Merge branch 'master' of ssh://sun@ryoko:22003/home/sun/GIT/quotesite

[quotesite.git] / quotesite / iBashGrabber.pm

diff --git a/quotesite/iBashGrabber.pm b/quotesite/iBashGrabber.pm
new file mode 100644 (file)
index 0000000..04bd07d
--- /dev/null
+++ b/quotesite/iBashGrabber.pm
@@ -0,0 +1,72 @@
+# (c) 2007 by Ralf Ertzinger <ralf@camperquake.de>
+# licensed under GNU GPL v2
+#
+# Grabber for german-bash.org
+
+package iBashGrabber;
+
+use GrabberBase;
+@ISA = qw(GrabberBase);
+
+use LWP::Simple qw(!get);
+use HTML::TokeParser;
+use Data::Dumper;
+use Encode;
+
+use strict;
+
+sub new {
+    my $class = shift;
+    my $self = $class->SUPER::new();
+
+    $self->{'NAME'} = 'ibash.de';
+    $self->{'PATTERNS'} = ['(http://(?:[-a-zA-Z0-9_.]+\.)*ibash.de/zitat_(\d+)\.html)'];
+
+    bless($self, $class);
+    $self->_prepare_parameters();
+
+    return $self;
+}
+
+sub _parse {
+    my $self = shift;
+    my $url = shift;
+    my $pattern = shift;
+    my $content;
+    my $metadata = {};
+    my $p;
+    my $t;
+
+    $url =~ m|$pattern|;
+    $url = $1;
+
+    $metadata->{'URL'} = $url;
+    $metadata->{'ID'} = $2;
+    $metadata->{'TYPE'} = 'quote';
+    $metadata->{'SOURCE'} = $self->{'NAME'};
+    $metadata->{'CONTENT'} = undef;
+
+    # Get the HTML file containing the quote
+    unless(defined($content = LWP::Simple::get(sprintf('http://www.ibash.de/zitat_%s.html', $2)))) {
+        $self->error('Could not download quote');
+        return undef;
+    }
+
+    $p = HTML::TokeParser->new(\$content);
+
+    while ($t = $p->get_tag('td')) {
+        if (exists($t->[1]->{'class'}) && ($t->[1]->{'class'} eq 'quote')) {
+            $metadata->{'CONTENT'} = $p->get_text('/td');
+            Encode::from_to($metadata->{'CONTENT'}, 'iso-8859-1', 'utf8');
+        }
+    }
+
+    unless(defined($metadata->{'CONTENT'})) {
+        $self->error('Could not extract quote content');
+        return undef;
+    }
+
+    return $metadata;
+}
+
+1;