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