X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=videosite%2FMyVideoGrabber.pm;h=c2685b782791a0de9ae3e85b887fe7ae473852ad;hb=09d45a4c4460dce9322f4528dad5123322ab6b38;hp=4b413870f1ab3ce47424799d259a3a769909ce76;hpb=911eeb36e674f916d08b04cd9c48bb33e96bf108;p=videosite.git diff --git a/videosite/MyVideoGrabber.pm b/videosite/MyVideoGrabber.pm index 4b41387..c2685b7 100644 --- a/videosite/MyVideoGrabber.pm +++ b/videosite/MyVideoGrabber.pm @@ -8,23 +8,22 @@ package videosite::MyVideoGrabber; use videosite::GrabberBase; @ISA = qw(videosite::GrabberBase); -use LWP::Simple qw(!get); -use HTML::Parser; +use HTML::TokeParser; use Data::Dumper; use strict; sub new { my $class = shift; - my $self = $class->SUPER::new(); - - $self->{'NAME'} = 'myvideo'; - $self->{'PATTERNS'} = ['(http://(?:[-a-zA-Z0-9_.]+\.)*myvideo.de/watch/(\d+))']; - - bless($self, $class); - $self->_prepare_parameters(); - - return $self; + my $self = $class->SUPER::new( + NAME => 'myvideo', + _SELFTESTURL => 'http://www.myvideo.de/watch/9191174/Battle_Worlds_Kronos_Alpha_Version_ausprobiert', + _SELFTESTTITLE => 'Battle Worlds: Kronos - Alpha-Version ausprobiert', + PATTERNS => ['(http://(?:[-a-zA-Z0-9_.]+\.)*myvideo.de/watch/(\d+))'], + @_, + ); + + return bless($self, $class); } sub _parse { @@ -33,10 +32,8 @@ 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; + my $t; $url =~ m|$pattern|; $url = $1; @@ -48,31 +45,24 @@ sub _parse { $metadata->{'TITLE'} = undef; $metadata->{'DLURL'} = undef; - unless(defined($content = LWP::Simple::get(sprintf('http://www.myvideo.de/watch/%s', $2)))) { + unless(defined($content = $self->simple_get(sprintf('http://www.myvideo.de/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'}; - $metadata->{'TITLE'} =~ s/\s*-\s*MyVideo$//; + $p = HTML::TokeParser->new(\$content); + while ($t = $p->get_tag('meta', 'link')) { + if ('meta' eq $t->[0]) { + if (exists($t->[1]->{property}) and ($t->[1]->{property} eq 'og:title')) { + $metadata->{'TITLE'} = $t->[1]->{content}; + } + } elsif ('link' eq $t->[0]) { + if (exists($t->[1]->{rel}) and ($t->[1]->{rel} eq 'image_src')) { + $metadata->{'DLURL'} = $t->[1]->{'href'}; + $metadata->{'DLURL'} =~ s,thumbs/[^/]*$,,; + $metadata->{'DLURL'} .= $metadata->{'ID'} . ".flv"; + $self->debug("Found URL: %s", $metadata->{'DLURL'}); } - } - } - - # Look for the download URL - foreach $e (@text) { - if ($e->[0] =~ m|SWFObject\(.*?\.swf\?([^\&]+)\&|) { - $metadata->{'DLURL'} = $1; } }