From f71d84d6376aaece61460af9711c09c6ec434444 Mon Sep 17 00:00:00 2001 From: Ralf Ertzinger Date: Sat, 5 Jan 2008 14:31:08 +0100 Subject: [PATCH] - Add metacafe grabber --- videosite/MetaCafeGrabber.pm | 81 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) create mode 100644 videosite/MetaCafeGrabber.pm diff --git a/videosite/MetaCafeGrabber.pm b/videosite/MetaCafeGrabber.pm new file mode 100644 index 0000000..bc798f3 --- /dev/null +++ b/videosite/MetaCafeGrabber.pm @@ -0,0 +1,81 @@ +package MetaCafeGrabber; + +use GrabberBase; +@ISA = qw(GrabberBase); + +use LWP::Simple qw(!get); +use HTML::Parser; +use Data::Dumper; + +use strict; + +sub new { + my $class = shift; + my $self = $class->SUPER::new(); + + $self->{'NAME'} = 'metacafe'; + $self->{'PATTERNS'} = ['(http://(?:[-a-zA-Z0-9_.]+\.)*metacafe.com/watch/(\d+)(?:\S+)?)']; + + bless($self, $class); + + return $self; +} + +sub _parse { + my $self = shift; + my $url = shift; + my $pattern = shift; + my $content; + my $metadata = {}; + my $p = HTML::Parser->new(api_version => 3); + my @accum; + my @text; + my $e; + + $url =~ m|$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://www.metacafe.com/watch/%s', $2)))) { + $self->error('Could not download %s', $url); + return undef; + } + + $p->handler(start => \@accum, "tagname, attr"); + $p->handler(text => \@text, "text"); + $p->report_tags(qw(meta script)); + $p->utf8_mode(1); + $p->parse($content); + + # Look for the title in the meta tags + foreach $e (@accum) { + if ('meta' eq $e->[0]) { + if ('title' eq $e->[1]->{'name'}) { + $metadata->{'TITLE'} = $e->[1]->{'content'}; + $metadata->{'TITLE'} =~ s/^Metacafe\s*-\s*//; + } + } + } + + # Look for the download URL + foreach $e (@text) { + if ($e->[0] =~ m|\.addParam\("flashvars", ".*\&mediaURL=([^\&]+)\&.*"|) { + $metadata->{'DLURL'} = $1; + } + } + + unless(defined($metadata->{'DLURL'}) && defined($metadata->{'TITLE'})) { + $self->error('Could not determine download URL'); + return undef; + } + + return $metadata; +} + +1; -- 1.8.3.1