#! /usr/bin/perl
#
#        $Id: dbconv,v 1.14 2007/11/29 04:40:02 doi Exp doi $
#
#        dbconv - convert seiyuu databases
#
#        this program will convert the format of the seiyuu databases,
#        which are maintained by Hitoshi Doi <doi@usagi.org>.
#        it is mainly used for creating parts of the WWW seiyuu pages.
#
#        the seiyuu information is available via WWW at:
#                http://www.usagi.org/doi/seiyuu/
#
#        original hack by Hitoshi Doi (doi@usagi.org), 1995.10.17
#
#        this program is public domain.  do whatever you want with it.
#        this program is offered "as is", and does not have any warrantee.
#
#        ---------------------------------------------------------------
#        modifications
#
#        1996.03.18        Hitoshi Doi
#                support title format version 2
#                generate indexed WWW info
#
#        1996.05.25        Hitoshi Doi
#                added -listall option
#
#        2001.02.09        Hitoshi Doi
#                support continuation lines
#        ---------------------------------------------------------------

$me = $0;
$me =~ s/.*\///;

$print_out = 'VA';
$wwwtype = '';
$wwwdir = '';
$listall = '';
$listall2 = '';
$do_mp3 = 0;
$MP3_HOME = 'http://mp3.usagi.org/';

while ($ai = shift(@ARGV)) {
        if ($ai =~ m/-dbformat/i) {
                $print_out = 'DB';
        }
        elsif ($ai =~ m/^-dbprefix/i) {
                $DB_prefix = shift(@ARGV);
        }
        elsif ($ai =~ m/^-vaformat/i) {
                $print_out = 'VA';
        }
        elsif ($ai =~ m/^-db/i) {
                $dbfile = shift(@ARGV);
        }
        elsif ($ai =~ m/^-www/i) {
                $wwwtype = shift(@ARGV);
                $wwwdir = shift(@ARGV);
                $print_out = '';
        }
        elsif ($ai eq '-listall') {
                $listall = shift(@ARGV);
                $print_out = '';
        }
        elsif ($ai eq '-listall2') {
                $listall2 = shift(@ARGV);
                $print_out = '';
        }
        elsif ($ai =~ m/^-help/i) {
                print "$me - convert seiyuu databases

usage: $me -db FILE
options: -dbformat          print out in DB format
         -vaformat          print out in VA format [default]
         -www cd SUBDIR     generate indexed WWW info for CDs
         -www title SUBDIR  generate indexed WWW info for titles
         -www role SUBIDR   generate indexed WWW info for roles
         -listall PATH      list all in HTML format, preceeded by PATH
         -listall2 PATH     list all in HTML TABLE format, preceeded by PATH
";
                exit 0;
        }
        elsif ($ai =~ m/^-debug/i) {
                $debug = 1;
        }
        elsif ($ai =~ m/^-mp3/i) {
                $do_mp3 = 1;
        }
}

if (!$dbfile) {
        print "$me: the seiyuu database is undefined.
please specify the database with the -db option.\n";
        exit -1;
}
$ev_db = "seiyuu database $dbfile";
if (!open(DB, "<$dbfile")) {
        print "$me: can't open the $ev_db.\n";
        exit -1;
}

$style = " onMouseOver=\"this.style.background='pink'\"";
$style .= " onMouseOut=\"this.style.background=''\"";

&read_profile;
#
# go through the database and do processing
#
$DB_format = '';
while ($l = <DB>) {
        chop($l);
        next if ($l =~ m/^$/);
        next if ($l =~ m/^#/);

        while ($l =~ m/\\$/) {                # continuation
            $l =~ s/\\$//;
            $l .= <DB>;
            chop($l);
        }

        #
        # new title
        #
        if ($l =~ m/^[A-Za-z0-9\"]/) {
            $cur_title = $l;
            $cur_url = '';
            $cur_cover = '';
            next;
        }

