X-Git-Url: https://git.camperquake.de/gitweb.cgi?p=quotesite.git;a=blobdiff_plain;f=quotesite%2FTwitterGrabber.pm;h=ef4887a248129c26ec9f8edfbba7c87b4308c6fc;hp=57e0ec496e530c01d0f939826982005f897d733a;hb=18eb17221c111a8442538c54d2527e8c06b25803;hpb=89a192885c8132e2fe1076faa442e46627e31ba7 diff --git a/quotesite/TwitterGrabber.pm b/quotesite/TwitterGrabber.pm index 57e0ec4..ef4887a 100644 --- a/quotesite/TwitterGrabber.pm +++ b/quotesite/TwitterGrabber.pm @@ -3,14 +3,15 @@ # # Grabber for twitter -package TwitterGrabber; +package quotesite::TwitterGrabber; -use GrabberBase; -@ISA = qw(GrabberBase); +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; @@ -20,7 +21,7 @@ sub new { my $self = $class->SUPER::new(); $self->{'NAME'} = 'twitter.com'; - $self->{'PATTERNS'} = ['(http://twitter.com/[^/]+/status/(\d+))']; + $self->{'PATTERNS'} = ['(https?://twitter.com/(?:\#\!/)?[^/]+/status(?:es)?/(\d+))']; bless($self, $class); $self->_prepare_parameters(); @@ -57,7 +58,7 @@ sub _parse { return undef; } - $metadata->{'CONTENT'} = $t->{'text'}; + $metadata->{'CONTENT'} = decode_entities($t->{'text'}); $metadata->{'ID'} = $t->{'user'}->{'screen_name'} . '/' . $metadata->{'ID'}; unless(defined($metadata->{'CONTENT'})) {