From: Ralf Ertzinger <sun@lain.camperquake.de>
Date: Sun, 30 Dec 2007 15:42:06 +0000 (+0100)
Subject: - Add grabber for DailyMotion
X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=commitdiff_plain;h=b97a01912322f6ebccf2713c3bbbcc266f02c7b5;p=videosite.git

- Add grabber for DailyMotion
---

diff --git a/videosite/DailyMotionGrabber.pm b/videosite/DailyMotionGrabber.pm
new file mode 100644
index 0000000..d08e71b
--- /dev/null
+++ b/videosite/DailyMotionGrabber.pm
@@ -0,0 +1,84 @@
+#
+# download strategy revised using
+# http://www.kde-apps.org/content/show.php?content=41456
+
+package DailyMotionGrabber;
+
+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'} = 'youtube';
+    $self->{'PATTERNS'} = ['(http://(?:[-a-zA-Z0-9_.]+\.)*.dailymotion.com/(?:[^/]+/)*video/([-a-zA-Z0-9_]+)'];
+
+    bless($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'} = 'dailymotion';
+    $metadata->{'TITLE'} = undef;
+    $metadata->{'DLURL'} = undef;
+
+    unless(defined($content = LWP::Simple::get(sprintf('http://www.dailymotion.com/video/%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'};
+            }
+        }
+    }
+
+    # Look for the download URL
+    foreach $e (@text) {
+        if ($e->[0] =~ m|\.add_variable("url", "([^\"]+)")|) {
+            $metadata->{'DLURL'} = $1;
+        }
+    }
+
+    unless(defined($metadata->{'DLURL'}) && defined($metadata->{'TITLE'})) {
+        $self->error('Could not determine download URL');
+        return undef;
+    }
+
+    return $metadata;
+}
+
+1;