HTML::LinkExtractor - Extract links from an HTML document

HTML::LinkExtractor is used for extracting links from HTML. It is very similar to HTML::LinkExtor, except that besides getting the URL, you also get the link-text.

Example ( please run the examples ):

    use HTML::LinkExtractor;
    use Data::Dumper;

    my $input = q{If <a href="http://perl.com/"> I am a LINK!!! </a>};
    my $LX = new HTML::LinkExtractor();

    $LX->parse(\$input);

    print Dumper($LX->links);
    __END__
    # the above example will yield
    $VAR1 = [
              {
                '_TEXT' => '<a href="http://perl.com/"> I am a LINK!!! </a>',
                'href' => bless(do{\(my $o = 'http://perl.com/')}, 'URI::http'),
                'tag' => 'a'
              }
            ];

CWHTML::LinkExtractor will also correctly extract nested link-type tags.

    ## the demo
    perl LinkExtractor.pm
    perl LinkExtractor.pm file.html othefile.html

    ## or if the module is installed, but you don't know where

    perl -MHTML::LinkExtractor -e" system $^X, $INC{q{HTML/LinkExtractor.pm}} "
    perl -MHTML::LinkExtractor -e' system $^X, $INC{q{HTML/LinkExtractor.pm}} '

    ## or

    use HTML::LinkExtractor;
    use LWP qw( get ); #     use LWP::Simple qw( get );

    my $base = 'http://search.cpan.org';
    my $html = get($base.'/recent');
    my $LX = new HTML::LinkExtractor();

    $LX->parse(\$html);

    print qq{<base href="$base">\n};

    for my $Link( @{ $LX->links } ) {
    ## new modules are linked  by /author/NAME/Dist
        if( $$Link{href}=~ m{^\/author\/\w+} ) {
            print $$Link{_TEXT}."\n";
        }
    }

    undef $LX;
    __END__

    ## or

    use HTML::LinkExtractor;
    use Data::Dumper;

    my $input = q{If <a href="http://perl.com/"> I am a LINK!!! </a>};
    my $LX = new HTML::LinkExtractor(
        sub {
            print Data::Dumper::Dumper(@_);
        },
        'http://perlFox.org/',
    );

    $LX->parse(\$input);
    $LX->strip(1);
    $LX->parse(\$input);
    __END__

    #### Calculate to total size of a web-page
    #### adds up the sizes of all the images and stylesheets and stuff

    use strict;
    use LWP; #     use LWP::Simple;
    use HTML::LinkExtractor;
                                                        #
    my $url  = shift || 'http://www.google.com';
    my $html = get($url);
    my $Total = length $html;
                                                        #
    print "initial size $Total\n";
                                                        #
    my $LX = new HTML::LinkExtractor(
        sub {
            my( $X, $tag ) = @_;
                                                        #
            unless( grep {$_ eq $tag->{tag} } @HTML::LinkExtractor::TAGS_IN_NEED ) {
                                                        #
    print "$$tag{tag}\n";
                                                        #
                for my $urlAttr ( @{$HTML::LinkExtractor::TAGS{$$tag{tag}}} ) {
                    if( exists $$tag{$urlAttr} ) {
                        my $size = (head( $$tag{$urlAttr} ))[1];
                        $Total += $size if $size;
    print "adding $size\n" if $size;
                    }
                }
            }
        },
        $url,
        0
    );
                                                        #
    $LX->parse(\$html);
                                                        #
    print "The total size of \n$url\n is $Total bytes\n";
    __END__

Accepts 3 arguments, all of which are optional. If for example you want to pass a CW$baseUrl, but don't want to have a callback invoked, just put CWundef in place of a subref.

This is the only class method.

1

a callback ( a sub reference, as in CWsub{}, or CW\&sub) which is to be called each time a new LINK is encountered ( for CW@HTML::LinkExtractor::TAGS_IN_NEED this means after the closing tag is encountered ) The callback receives an object reference(CW$LX) and a link hashref.

2

and a base URL ( URI->new, so its up to you to make sure it's valid which is used to convert all relative URI's to absolute ones.