        $l =~ s/^[ ]+//i;                # remove leading spaces
        #
        # special keywords
        #
        if ($l =~ m/^URL/) {
            $l =~ s/^URL[ ]*//i;
            $cur_url = $l;
            if ($listall) {
                print "<LI><A HREF=\"$listall/$cur_url\">";
                print "$cur_title</A>\n";
            }
            if ($listall2) {
                my @x = split(/,/, $cur_title);
                print "<TR $style valign=top><TD>$x[1]</TD><TD>";
                print "<A HREF=\"$listall2/$cur_url\">$x[0]</A></TD></TR>\n";
            }
            next;
        }
        if ($l =~ m/^COVER/) {
            $l =~ s/^COVER[ ]*//i;
            $cur_cover = $l;
            next;
        }
        if ($l =~ m/^MP3/) {
            next if (!$do_mp3);
            $l =~ s/^MP3[ ]*//i;
            $MP3 = $l;
            next;
        }
        #
        # character :: voice actor info
        #
        ($chara, $val) = split(/::/, $l);
        ($va, $debut) = split(/--/, $val);
        $chara =~ s/[ ]*$//;
        $chara =~ s/^[ ]*//;
        $va =~ s/^[ ]*//;
        $va =~ s/[ ]*$//;
        $eps = '';
        if ($va =~ m/\{.*\}/) {
            $eps = $va;
            $va =~ s/[ ]*\{.*\}.*//;
            $eps =~ s/.*\{//;
            $eps =~ s/\}.*//;
        }
        $debut =~ s/^[ ]*//;

        next if (!$va);

        #
        # create WWW info
        #
        if ($wwwdir) {
            $val = $va;
            $val =~ y/A-Z/a-z/;
            $s_titles{$val} .= "$cur_title\t";
            $s_url{$val} .= "$cur_url\t";
            $s_role{$val} .= "$chara\t";
            $s_eps{$val} .= "$eps\t";
            $s_debut{$val} .= "$debut\t";
        }
        #
        # to VA format
        #
        if ($print_out eq 'VA') {
            if ($debut) {
                $VA_data{$va} .= " $cur_title -- $debut\n  $chara";
            }
            else {
                $VA_data{$va} .= " $cur_title\n  $chara";
            }
            if ($eps) {
                $VA_data{$va} .= " {$eps}";
            }
            $VA_data{$va} .= "\n";
        }
        #
        # to DB format
        #  [debut] char - title {eps} - va (kanji)
        #
        if ($print_out eq 'DB') {
            $DB_format = "$DB_prefix ";
            $DB_format .= "[$debut] " if ($debut);
            if ($MP3) {
                $DB_format .= "<A HREF=\"$MP3_HOME/$MP3\">";
                $DB_format .= "$chara</A> - ";
            }
            else {
                $DB_format .= "$chara - ";
            }
            if ($cur_url) {
                $DB_format .= "<A HREF=\"";
                if ($cur_url =~ m/^\//) {
                    $DB_format .= $cur_url;
                }
                else {
                    $DB_format .= "/doi/$cur_url";
                }
                ($tt, $dt) = split(/,/, $cur_title);
                $DB_format .= "\">$tt</A> $dt";
            }
            else {
                $DB_format .= "$cur_title";
            }
            if ($eps) {
                $DB_format .= " {$eps}";
            }
            $va .= " ($kanji{$va})" if ($kanji{$va});
            $DB_format .= " - $va\n";
            print $DB_format;
            $DB_format = '';
        }
        $MP3 = '';
}
close(DB);

#
# print out VA format
#
if ($print_out eq 'VA') {
    foreach $k (sort keys(%VA_data)) {
        print "$k\n$VA_data{$k}\n";
    }
}

if ($wwwdir) {
    &write_wwwcd if ($wwwtype eq 'cd');
    &write_wwwtitle if ($wwwtype eq 'title');
    &write_wwwrole if ($wwwtype eq 'role');
}

exit 0;

#
# subroutines
#

sub read_profile
{
    $pf = '/home/doi/www/seiyuu/db/DB.profile';
    open(PF, "<$pf");
    my @all = <PF>;
    close(PF);
    %kanji = ();
    foreach $pl (@all) {
        chop($pl);
        ($r, $k) = split(/\t/, $pl, 2);
        $kanji{$r} = $k;
    }
}

