X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=videosite%2FYouTubeGrabber.pm;h=dc6bba72c9ab1d5d3d58a064414e5d5fe09b50bc;hb=56e21ccb455187b27d7ec4c8c12aa1bf2f61e22f;hp=e2c6e703d413764696c3978f83f207cfb7e8b5eb;hpb=3fc523f845e7b98c22b00411cb57ab9e261880ea;p=videosite.git diff --git a/videosite/YouTubeGrabber.pm b/videosite/YouTubeGrabber.pm index e2c6e70..dc6bba7 100644 --- a/videosite/YouTubeGrabber.pm +++ b/videosite/YouTubeGrabber.pm @@ -12,7 +12,7 @@ use GrabberBase; @ISA = qw(GrabberBase); use LWP::Simple qw(!get); -use HTML::Parser; +use HTML::TokeParser; use Data::Dumper; use strict; @@ -37,10 +37,9 @@ sub _parse { my $pattern = shift; my $content; my $metadata = {}; - my $p = HTML::Parser->new(api_version => 3); - my @accum; - my @text; + my $p; my $e; + my $tag; $url =~ m|$pattern|; $url = $1; @@ -57,34 +56,28 @@ sub _parse { return undef; } - if ($content =~ /This video has been removed due to terms of use violation/) { - $self->error('Video has been removed'); - return undef; - } + $p = HTML::TokeParser->new(\$content); - $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'} = $e->[1]->{'content'}; + while ($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'}); } - } - } - - # Look for the download URL - foreach $e (@text) { - if ($e->[0] =~ 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'}); + } 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; + } } }