    $ALinkP{href} = URI->new_abs( $ALink{href}, $base );

3

A boolean (just stick with 1). See the example in DESCRIPTION. Normally, you'd get back _TEXT that looks like

    '_TEXT' => '<a href="http://perl.com/"> I am a LINK!!! </a>',

If you turn this option on, you'll get the following instead

    '_TEXT' => ' I am a LINK!!! ',

The private utility function CW_stripHTML does this by using HTML::TokeParsers method get_trimmed_text. You can turn this feature on an off by using CW$LX->strip(undef || 0 || 1) Each time you call CWparse, you should pass it a CW$filename a CW*FILEHANDLE or a CW\$FileContent

Each time you call CWparse a new CWHTML::TokeParser object is created and stored in CW$this->{_tp}.

You shouldn't need to mess with the TokeParser object. Only after you call CWparse will this method return anything. This method returns a reference to an ArrayOfHashes, which basically looks like (Data::Dumper output)

    $VAR1 = [ { tag => 'img', src => 'image.png' }, ];

Please note that if yo provide a callback this array will be empty. If you pass in CWundef (or nothing), returns the state of the option. Passing in a true or false value sets the option.

If you wanna know what the option does see CW$LX->new([\&callback, [$baseUrl, [1]]])

Take a look at CW%HTML::LinkExtractor::TAGS to see what I consider to be link-type-tag.

Take a look at CW@HTML::LinkExtractor::VALID_URL_ATTRIBUTES to see all the possible tag attributes which can contain URI's (the links!!)

Take a look at CW@HTML::LinkExtractor::TAGS_IN_NEED to see the tags for which the CW'_TEXT' attribute is provided, like CW<a href="#"> TEST </a>

I took at look at CW%HTML::Tagset::linkElements and the following URL's

    http://www.blooberry.com/indexdot/html/tagindex/all.htm

    http://www.blooberry.com/indexdot/html/tagpages/a/a-hyperlink.htm
    http://www.blooberry.com/indexdot/html/tagpages/a/applet.htm
    http://www.blooberry.com/indexdot/html/tagpages/a/area.htm

    http://www.blooberry.com/indexdot/html/tagpages/b/base.htm
    http://www.blooberry.com/indexdot/html/tagpages/b/bgsound.htm

    http://www.blooberry.com/indexdot/html/tagpages/d/del.htm
    http://www.blooberry.com/indexdot/html/tagpages/d/div.htm

    http://www.blooberry.com/indexdot/html/tagpages/e/embed.htm
    http://www.blooberry.com/indexdot/html/tagpages/f/frame.htm

    http://www.blooberry.com/indexdot/html/tagpages/i/ins.htm
    http://www.blooberry.com/indexdot/html/tagpages/i/image.htm
    http://www.blooberry.com/indexdot/html/tagpages/i/iframe.htm
    http://www.blooberry.com/indexdot/html/tagpages/i/ilayer.htm
    http://www.blooberry.com/indexdot/html/tagpages/i/inputimage.htm

    http://www.blooberry.com/indexdot/html/tagpages/l/layer.htm
    http://www.blooberry.com/indexdot/html/tagpages/l/link.htm

    http://www.blooberry.com/indexdot/html/tagpages/o/object.htm

    http://www.blooberry.com/indexdot/html/tagpages/q/q.htm

    http://www.blooberry.com/indexdot/html/tagpages/s/script.htm
    http://www.blooberry.com/indexdot/html/tagpages/s/sound.htm

    And the special cases

    <!DOCTYPE HTML SYSTEM "http://www.w3.org/DTD/HTML4-strict.dtd">
    http://www.blooberry.com/indexdot/html/tagpages/d/doctype.htm
    '!doctype'  is really a process instruction, but is still listed
    in %TAGS with 'url' as the attribute

    and

    <meta HTTP-EQUIV="Refresh" CONTENT="5; URL=http://www.foo.com/foo.html">
    http://www.blooberry.com/indexdot/html/tagpages/m/meta.htm
    If there is a valid url, 'url' is set as the attribute.
    The meta tag has no 'attributes' listed in %TAGS.

HTML::LinkExtor, HTML::TokeParser, HTML::Tagset.

D.H (PodMaster)

Please use http://rt.cpan.org/ to report bugs.

Just go to http://rt.cpan.org/NoAuth/Bugs.html?Dist=HTML-Scrubber to see a bug list and/or repot new ones.

Copyright (c) 2003, 2004 by D.H. (PodMaster). All rights reserved.

This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself. The LICENSE file contains the full text of the license.