Files
oddmuse/strip.pl
2003-12-24 04:24:52 +00:00

38 lines
1.1 KiB
Perl
Executable File

#!/usr/bin/perl
use CGI qw/:standard/;
use CGI::Carp qw(fatalsToBrowser);
use LWP::UserAgent;
if (not param('url')) {
print header(),
start_html('Link Stripping'),
h1('Link Stripping'),
p('Transforms HTML into a plain-text list of link texts, one link per line.',
'For example, the HTML <a href="foo.html">bar</a> will be transformed to',
'the text "bar".'),
start_form(-method=>'GET'),
p('HTML feed: ', textfield('url'), submit()),
end_form(),
end_html();
exit;
}
print header(-type=>'text/plain; charset=UTF-8');
$ua = LWP::UserAgent->new;
$request = HTTP::Request->new('GET', param('url'));
$response = $ua->request($request);
$data = $response->content;
$p = HTML::Parser->new(api_version => 3);
$p->handler( start => \&start_handler, "tagname,self");
%pages = ();
$p->parse($data);
$p->eof; # signal end of document
print join("\n", sort keys %pages), "\n";
sub start_handler {
return if shift ne "a";
my $self = shift;
$self->handler(text => sub { $pages{(shift)} = 1 }, "dtext");
$self->handler(end => sub { $self->handler(text => ""); });
}