#
# write data for WWW use (for CD)
#
sub write_wwwcd
{
    my(@e_name, @e_url, @e_role);
    my($en, $eu, $er);
    my $www_index = "$wwwdir/Index";
    my $www_out;

    system "/bin/mv $wwwdir $wwwdir.old";
    mkdir($wwwdir, 0755);
    open(INDEX, ">$www_index");
    my $i = 0;
    foreach $k (keys(%s_titles)) {
        printf INDEX "%05d:%s\n", $i, $k;
        $www_out = sprintf("$wwwdir/%05d", $i);
        open(WWWOUT, ">$www_out");
        @e_name = split(/\t/, $s_titles{$k});
        @e_url = split(/\t/, $s_url{$k});
        @e_role = split(/\t/, $s_role{$k});
        @e_eps = split(/\t/, $s_eps{$k});
        while ($en = shift(@e_name)) {
            $eu = shift(@e_url);
            $er = shift(@e_role);
            $ep = shift(@e_eps);
            my @x = split(/,/, $en);
            print WWWOUT "<TR valign=top><TD>$x[1]</TD><TD>";
            if ($eu) {
                if ($eu =~ m/^\//) {
                    print WWWOUT "<A HREF=\"$eu\">";
                }
                else {
                    print WWWOUT "<A HREF=\"../$eu\">";
                }
            }
            print WWWOUT "$x[0]";
            if ($eu) {
                print WWWOUT "</A>";
            }
            print WWWOUT "</TD><TD>" . $er;
            print WWWOUT " {$ep}" if ($ep);
            print WWWOUT "</TD></TR>\n";
        }
        close(WWWOUT);
        $i++;
    }
    close(INDEX);
}

#
# write data for WWW use (for titles)
#
sub write_wwwtitle
{
    my(@e_name, @e_url, @e_role);
    my($en, $eu, $er);
    my $www_index = "$wwwdir/Index";
    my $www_out;

    system "/bin/mv $wwwdir $wwwdir.old";
    mkdir($wwwdir, 0755);
    open(INDEX, ">$www_index");
    my $i = 0;
    foreach $k (keys(%s_titles)) {
        printf INDEX "%05d:%s\n", $i, $k;
        $www_out = sprintf("$wwwdir/%05d", $i);
        open(WWWOUT, ">$www_out");
        @e_name = split(/\t/, $s_titles{$k});
        @e_url = split(/\t/, $s_url{$k});
        while ($en = shift(@e_name)) {
            $eu = shift(@e_url);
            print WWWOUT "<LI>";
            if ($eu) {
                if ($eu =~ m/^\//) {
                    print WWWOUT "<A HREF=\"$eu\">";
                }
                else {
                    print WWWOUT "<A HREF=\"../$eu\">";
                }
            }
            print WWWOUT "$en";
            if ($eu) {
                print WWWOUT "</A>";
            }
            print WWWOUT "\n";
        }
        close(WWWOUT);
        $i++;
    }
    close(INDEX);
}

#
# write data for WWW use (for roles)
#
sub write_wwwrole
{
    my(@e_name, @e_url, @e_role);
    my($en, $eu, $er);
    my $www_index = "$wwwdir/Index";
    my $www_out;

    system "/bin/mv $wwwdir $wwwdir.old";
    mkdir($wwwdir, 0755);
    open(INDEX, ">$www_index");
    my $i = 0;
    foreach $k (keys(%s_titles)) {
        printf INDEX "%05d:%s\n", $i, $k;
        $www_out = sprintf("$wwwdir/%05d", $i);
        open(WWWOUT, ">$www_out");
        @e_name = split(/\t/, $s_titles{$k});
        @e_url = split(/\t/, $s_url{$k});
        @e_role = split(/\t/, $s_role{$k});
        @e_eps = split(/\t/, $s_eps{$k});
        @e_debut = split(/\t/, $s_debut{$k});
        while ($en = shift(@e_name)) {
            $eu = shift(@e_url);
            $er = shift(@e_role);
            $ee = shift(@e_eps);
            $ed = shift(@e_debut);
            print WWWOUT "<TR valign=top><TD> ";
            if ($eu) {
                if ($eu =~ m/^\//) {
                    print WWWOUT "<A HREF=\"$eu\">";
                }
                else {
                    print WWWOUT "<A HREF=\"../$eu\">";
                }
            }
            print WWWOUT "$en";
            if ($eu) {
                print WWWOUT "</A>";
            }
            if ($ee) {
                print WWWOUT " {$ee}";
            }
            if ($ed) {
                print WWWOUT " <b>[$ed]</b>";
            }
            print WWWOUT " </TD><TD>$er</TD></TR>\n";
        }
        close(WWWOUT);
        $i++;
    }
    close(INDEX);
}