# (c) 2007 by Ralf Ertzinger # licensed under GNU GPL v2 # # Grabber for youtube.com/de/... # # download strategy revised using # http://www.kde-apps.org/content/show.php?content=41456 package YouTubeGrabber; use GrabberBase; @ISA = qw(GrabberBase); use LWP::Simple qw(!get); use HTML::TokeParser; use Data::Dumper; use strict; sub new { my $class = shift; my $self = $class->SUPER::new(); $self->{'NAME'} = 'youtube'; $self->{'PATTERNS'} = ['(http://(?:[-a-zA-Z0-9_.]+\.)*youtube.(?:com|de|co.uk)/watch\?(?:.+=.+&)*v=([-a-zA-Z0-9_]+))', '(http://(?:[-a-zA-Z0-9_.]+\.)*youtube.(?:com|de|co.uk)/v/([-a-zA-Z0-9_]+))']; bless($self, $class); $self->_prepare_parameters(); return $self; } sub _parse { my $self = shift; my $url = shift; my $pattern = shift; my $content; my $metadata = {}; my $p; my $e; my $tag; $url =~ m|$pattern|; $url = $1; $metadata->{'URL'} = $url; $metadata->{'ID'} = $2; $metadata->{'TYPE'} = 'video'; $metadata->{'SOURCE'} = $self->{'NAME'}; $metadata->{'TITLE'} = undef; $metadata->{'DLURL'} = undef; unless(defined($content = LWP::Simple::get(sprintf('http://youtube.com/watch?v=%s', $2)))) { $self->error('Could not download %s', $url); return undef; } $p = HTML::TokeParser->new(\$content); while ($tag = $p->get_tag('div', 'meta', 'script')) { if ('meta' eq $tag->[0]) { if ('title' eq $tag->[1]->{'name'}) { $metadata->{'TITLE'} = $tag->[1]->{'content'}; $self->debug('Title found: %s', $metadata->{'TITLE'}); } } elsif ('script' eq $tag->[0]) { $e = $p->get_text(); if ($e =~ m|/watch_fullscreen\?(.+)\x27|) { my %args = map { split(/=/, $_, 2); } split(/&/, $1); $metadata->{'DLURL'} = sprintf('http://www.youtube.com/get_video.php?video_id=%s&t=%s', $metadata->{'ID'}, $args{'t'}); $self->debug('URL found: %s', $metadata->{'DLURL'}); } } elsif ('div' eq $tag->[0]) { if ('errorBox' eq $tag->[1]->{'class'}) { $self->error("Could not get video data for youtube %s: %s", $metadata->{'ID'}, $p->get_trimmed_text()); return undef; } } } unless(defined($metadata->{'DLURL'}) && defined($metadata->{'TITLE'})) { $self->error('Could not determine download URL'); return undef; } return $metadata; } 1;