=head1 NAME

WWW::TV::Series - Parse TV.com for TV Series information.

=head1 SYNOPSIS

  use WWW::TV::Series qw();
  my $series = WWW::TV::Series->new(name => 'Prison Break');

  my @episodes = $series->episodes;
  print $series->summary;

=head1 DESCRIPTION

The L<WWW::TV::Series> module parses TV.com series information using
L<LWP::UserAgent>.

=head1 METHODS

=cut

package WWW::TV::Series;
use strict;
use warnings;

our $VERSION = '0.14';

use Carp qw(croak);
use LWP::UserAgent qw();

=head2 new

    The new() method is the constructor. It takes the id of the show if
    you have previously looked that up, or the name of the show which
    will be used to perform a search and the id will be taken from the
    first result.

    Optional parameters let you set the season number or LWP user agent.

        # default usage
        my $series = WWW::TV::Series->new(name => 'Prison Break');
        my $series = WWW::TV::Series->new(id => 31635);

        # change user-agent from the default of "libwww-perl/#.##"
        my $series = WWW::TV::Series->new(id => 31635, agent => 'WWW::TV');

    It is recommended that you lookup the show first and use the ID,
    otherwise you just don't know what will be returned.

    The constructor also takes a single scalar as an argument and does
    it's best to figure out what you want. But due to some shows being
    all digits as a name (e.g. "24"), use of this is not recommended
    (and in future may be deprecated).

=cut

sub new {
    my $class = ref $_[0] ? ref(shift) : shift;

    my %data;

    if (@_ == 1) {
        # If they gave us a plain scalar argument, try our best to figure out
        # what it is. Of course this dies in the arse if you want to search
        # for a program with a name like '24'.
        if ($_[0] =~ /^\d+$/) {
            $data{id} = shift;
        }
        else {
            $data{name} = shift;
        }
    }
    elsif (scalar(@_) % 2 == 0) {
        %data = @_;
    }

    $data{agent} = $class->agent($data{agent});
    $data{site}  = $class->site ($data{site});
 
    $data{id} = $class->_get_first_search_result($data{name}, $data{agent}, $data{site})
        if exists $data{name};

    croak 'No id or name given to constructor' unless exists $data{id};
    croak "Invalid id: $data{id}" unless $data{id} =~ /^\d+$/;

    return bless {
        id      => $data{id},
        _season => $data{season} || 0,
        _agent  => $data{agent},
        _site   => $data{site},
        filled  => { id => 1 },
    }, $class;
}

sub _get_first_search_result {
    my ($class, $name, $agent, $site) = @_;

    my $ua = LWP::UserAgent->new( agent => $agent );
    my $rc = $ua->get(
        sprintf("http://%s.tv.com/search.php?type=Search&stype=ajax_search&search_type=program&qs=%s",
        	$site, $name)
    );
    croak "Unable to get search results for $name" unless $rc->is_success;

    for (split /\n/, $rc->content) {
        next unless m{<a href="http://\w+.tv.com/.*?show/(\d+)/summary.html};
        return $1;
    }
    croak 'Unable to find a show in the search results.';
}

=head2 summary

    Returns a string containing basic information about this series.

=cut

sub summary {
    my $self = shift;

    unless (exists $self->{filled}->{summary}) {
        ($self->{summary}) = $self->_html =~ m{
            <span\sclass="long">(.*?)</span>
        }smx;
        $self->{summary} =~ s/<br ?\/?>/\n/g;
        $self->{summary} =~ s/<a href="[^"]+">.*?<\/a>//g;
        $self->{summary} =~ s/^\s*//;
        $self->{summary} =~ s/\s*$//;
        $self->{filled}->{summary} = 1;
    }

    return $self->{summary};
}

=head2 genres

    Returns a list of all the genres that TV.com have categorised this series as.

    # in scalar context, returns a comma-delimited string
    my $genres = $series->genres;

    # in array context, returns an array
    my @genres = $series->genres;

=cut

sub genres {
    my $self = shift;

    unless (exists $self->{filled}->{genres}) {
        my ($genres_row) = $self->_html =~ m{
            <span\sclass="genres">(<a\shref=.*</a>)</span>
        }x;

        $self->{genres} =
            join(
                ', ',
                map { s/\s*<a href="[^"]+">(.*?)<\/a>\s*/$1/; $_ }
                split(/,/, $genres_row)
            );

        my @genres = split(/, /, $self->{genres});
        $self->{genres} = \@genres;
        $self->{filled}->{genres} = 1;
    }

    return wantarray ? @{$self->{genres}} : join(', ', @{$self->{genres}});
}

=head2 cast

    Returns a list of the cast members. The order is the same as they
    appear on TV.com, which is most likely nothing to go by, but
    in most cases is the main cast order.

    # in scalar context, returns a comma-delimited string
    my $cast = $series->cast;

    # in array context, returns an array
    my @cast = $series->cast;

=cut

sub cast {
    my $self = shift;

    unless (exists $self->{filled}->{cast}) {
        my ($cast_line) = $self->_html =~ m{<ul><li\s* ><div class="wrap"><span class="cast_member">(.*?)</ul>};
        my @cast;
        for my $person (split /<\/li>/, $cast_line) {
            next unless $person =~ m{<a href="[^"]+">(.*?)</a>};
            push @cast, $1;
        }
        $self->{cast} = \@cast;
        $self->{filled}->{cast} = 1;
    }

    return wantarray ? @{$self->{cast}} : join(', ', @{$self->{cast}});
}

=head2 name

    Returns a string containing the name of the series.

