X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=videosite%2FBreakGrabber.pm;h=0619da5c2650a124ee2c7bbb1ebbebc5a9a47815;hb=400621043a9ee2217a07e2afe5cedf23a2491d69;hp=4870da8be995885e9499d422b7a7848d6efc160a;hpb=c9eaebd4eec29a38c2b89c581438ce94e32b1814;p=videosite.git diff --git a/videosite/BreakGrabber.pm b/videosite/BreakGrabber.pm index 4870da8..0619da5 100644 --- a/videosite/BreakGrabber.pm +++ b/videosite/BreakGrabber.pm @@ -8,8 +8,7 @@ package BreakGrabber; use GrabberBase; @ISA = qw(GrabberBase); -use LWP::Simple qw(!get); -use HTML::Parser; +use HTMLHelper; use Data::Dumper; use strict; @@ -34,10 +33,8 @@ sub _parse { my $pattern = shift; my $content; my $metadata = {}; - my $p = HTML::Parser->new(api_version => 3); - my @accum; - my @text; - my $e; + my $p = HTMLHelper->new(); + my $n; $url =~ m|$pattern|; $url = $1; @@ -49,28 +46,20 @@ sub _parse { $metadata->{'TITLE'} = undef; $metadata->{'DLURL'} = undef; - unless(defined($content = LWP::Simple::get($1))) { + unless(defined($content = $p->load($1))) { $self->error('Could not download %s', $url); return undef; } - $p->handler(start => \@accum, "tagname, attr"); - $p->report_tags(qw(meta)); - $p->utf8_mode(1); - $p->parse($content); - - # Look for the title in the meta tags - foreach $e (@accum) { - if ('meta' eq $e->[0]) { - if ('embed_video_title' eq $e->[1]->{'name'}) { - $metadata->{'TITLE'} = $e->[1]->{'content'}; - } - - if ('embed_video_thumb_url' eq $e->[1]->{'name'}) { - $metadata->{'DLURL'} = $e->[1]->{'content'}; - $metadata->{'DLURL'} =~ s/\.jpg$/\.flv/; - } - } + $n = $p->findnodes('meta[@name="embed_video_title"]'); + if (defined($n)) { + $metadata->{'TITLE'} = $n->{'content'}; + } + + $n = $p->findnodes('meta[@name="embed_video_thumb_url"]'); + if (defined($n)) { + $metadata->{'DLURL'} = $n->{'content'}; + $metadata->{'DLURL'} =~ s/\.jpg$/\.flv/; } unless(defined($metadata->{'DLURL'}) && defined($metadata->{'TITLE'})) {