--- name: "RDF::TrineShortcuts" section: "3pm" platform: "linux" package: "librdf-trineshortcuts-perl" version: "0.104-1" release: "Debian 7 (Wheezy)" platforms: ["linux"] url: "https://new.carta.tech/RDF::TrineShortcuts" --- # RDF::TrineShortcuts(3pm) totally unauthorised module for cheats and charlatans ## NAME RDF::TrineShortcuts - totally unauthorised module for cheats and charlatans ## SYNOPSIS ``` use RDF::TrineShortcuts; my $model = rdf_parse('http://example.com/data.rdf'); my $query = 'ASK { ?person a . }'; if (rdf_query($query, $model)) { print "Document describes a person.\n"; } else { print "Document doesn't describe a person.\n"; print "What does it describe? Let's see...\n"; print rdf_string($model); } ``` ## DESCRIPTION This module exports three functions which simplify frequently performed tasks using RDF::Trine and RDF::Query. (A number of other functions are also available but not exported by default.) - • `"rdf_parse($data)"` - • `"rdf_string($model, $format)"` - • `"rdf_query($sparql, $endpoint_or_model)"` In addition, because it calls "use RDF::Trine", "use RDF::Query", and "use RDF::Query::Client", your code doesn't need to. ### Main Functions - "rdf\_parse($data)" `$data` can be some serialised RDF (in RDF/XML, Turtle, RDF/JSON, or any other format that RDF::Trine::Parser supports); or a URI (string or URI object); or an HTTP::Message object; or a hashref (as per RDF::Trine::Model's add\_hashref method); or a file name or an open file handle; or an RDF::Trine::Iterator::Graph. Essentially it could be anything you could reasonably expect to grab RDF from. It can be undef. If `$input` is a blessed object that rdf\_parse is unable to natively deal with, it will attempt to call `"$input->TO_RDF"` and deal with the result instead. (This is similar in spirit to the JSON module's convert\_blessed functionality.) The function returns an RDF::Trine::Model. There are additional optional named arguments, of which the two most useful are probably 'base', which sets the base URI for any relative URI references; and 'type', which indicates the media type of the input (though the function can usually guess this quite reliably). ``` $model = rdf_parse($input, 'base' => 'http://example.com/', 'type' => 'application/rdf+xml'); ``` Other named arguments include 'model' to provide an existing RDF::Trine::Model to add statements to; and 'context' for providing a context/graph URI (which may be a string, URI object or RDF::Trine::Node). - "rdf\_string($model, $format)" Serialises an RDF::Trine::Model to a string. `$model` is the model to serialise. If `$model` is not an RDF::Trine::Model object, then it's automatically passed through rdf\_parse first. `$format` is the format to use. One of 'RDFXML' (the default), 'RDFJSON', 'Turtle', 'Canonical NTriples' or 'NTriples'. If `$format` is not one of the above, then the function will try to guess what you meant. Preferred namespace names can be provided as a named argument: ``` print rdf_string($model, 'turtle', namespaces => { foo=>'http://example.com/vocabs/foo#' }); ``` You can find the relevant Internet media type like this: ``` my $type; my $string = rdf_string($model, 'rdfxml', media_type=>\$type); print $cgi->header($type), $string and exit; ``` - "rdf\_query($sparql, $endpoint)" `$sparql` is a SPARQL query to be run at `$endpoint`. `$endpoint` may be either an endpoint URI (string or URI object) or a model supported by RDF::Query (e.g. an RDF::Trine::Model.) Query languages other than SPARQL may be used (see \ for a list of supported languages). e.g. ``` rdf_query("SELECT ?s, ?p, ?o WHERE (?s, ?p, ?o)" $model, query_lang=>'rdql'); ``` Options query\_base, query\_update and query\_load\_data correspond to the base, update and load\_data options passed to RDF::Query's constructor. If the SPARQL query returns a boolean (i.e. an ASK query), then this function returns a boolean. If the query returns a graph (i.e. CONSTRUCT or DESCRIBE), then this function returns an RDF::Trine::Model corresponding to the graph. Otherwise (i.e. SELECT) it returns an RDF::Trine::Iterator object. For queries which return a graph, an optional `$model` parameter can be passed containing an existing RDF::Trine::Model to add statements to: ``` rdf_query("CONSTRUCT {?s ?p ?o} WHERE {?s ?p ?o}", 'http://example.com/sparql', model => $model); ``` This function can expand a small set of commonly used prefixes. For example: ``` $result = rdf_query('SELECT ?id ?name {?id foaf:name ?name}', $model); ``` The hashref `$RDF::TrineShortcuts::Namespaces` is consulted for expansions. ### Additional Functions These are not exported by default, so need to be imported explicitly, e.g. ``` use RDF::TrineShortcuts qw(:default rdf_node rdf_statement); ``` - "rdf\_node($value, %args)" Creates an RDF::Trine::Node object. Will attempt