDERR "Unknown command given.\n";
    pod2usage(1);
    exit(1);
}
__END__

=head1 NAME

abcde-musicbrainz-tool - Musicbrainz query tool

=head1 SYNOPSIS

 abcde-musicbrainz-tool [options]

 Options:
   --command {id|data|calcid} mode of operation (default: id)
   --device <DEV>             read from CD-ROM device DEV (default: /dev/cdrom)
   --discid <ID>              Disc ID to query with --command data.
   --discinfo <F> <L> <LI> <LO> <TRK1OFF> [<TRK2OFF> [...]]
                              Disc information for --command calcid.
   --workdir <DIR>            working directory (default: /tmp)
   --help                     print option summary
   --man                      full documentation

=head1 OPTIONS

=over 8

=item B<--command> I<{id|data|calcid}>

Select mode of operation:

=over 8

=item B<id>

Read the disc-ID from the disc in the given device, and print it, the number of tracks, their start sectors, and the duration of the disc in seconds, to stdout. Format:

 ID TRACKCOUNT OFFSET1 [OFFSET2 [...]] LENGTH_S

=item B<data>

Query MusicBrainz web service and store data into the workdir into cddbread.1, cddbread.2, ... files in the workdir.

=item B<calcid>

Calculate MusicBrainz ID from given B<--discinfo> data.

=back

=item B<--device>

Specify CD-ROM drive's device name, to read ID from with B<--command id>.

=item B<--discid>

Supply disc ID for B<--command data>.

=item B<--discinfo> I<<first track> <last track> <lead-in sector> <lead-out sector> <track1 offset> [<track2 offset> [...]]>

Supply disc information for B<--command calcid>.

=item B<--workdir> I<directory>

The cddbread.* output files from B<--command data> go into this directory.

=item B<--help>

Print a brief help message and exit.

=item B<--man>

Display full manual and exit.

=back