Spark Usage

Javadocs:

• spark-api
• spark-spi
• spark-protocol

A simple example

This is an example using the SPARQL protocol implementation (spark-protocol) of the Spark API (spark-api) showcasing the normal flow of execution (sans error handling):

DataSource myDS = new ProtocolDataSource("http://DBpedia.org/sparql");
Connection conn = myDS.getConnection(NoCredentials.INSTANCE); 
Command query = conn.createCommand("SELECT ?p ?o WHERE { ?s ?p ?o }"); 
Solutions solutions = query.executeQuery(); 
System.out.println("vars = " + solutions.getVariables()); 
int row = 0; 
for(Map solution : solutions) { 
  System.out.println("Row " + (row++) + ": " + solution); 
} 
solutions.close(); 
query.close(); 
conn.close();

Very similar to JDBC, a DataSource is used to represent a remote SPARQL processor. Connections are created by the DataSource. Commands representing a SPARQL execcution are created from the Connection. Commands when executed return a Result, most often a Solutions object, which is an iterable cursor over the series of solutions.

API overview

This is an overview of the main packages and classes in the Spark API:

spark.api:

• DataSource - the provider implements this class and provides JavaBean style getters and setters to configure the Datasource. A factory for Connections.
• Connection - a Connection to a remote SPARQL processor (the underlying implementation may not necessarily maintain a persistent connection). A factory for Commands and access to the ServiceDescription for the Connection.
• Command - a particular SPARQL command for execution. Executing it produces a Result.
• Result - the super-interface for any kind of result from a command execution.
• Cursored Result - a Result that represents a sequence of result items. The cursor can only be traversed as an iterator from beginning to end in one pass.
• Solutions - a cursored series of solutions from a SPARQL SELECT query

spark.api.rdf:

• RDFNode - super-interface of all RDF data nodes
• Resource - super-interface of all RDF resources
• NamedNode - represents a named resource (usually represented by an IRI or URI)
• BlankNode - represents an anonymous resource, which may have a label
• Literal - super-interface of all RDF literal values. All literal values have a lexical representation.
• PlainLiteral - an untyped plain literal with an optional language tag
• TypedLiteral - a typed literal with an XSD type
• Triple - an RDF Triple consisting of a subject, predicate, and object