=cut

sub name {
    my $self = shift;

    unless (exists $self->{filled}->{name}) {
        ($self->{name}) = $self->_html =~ m{
            <div\sclass="content_title".*?>\n\n?
            <h1>(.*?):\s*<span>Summary</span></h1>\n
        }x;
        $self->{filled}->{name} = 1;
    }

    return $self->{name};
}

=head2 image

    Returns the url of an image that can be used to identify this series.

=cut

sub image {
    my $self = shift;

    unless (exists $self->{filled}->{image}) {
        ($self->{image}) = $self->_html =~ m{
          <div\sid="topslot">\s*\n
          (?:<a\shref="[^"]+">\n)?
          \s*<img\ssrc="([^"]+)"
        }x;
        $self->{filled}->{image} = 1;
    }

    return $self->{image};
}

=head2 episodes

    Returns an array of L<WWW::TV::Episode> objects in order.

    # All episodes
    my @episodes = $series->episodes;

    # Episodes for season 2 only
    my @episodes = $series->episodes( season => 2 );

=cut

sub episodes {
    my $self = shift;

    my %args;
    if (scalar(@_) % 2 == 0) {
        %args = @_;
    }

    my $season = exists $args{season} ? $args{season} : $self->{_season};

    unless ($self->{filled}->{episodes}->{$season}) {
        my $ua = LWP::UserAgent->new(agent => $self->agent);
        my $rc = $ua->get($self->episode_url($season));
        croak sprintf('Unable to fetch episodes for series %d, season %d', $self->id, $season)
            unless $rc->is_success;

        require WWW::TV::Episode;
        my ($episode_line) = $rc->content =~ m{<th class="ep_title">(.*)\n};
        my @episodes =
            grep { defined }
            map {
                my $ep;
                if (m#<a href=".*/episode/(\d+)/summary\.html[^"]*">(.*?)</a>#) {
                    $ep = WWW::TV::Episode->new(id => $1, name => $2, agent => $self->agent);
                }
                $ep;
            } split /<\/div>/, $episode_line;

        $self->{episodes}->{$season} = \@episodes;
        $self->{filled}->{episodes}->{$season} = 1;
    }

    return @{$self->{episodes}->{$season}};
}

sub _html {
    my $self = shift;

    unless ($self->{filled}->{html}) {
        my $ua = LWP::UserAgent->new (agent => $self->agent);
        my $rc = $ua->get($self->url);
        croak sprintf('Unable to fetch page for series %s', $self->id)
            unless $rc->is_success;

        $self->{html} =
            join(
                "\n",
                map { s/^\s*//; s/\s*$//; $_ }
                split /\n/, $rc->content
            );

        $self->{filled}->{html} = 1;
    }

    return $self->{html};
}

=head2 id

    The ID of this series, according to TV.com

=cut

sub id {
    my $self = shift;

    return $self->{id};
}

=head2 agent ($value)

Returns the current user agent setting, and sets to $value if provided.

=cut

sub agent {
    my $self = shift;   # may be called as $self or $class
    my $value = shift;

    if (ref $self) {
        if (defined $value) {    
            $self->{_agent} = $value;
        }
        return ($self->{_agent} || LWP::UserAgent::_agent);
    } else {
        return ($value || LWP::UserAgent::_agent);
    }
}

=head2 site ($value)

Returns the current mirror site setting, and sets to $value if provided.

Default site is "www"; other options include: us, uk, au

=cut

sub site {
    my $self = shift;  # may be called as $self or $class
    my $value = shift;

    if (ref $self) {
        if (defined $value) {    
			if ($value =~ m#^(au|uk|us|www|)$#i) {
				$self->{_site} = $value;
			} else {
				warn "Ignoring unknown site value: [$value]\n";
			}
        }
        return ($self->{_site} || 'www');
    } else {
        return ($value || 'www');
    }
}

=head2 url

    Returns the url that was used to create this object.

=cut

sub url {
    my $self = shift;

    return sprintf('http://%s.tv.com/show/%d/summary.html', $self->{_site}, $self->id);
}

=head2 episode_url ($season)

    Returns the url that is used to get the episode listings for this
    series.

    $season is optional ; defaults to "all"

=cut

sub episode_url {
    my $self = shift;
    my $season = shift || 'All';  # 0 == ALL seasons

    return sprintf(
        'http://%s.tv.com/show/%d/episode_listings.html?season=%s',
        $self->{_site}, $self->id, $season
    );
}

1;

__END__

=head1 SEE ALSO

L<WWW::TV::Episode>

=head1 KNOWN ISSUES

There isn't yet any caching support. I don't see a need for it, but if you feel
the need to implement it then don't let me stop you.

There also isn't support for proxy servers yet. LWP should use it from your
environment if you really need it, but who still uses them anyway? Isn't it all
done transparently these days.

=head1 BUGS

Please report any bugs or feature requests through the web interface
at L<http://rt.cpan.org/Dist/Display.html?Queue=WWW-TV>.

=head1 AUTHORS

Danial Pearce C<cpan@tigris.id.au>

Stephen Steneker C<stennie@cpan.org>

=head1 LICENCE AND COPYRIGHT

Copyright (c) 2006-2008 Danial Pearce C<cpan@tigris.id.au>. All rights reserved.

Some parts copyright 2007-2008 Stephen Steneker C<stennie@cpan.org>.