C///

\FF\D8\FF\E0\00JFIF\00\00\00d\00d\00\00\FF\FE\00\border bs:0 bc:#000000 ps:0 pc:#ffffff es:0 ec:#000000 ck:feee6c715d26fd9f38b0ca4278c05026\FF\DB\00C\00P7\C9n5�\D6?\BD�\9Ds\EBp\9F[`8m\B7)o\B5\E8\E6I\99\FE3]]A2\BA\8Cw\D6E\93\\DEv\C8\009\F2\F1NI?uc\\F5\EA\96k\xN<~buv\EA\C8\D7 \8B\84\CEcxI\BBg\AE\9E=\D6+n\EC\80\C8A\8C\AE\EB\CF\D5\DA\E9"2\A4\B9j5\EB\F3W\B63\96\B30Yu\DA\FC8\ED\DF\E7Ms\FB\F1\8E\B3\FA\EA\E8\E6(\883zs\F2_\8DFk\8Bh \00\8C\DCw\D3R\B5+6X\BA\B2\C4j\AB0\B4\FCMw\C2I\8E\9B\E3\A9~9u\FA\D3l\80\C8%p\EE\FDn2� \00$\FEj\C4e\A9\DB\~\95\A7\A5\80EK\BB\8DDsP\00@@AD'k\CF\E8\DB\D2(\80\9AK\D3\85\D6lb\F2\BA\8C*\80\00)\95 59\A3R:\F3\CE"\B6\80\88\00\00i1u4�\E9\F2�\A6\A2\FACM\93WMb*\E0*\00\00\00\00\00(\A8\80\00\00\80\00\00\00\00\00\00\00\00\00\FF\D9 C/// package HTML::LinkExtor; require HTML::Parser; @ISA = qw(HTML::Parser); $VERSION = "3.69"; =head1 NAME HTML::LinkExtor - Extract links from an HTML document =head1 SYNOPSIS require HTML::LinkExtor; $p = HTML::LinkExtor->new(\&cb, "http://www.perl.org/"); sub cb { my($tag, %links) = @_; print "$tag @{[%links]}\n"; } $p->parse_file("index.html"); =head1 DESCRIPTION I is an HTML parser that extracts links from an HTML document. The I is a subclass of I. This means that the document should be given to the parser by calling the $p->parse() or $p->parse_file() methods. =cut use strict; use HTML::Tagset (); # legacy (some applications grabs this hash directly) use vars qw(%LINK_ELEMENT); *LINK_ELEMENT = \%HTML::Tagset::linkElements; =over 4 =item $p = HTML::LinkExtor->new =item $p = HTML::LinkExtor->new( $callback ) =item $p = HTML::LinkExtor->new( $callback, $base ) The constructor takes two optional arguments. The first is a reference to a callback routine. It will be called as links are found. If a callback is not provided, then links are just accumulated internally and can be retrieved by calling the $p->links() method. The $base argument is an optional base URL used to absolutize all URLs found. You need to have the I module installed if you provide $base. The callback is called with the lowercase tag name as first argument, and then all link attributes as separate key/value pairs. All non-link attributes are removed. =cut sub new { my($class, $cb, $base) = @_; my $self = $class->SUPER::new( start_h => ["_start_tag", "self,tagname,attr"], report_tags => [keys %HTML::Tagset::linkElements], ); $self->{extractlink_cb} = $cb; if ($base) { require URI; $self->{extractlink_base} = URI->new($base); } $self; } sub _start_tag { my($self, $tag, $attr) = @_; my $base = $self->{extractlink_base}; my $links = $HTML::Tagset::linkElements{$tag}; $links = [$links] unless ref $links; my @links; my $a; for $a (@$links) { next unless exists $attr->{$a}; (my $link = $attr->{$a}) =~ s/^\s+//; $link =~ s/\s+$//; # HTML5 push(@links, $a, $base ? URI->new($link, $base)->abs($base) : $link); } return unless @links; $self->_found_link($tag, @links); } sub _found_link { my $self = shift; my $cb = $self->{extractlink_cb}; if ($cb) { &$cb(@_); } else { push(@{$self->{'links'}}, [@_]); } } =item $p->links Returns a list of all links found in the document. The returned values will be anonymous arrays with the following elements: [$tag, $attr => $url1, $attr2 => $url2,...] The $p->links method will also trunca