neo4j-sparql-extension

Neo4j unmanaged extension for RDF storage and SPARQL 1.1 query features.

Installation

Download the latest release from the releases page and place it inside the /plugins/ directory of the Neo4j server installation.

To enable the extension add it to the org.neo4j.server.thirdparty_jaxrs_classes key in the /conf/neo4j-server.properties file. For example:

org.neo4j.server.thirdparty_jaxrs_classes=de.unikiel.inf.comsys.neo4j=/rdf

The RDF/SPARQL extension is then avaiable as /rdf resource on the Neo4j server.

Please note that if there is any data in the database that was not imported using the /rdf/graph resource the plugin might crash, because the plugin expects the data to be stored in a special way to support RDF storage in Neo4j.

SPARQL Protocol (SPARQL 1.1 Queries)

Base resource: /rdf/query

Use this resource to execute SPARQL queries.

$ curl -v -X POST localhost:7474/rdf/query \
       -H "Content-Type: application/sparql-query" \
       -H "Accept: application/sparql-results+json" \
       -d "SELECT ?s ?p ?o WHERE { ?s ?p ?o } LIMIT 5"

See SPARQL 1.1 Protocol "2.1 query operation".

SPARQL Graph Protocol

Base resource: /rdf/graph

Use this resource to add, replace or delete RDF data.

$ curl -v -X PUT \
       localhost:7474/rdf/graph \
       -H "Content-Type:text/turtle" --data-binary @data.ttl

$ curl -v localhost:7474/rdf/graph

See SPARQL 1.1 Graph Store HTTP Protocol "5 Graph Management Operations".

SPARQL Protocol (SPARQL 1.1 Update Queries)

Base resource: /rdf/update

Use this resource to execute SPARQL update queries.

$ curl -v -X POST localhost:7474/rdf/query \
       -H "Content-Type: application/sparql-update" \
       -d "@prefix dc: <http://purl.org/dc/elements/1.1/> . \
           @prefix ns: <http://example.org/ns#> . \
           <http://example/book1> ns:price 42 ."

See SPARQL 1.1 Protocol "2.1 update operation".

OWL-2 Inference

The plugin supports (limited) OWL-2 reasoning using query rewriting of SPARQL algebra expressions. For a list of supported axioms, see the inference wiki page.

To use inference the TBox must be uploaded to the special graph urn:sparqlextension:tbox:

$ curl -v -X PUT \
       localhost:7474/rdf/graph\?graph=urn%3Asparqlextension%3Atbox \
       -H "Content-Type:text/turtle" --data-binary @tbox.ttl

Now it is possible to send SPARQL queries that additionally return inferrend solutions. There are two ways to enable inference:

Using a Query Parameter

Just send your SPARQL query to /rdf/query and add a query parameter inference=true:

$ curl -v -X POST localhost:7474/rdf/query\?inference=true \
       -H "Content-Type: application/sparql-query" \
       -H "Accept: application/sparql-results+json" \
       -d "SELECT ?s ?p ?o WHERE { ?s ?p ?o } LIMIT 5"

Using the Inference Resource

Send your SPARQL query to /rdf/query/inference:

$ curl -v -X POST localhost:7474/rdf/query/inference \
       -H "Content-Type: application/sparql-query" \
       -H "Accept: application/sparql-results+json" \
       -d "SELECT ?s ?p ?o WHERE { ?s ?p ?o } LIMIT 5"

Chunked Imports

If you want to import a large amount of RDF data, you should enable the chunked import. The import will be split into smaller chunks and each chunk will be committed seperately to the database. To enable a chunked import set the query parameter chunked=true when using the /rdf/graph resource.

Configuration

To change the configuration add a sparql-extension.properties file in the /conf folder of the Neo4j server installation.

The default configuration is as follows:

de.unikiel.inf.comsys.neo4j.query.timeout = 120
de.unikiel.inf.comsys.neo4j.query.patterns = p,c,pc
de.unikiel.inf.comsys.neo4j.inference.graph = urn:sparqlextension:tbox
de.unikiel.inf.comsys.neo4j.chunksize = 1000

License

GPLv3