Functions¶

"iri", "literal", "blank", "variable"

As per the similarly named functions exported by RDF::Trine itself.

These are wrapped with a very tiny bit of DWIMmery. A blessed URI object passed to "iri" will be handled properly; a blessed URI object passed to "literal" will default the datatype to xsd:anyURI. A string starting with "_:" passed to either "iri" or "blank" will correctly create a blank node. A string starting with "?" passed to either "blank" or "variable" will correctly create a variable. If any of them are passed an existing RDF::Trine::Node, it will be passed through untouched.

Other than that, no magic.

"curie"

Like "iri" but passes strings through RDF::NS::Trine.

statement(@nodes)

As per the similarly named function exported by RDF::Trine itself.

Again, a tiny bit of DWIMmery: blessed URI objects are passed through "iri" and unblessed scalars (i.e. strings) are assumed to be literals.

"store"

As per the similarly named function exported by RDF::Trine itself.

"model"

Returns a new RDF::Trine::Model. May be passed a store as a parameter.

"parse($source, %options)"

Parses the source and returns an RDF::Trine::Model. The source may be:

  my $model = parse('/tmp/data.ttl', as => 'Turtle');
  my $data   = iri('http://example.com/data.nt');
  my $parser = RDF::Trine::Parser::NTriples->new;
  my $model  = model();
  
  parse($data, using => $parser, into => $model);

"serialize($data, %options)"

Serializes the data (which can be an RDF::Trine::Model or an RDF::Trine::Iterator) and returns it as a string.

The "serializer" option can be used to provide a blessed RDF::Trine::Serializer object to use; the "type" option can be used instead to provide a type hint. The "output" option can be used to provide a filehandle, IO::All, Path::Class::File or file name to write to instead of returning the results as a string.

"to" and "file" are aliases for "output"; "type", "using" and "as" are aliases for "serializer".

Examples:

  print serialize($model, as => 'Turtle');
  my $file = Path::Class::File->new('/tmp/data.nt');
  serialize($iterator, to => $file, as => 'NTriples');
    

Array References¶

In addition to the above interface, each function supports being called with a single arrayref argument. In those cases, the arrayref is dereferenced into an array, and treated as a list of arguments. That is, the following are equivalent:

  foo($bar, $baz);
  foo([$bar, $baz]);

This is handy if you're writing a module of your own and wish to accept some RDF data:

  sub my_method {
    my ($self, $rdf, $foo) = @_;
    $rdf = parse($rdf);
    
    ....
  }

Your method can now be called like this:

  $object->my_method($model, 'foo');
  
  $object->my_method($url, 'foo');
  
  $object->my_method(
      [ $filehandle, as => 'Turtle', base => $uri ],
      'foo',
  );

Export¶

By default, nothing is exported. You need to request things:

  use RDF::TrineX::Functions qw< iri literal blank statement model >;

Thanks to Sub::Exporter, you can rename functions:

  use RDF::TrineX::Functions
    qw< literal statement model >,
    blank => { -as => 'bnode' },
    iri   => { -as => 'resource' };

If you want to export everything, you can do:

  use RDF::TrineX::Functions -all;

To export just the functions which generate RDF::Trine::Node objects:

  use RDF::TrineX::Functions -nodes;

Or maybe even:

  use RDF::TrineX::Functions -nodes => { -suffix => '_node' };

If you want to export something roughly compatible with the old RDF::TrineShortcuts, then there's:

  use RDF::TrineX::Functions -shortcuts;

When exporting the "serialize" function you may set a default format:

  use RDF::TrineX::Functions
      serialize => { -type => 'NTriples' };

This will be used when "serialize" is called with no explicit type given.

Pseudo-OO interface¶

"new"

This acts as a constructor, returning a new RDF::TrineX::Functions object.

All the normal functions can be called as methods:

 my $R = RDF::TrineX::Functions->new;
 my $model = $R->model;

There's no real advantage to using this module as an object, but it can help you avoid namespace pollution.