NAME XML::RSS::LibXML - XML::RSS with XML::LibXML SYNOPSIS use XML::RSS::LibXML; my $rss = XML::RSS::LibXML->new; $rss->parsefile($file); print "channel: $rss->{channel}->{title}\n"; foreach my $item (@{ $rss->{items} }) { print " item: $item->{title} ($item->{link})\n"; } # Add custom modules $rss->add_module(uri => $uri, prefix => $prefix); # See docs for XML::RSS for these $rss->channel(...); $rss->add_item(...); $rss->image(...); $rss->textinput(...); $rss->save(...); $rss->as_string($format); DESCRIPTION XML::RSS::LibXML uses XML::LibXML (libxml2) for parsing RSS instead of XML::RSS' XML::Parser (expat), while trying to keep interface compatibility with XML::RSS. XML::RSS is an extremely handy tool, but it is unfortunately not exactly the most lean or efficient RSS parser, especially in a long-running process. So for a long time I had been using my own version of RSS parser to get the maximum speed and efficiency - this is the re-packaged version of that module, such that it adheres to the XML::RSS interface. Use this module when you have severe performance requirements working with RSS files. COMPATIBILITY There seems to be a bit of confusion as to how compatible XML::RSS::LibXML is with XML::RSS: XML::RSS::LibXML is NOT 100% compatible with XML::RSS. For instance XML::RS::LibXML does not do a complete parsing of the XML document because of the way we deal with XPath and libxml's DOM (see CAVEATS below) On top of that, I originally wrote XML::RSS::LibXML as sort of a fast replacement for XML::RAI, which looked cool in terms of abstracting the various modules. And therefore versions prior to 0.02 worked more like XML::RAI rather than XML::RSS. That was a mistake in hind sight, so it has been addressed (Since XML::RSS::LibXML version 0.08, it even supports writing RSS :) From now on XML::RSS::LibXML will try to match XML::RSS's functionality as much as possible in terms of parsing RSS feeds. Please send in patches and any tests that may be useful! PARSED STRUCTURE Once parsed the resulting data structure resembles that of XML::RSS. However, as one addition/improvement, XML::RSS::LibXML uses a technique to allow users to access complex data structures that XML::RSS doesn't support as of this writing. For example, suppose you have a tag like the following: ... foo bar baz All of the fields in this construct can be accessed like so: $rss->channel->{tag} # "foo bar baz" $rss->channel->{tag}{attr1} # "val1" $rss->channel->{tag}{attr2} # "val2" See XML::RSS::LibXML::MagicElement for details. METHODS new(%args) Creates a new instance of XML::RSS::LibXML. You may specify a version in the constructor args to control which output format as_string() will use. XML::RSS::LibXML->new(version => '1.0'); You can also specify the encoding that you expect this RSS object to use when creating an RSS string XML::RSS::LiBXML->new(encoding => 'euc-jp'); parse($string) Parse a string containing RSS. parsefile($filename) Parse an RSS file specified by $filename channel(%args) add_item(%args) image(%args) textinput(%args) These methods are used to generate RSS. See the documentation for XML::RSS for details. Currently RSS version 0.9, 1.0, and 2.0 are supported. as_string($format) Return the string representation of the parsed RSS. If $format is true, this flag is passed to the underlying XML::LibXML object's toString() method. By default, $format is true. add_module(uri => $uri, prefix => $prefix) Adds a new module. You should do this before parsing the RSS. XML::RSS::LibXML understands a few modules by default: rdf => "http://www.w3.org/1999/02/22-rdf-syntax-ns#", dc => "http://purl.org/dc/elements/1.1/", syn => "http://purl.org/rss/1.0/modules/syndication/", admin => "http://webns.net/mvcb/", content => "http://purl.org/rss/1.0/modules/content/", cc => "http://web.resource.org/cc/", taxo => "http://purl.org/rss/1.0/modules/taxonomy/", So you do not need to add these explicitly. save($file) Saves the RSS to a file items() Syntactic sugar to allow statement like this: foreach my $item ($rss->items) { ... } Instead of foreach my $item (@{$rss->{items}}) { ... } In scalar context, returns the reference to the list of items. PERFORMANCE Here's a simple benchmark using benchmark.pl in this distribution: daisuke@localhost XML-RSS-LibXML$ perl -Mlib=lib benchmark.pl index.rdf Rate rss rss_libxml rss 4.40/s -- -86% rss_libxml 32.2/s 633% -- CAVEATS Only first level data under and tags are examined. So if you have complex data, this module will not pick it up. For most of the cases, this will suffice, though. Some of the structures will need to be handled via XML::RSS::LibXML::MagicElement. For example, XML::RSS's SYNOPSIS shows a snippet like this: $rss->add_item(title => "GTKeyboard 0.85", # creates a guid field with permaLink=true permaLink => "http://freshmeat.net/news/1999/06/21/930003829.html", # alternately creates a guid field with permaLink=false # guid => "gtkeyboard-0.85 enclosure => { url=> 'http://example.com/torrent', type=>"application/x-bittorrent" }, description => 'blah blah' ); However, the enclosure element will need to be an object: enclosure => XML::RSS::LibXML::MagicElement->new( attributes => { url => 'http://example.com/torrent', type=>"application/x-bittorrent" }, ); Also, some elements such as permaLink elements are not really parsed such that it can be serialized and parsed back and force. I could fix this, but that would break some compatibility with XML::RSS TODO Tests. Currently tests are simply stolen from XML::RSS. It would be nice to have tests that do more extensive testing for correctness SEE ALSO XML::RSS, XML::LibXML, XML::LibXML::XPathContext AUTHORS Copyright (c) 2005 Daisuke Maki , Tatsuhiko Miyagawa . All rights reserved. Development partially funded by Brazil, Ltd.