# (c) 2010 by Ralf Ertzinger # licensed under GNU GPL v2 # # Grabber for twitter package quotesite::TwitterGrabber; use quotesite::GrabberBase; @ISA = qw(quotesite::GrabberBase); use LWP::Simple qw(!get); use Data::Dumper; use XML::Simple; use HTML::Entities qw(decode_entities); use Encode; use strict; sub new { my $class = shift; my $self = $class->SUPER::new(); $self->{'NAME'} = 'twitter.com'; $self->{'PATTERNS'} = ['(https?://twitter.com/(?:\#\!/)?[^/]+/status(?:es)?/(\d+))']; bless($self, $class); $self->_prepare_parameters(); return $self; } sub _parse { my $self = shift; my $url = shift; my $pattern = shift; my $content; my $metadata = {}; my $p = XML::Simple->new(); my $t; $url =~ m|$pattern|; $url = $1; $metadata->{'URL'} = $url; $metadata->{'ID'} = $2; $metadata->{'TYPE'} = 'quote'; $metadata->{'SOURCE'} = $self->{'NAME'}; $metadata->{'CONTENT'} = undef; # Get the XML file containing the quote unless(defined($content = LWP::Simple::get(sprintf('http://api.twitter.com/1/statuses/show/%s.xml', $2)))) { $self->error('Could not download quote'); return undef; } unless(defined($t = $p->XMLin($content))) { $self->error('Could not parse XML metadata'); return undef; } $metadata->{'CONTENT'} = decode_entities($t->{'text'}); $metadata->{'ID'} = $t->{'user'}->{'screen_name'} . '/' . $metadata->{'ID'}; unless(defined($metadata->{'CONTENT'})) { $self->error('Could not extract quote content'); return undef; } return $metadata; } 1;