inthenews

#!/usr/bin/perl
use strict;
use warnings;
use MIME::Lite;
use HTML::TreeBuilder;
use LWP::Simple qw(get $ua);
use YAML;
use Data::Dumper;
#binmode STDOUT, ":utf8";

# Wikipedia blocks LWP::Simple's default User-Agent
# http://stackoverflow.com/questions/24546/why-cant-i-fetch-wikipedia-pages-with-lwpsimple
$ua->agent('Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.0.18) Gecko/2010020219 Firefox/3.0.18');

my $url = 'http://en.wikipedia.org/wiki/Template:In_the_news';

#my $file = '/tmp/Template:In_the_news';
#my $tree = HTML::TreeBuilder->new->parse_file($file);

my $content = get("$url") or die "Couldn't download $url";

my $tree = HTML::TreeBuilder->new->parse($content)->eof;

my $body_content = $tree->look_down( id => 'mw-content-text' )->look_down( _tag => 'ul' );

my $data;

for my $story ( $body_content->look_down( _tag => 'li' ) ) {
    my $as_text = $story->as_trimmed_text;
    push( @$data, "$as_text" );
}

my $html;

for my $story ( $body_content->look_down( _tag => 'ul' ) ) {
    my $as_html = $story->as_HTML;
    $html .= $as_html;
}

$tree->delete();

$html =~ s{(<a href=")}{$1http://en.wikipedia.org}g;

my $msg = MIME::Lite->new(
    From       => '"greptilian.com" <philipdurbin@gmail.com>',
    'Reply-To' => 'philipdurbin@gmail.com',
    To         => 'philipdurbin@gmail.com',
    Subject    => 'Wikipedia - In the News',
    Type       => 'text/html',
    Data       => $html,
);

$msg->send();