to automatically determine whether `$value` is a blank node, resource, literal or variable, but an optional named argument 'type' can be used to explicitly indicate this. For literals, named arguments 'datatype' and 'lang' are allowed. If 'datatype' is not a URI, then it's assumed to be an XSD datatype. ``` $node = rdf_node("Hello", type=>'literal', lang=>'en'); ``` For resources, the named argument 'base' is allowed. If `$value` is undef, then it would normally be treated like a zero-length string. By setting the argument 'passthrough\_undef' to 1, you can allow it to pass thorugh and return undef. This function can expand a small set of commonly used prefixes. For example: ``` $node = rdf_node('foaf:primaryTopic'); ``` The hashref `$RDF::TrineShortcuts::Namespaces` is consulted for expansions. This function is not exported by default, but can be exported using the tag ':nodes' or ':all'. ``` use RDF::TrineShortcuts qw(:default :nodes); ``` - "rdf\_literal($value, %args)", "rdf\_blank($value, %args)", "rdf\_resource($value, %args)", "rdf\_variable($value, %args)" Shortcuts for rdf\_node($value, type=\>'literal', `%args`) and so on. The rdf\_resource function will create a blank node resource if `$value` begins '\_:'. These functions are not exported by default, but can be exported using the tag ':nodes' or ':all'. ``` use RDF::TrineShortcuts qw(:all); ``` - "rdf\_statement($s, $p, $o, \[$g\])", "rdf\_statement($ntriple, \[$g\])" Returns an RDF::Trine::Statement. Parameters `$s`, `$p`, `$o` and `$g` can each be either a plain string that could be passed to rdf\_node, or an arrayref of rdf\_node parameters, or an RDF::Trine::Node. `$ntriple` is a single N-Triples statement. This function is not exported by default, but can be exported using the tag ':all'. ``` use RDF::TrineShortcuts qw(:all); ``` - "flatten\_node($node)" Converts a node back to a string. By default, blank nodes and variables are stringified to their N-Triples and SPARQL representations; URIs are stringified without angled bracket delimiters; and literals to their literal values. Various options are available: 'resource\_as', 'blank\_as', 'variable\_as' and 'literal\_as' can each be set to 'ntriples', 'value' or 'default'. ``` print flatten_node($my_resource, resource_as=>'ntriples'); ``` This function is not exported by default, but can be exported using the tag ':flatten' or ':all'. ``` use RDF::TrineShortcuts qw(:default :flatten); ``` - "flatten\_iterator($iter)" Converts an iterator to a Perl list. In list context returns a list; in scalar context returns an arrayref instead. Each item in the list is, in the case of a bindings iterator, a hashref; or, in the case of a triple/quad iterator, an arrayref \[s, p, o, g\]. For boolean iterators, `$iter`-\>get\_boolean is returned. The nodes which are values in the hashref/arrayref are flattened with flatten\_node, unless flatten\_iterator is called with 'keep\_nodes'=\>1. ``` my @results = flatten_iterator($iter, keep_nodes=>1); ``` You can pass additional options for flatten\_node too: ``` my @results = flatten_iterator($iter, resource_as=>'ntriples'); ``` This function is not exported by default, but can be exported using the tag ':flatten' or ':all'. ``` use RDF::TrineShortcuts qw(:default :flatten); ``` ### Object-Oriented Interface RDF::TrineShortcuts has an alternative, object-oriented interface, not enabled by default. ``` use RDF::TrineShortcuts -methods; my $model = RDF::Trine::Model->temporary_model; # Alias for rdf_parse($some_turtle, model=>$model) $model->parse($some_turtle); # Alias for rdf_string($model, 'rdfxml'); my $rdfxml = $model->parse('rdfxml'); print $rdfxml; my $query = 'SELECT ?name WHERE { ?id foaf:name ?name . } '; # Alias for rdf_query($query, $model); my $result = $model->sparql($query); # Alias for flatten_iterator(); my @result = $result->flatten; ``` And so on. The following methods are set up: - • RDF::Trine::Model: `"parse"`, `"string"`, `"sparql"`. - • RDF::Trine::Node: `"flatten"`. - • RDF::Trine::Iterator: `"flatten"`. - • URI::http: `"sparql"`. - • URI: `"resource"`. Future versions of the RDF::Trine and URI packages may break this. It's a pretty dodgy feature. You can load the normal RDF::TrineShortcuts function-based interface in addition to the object-oriented interface like this: ``` use RDF::TrineShortcuts qw(:default -methods); ``` Or everything: ``` use RDF::TrineShortcuts qw(:all -methods); ``` ## BUGS Please report any bugs to \. ## SEE ALSO RDF::Trine, RDF::Query, RDF::Query::Client. \. This module is distributed with three command-line RDF tools. trapper is an RDF fetcher/parser/serialiser; toquet is a SPARQL query tool; trist is an RDF statistics tool. ## AUTHOR Toby Inkster \. ## COPYRIGHT Copyright 2010 Toby Inkster This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.