From 4b79658ad5745fd99983221b219619c22658d771 Mon Sep 17 00:00:00 2001 From: Ralf Ertzinger Date: Thu, 17 Apr 2008 17:19:03 +0200 Subject: [PATCH] - Add GoogleGrabber --- videosite/GoogleGrabber.pm | 86 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 86 insertions(+) create mode 100644 videosite/GoogleGrabber.pm diff --git a/videosite/GoogleGrabber.pm b/videosite/GoogleGrabber.pm new file mode 100644 index 0000000..c1fe9eb --- /dev/null +++ b/videosite/GoogleGrabber.pm @@ -0,0 +1,86 @@ +# (c) 2007 by Ralf Ertzinger +# licensed under GNU GPL v2 +# +# Grabber for video.google.com + +package GoogleGrabber; + +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'} = 'google'; + $self->{'PATTERNS'} = ['(http://video\.google\.com/videoplay\?docid=([-\d]+))']; + + bless($self, $class); + $self->_prepare_parameters(); + + return $self; +} + +sub _parse { + my $self = shift; + my $url = shift; + my $pattern = shift; + my $content; + my $metadata = {}; + my $p = 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://video.google.com/videoplay?docid=%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(embed div)); + $p->utf8_mode(1); + $p->parse($content); + + # Look for the title in the div tags + foreach $e (@accum) { + if ('div' eq $e->[0]) { + if ((exists($e->[1]->{'class'})) and ('title' eq $e->[1]->{'class'})) { + $metadata->{'TITLE'} = $e->[1]->{'title'}; + } + } + + if ('embed' eq $e->[0]) { + if ((exists($e->[1]->{'src'})) and ($e->[1]->{'src'} =~ m|^/googleplayer.swf\?\&videoUrl=([^&]+)\&|)) { + $metadata->{'DLURL'} = $1; + $metadata->{'DLURL'} =~ s/%(..)/chr(hex($1))/ge; + } + } + } + + unless(defined($metadata->{'DLURL'}) && defined($metadata->{'TITLE'})) { + $self->error('Could not determine download URL'); + return undef; + } + + return $metadata; +} + +1; -- 1.8.3.1