NAME HTTP::LRDD - link-based resource descriptor discovery SYNOPSIS use HTTP::LRDD; my $lrdd = HTTP::LRDD->new; my @descriptors = $lrdd->discover($resource); foreach my $descriptor (@descriptors) { my $description = $lrdd->parse($descriptor); # $description is an RDF::Trine::Model } DESCRIPTION Note: the LRDD specification has ceased to be, with some parts being merged into the host-meta Internet Draft. This CPAN module will go in its own direction, bundling up best-practice techniques for discovering links and descriptors for a given URI. Import Routine "use HTTP::LRDD (@predicates)" When importing HTTP::LRDD, you can optionally provide a list of predicate URIs (i.e. the URIs which rel values expand to). This may also include IANA-registered link types, which are short tokens rather than full URIs. If you do not provide a list of predicate URIs, then a sensible default set is used. Constructors "HTTP::LRDD->new(@predicates)" Create a new LRDD discovery object using the given predicate URIs. If @predicates is omitted, then the predicates passed to the import routine are used instead. "HTTP::LRDD->new_strict" Create a new LRDD discovery object using the 'describedby' and 'lrdd' IANA-registered predicates. "HTTP::LRDD->new_default" Create a new LRDD discovery object using the default set of predicates ('describedby', 'lrdd', 'wdrs:describedby', 'xhv:meta' and 'rdfs:seeAlso'). Public Methods "$lrdd->discover($resource_uri)" Discovers a descriptor for the given resource; or if called in a list context, a list of descriptors. A descriptor is a resource that provides a description for something. So, if the given resource URI was the web address for an image, then the descriptor might be the web address for a metadata file about the image. If the given URI was an e-mail address, then the descriptor might be a profile document for the person to whom the address belongs. The following sources are checked (in order) to find links to descriptors. * HTTP response headers ("Link" header; "303 See Other" status) * HTTP response message (RDF or RDFa) * https://HOSTNAME/.well-known/host-meta * http://HOSTNAME/.well-known/host-meta If none of the above is able to yield a link to a descriptor, then the resource URI itself may be returned if it is in a self-describing format (e.g. RDF). There is no guaranteed file format for the descriptor, but it is usually RDF, POWDER XML or XRD. This method can also be called without an object (as a class method) in which case, a temporary object is created automatically using "new". "$lrdd->parse($descriptor_uri)" Parses a descriptor in XRD or RDF (RDF/XML, RDFa, Turtle, etc). Returns an RDF::Trine::Model or undef if unable to process. This method can also be called without an object (as a class method) in which case, a temporary object is created automatically using "new". "$lrdd->process($resource_uri)" Performs the equivalent of "discover" and "parse" in one easy step. Calls "discover" in a non-list context, so only the first descriptor is used. "$lrdd->process_all($resource_uri)" Performs the equivalent of "discover" and "parse" in one easy step. Calls "discover" in a list context, so multiple descriptors are combined into the resulting graph. EXAMPLES Discover the hub address (PubSubHubub) for a feed: my $lrdd = HTTP::LRDD->new('hub'); my $hub = $lrdd->discover('http://example.net/feed.atom'); Discover an author link (rel="author") from an HTML page: my $lrdd = HTTP::LRDD->new('author'); my $author = $lrdd->discover('http://example.com/page.html'); (For RDF people, you should note that rel="author" is not semantically equivalent to the "foaf:maker" property but closer to the "foaf:maker/foaf:homepage" SPARQL 1.1 property path - i.e. the rel="author" link destination is not a URI for the author themselves, but a page about the author.) If that author resource is in a machine-readable format (e.g. RDF), then parse the data: my $author_data = $lrdd->parse($author); Or, you can combine "discover" and "parse": my $lrdd = HTTP::LRDD->new('author'); my $author_data = $lrdd->process('http://example.com/page.html'); Get metadata for an image: my $lrdd = HTTP::LRDD->new; my $data = $lrdd->process_all('http://example.org/flower.jpeg'); As we're not passing any arguments to the constructor, we can use a shortcut: my $data = HTTP::LRDD->process_all('http://example.org/flower.jpeg'); Find the title of the image: use RDF::TrineShortcuts qw/:default :flatten/; my @results = flatten_iterator(rdf_query( "SELECT ?t WHERE { <http://example.org/flower.jpeg> dc:title ?t }")); if (@results) { printf("The title is: %s\n", $results[0]->{'title'}); } else { warn "Could not find title for image."; } BUGS Please report any bugs to <http://rt.cpan.org/>. Note: many problems can stem from servers that send incorrect "Content-Type" headers. If you send an XRD file as "text/html", then this module will not guess what you're doing - it will assume the file is really HTML, and inspect it for RDFa. For host-meta files, this module is slightly more relaxed, as there's a strong assumption that they are XRD... but YOU SHOULD NOT RELY ON THIS. If you're running a server, use the correct media type. SEE ALSO HTTP::Link::Parser, XRD::Parser, XML::Atom::OWL WWW::Finger, RDF::TrineShortcuts. <http://www.perlrdf.org/>. AUTHOR Toby Inkster <tobyink@cpan.org>. COPYRIGHT AND LICENCE Copyright 2010-2011 Toby Inkster This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself. DISCLAIMER OF WARRANTIES THIS PACKAGE IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.