dld logo

Dockerizing Linked Data

Popular knowledge bases that provide SPARQL endpoints for the web are usually experiencing a high number of requests, which often results in low availability of their interfaces. A common approach to counter the availability issue is to run a local mirror of the knowledge base. Running a SPARQL endpoint is currently a complex task which requires a lot of effort and technical support for domain experts who just want to use the SPARQL interface.

With our approach of containerised knowledge base shipping we are introducing a simple to setup methodology for running a local mirror of an RDF knowledge base and SPARQL endpoint with interchangeable exploration components. The flexibility of the presented approach further helps maintaining the publication infrastructure for dataset projects.

Installation

We assume you have git, python (2.7 or 3.x, with pip) and a docker daemon running. For further details please check the requirements section.

$ git clone https://github.com/Dockerizing/dockerizing-bootstrap
$ cd dockerizing-bootstrap
$ git submodule init && git submodule update
$ sudo pip install -Ur requirements.txt

Run

$ cd examples/simple/
$ ../../dld.py
$ cd wd-dld
$ docker-compose up

See How To Use Bootstrap Script for more advanced setup.

Run Query

Now your SPARQL endpoint should be available at http://localhost:8895/sparql.
Try to execute a simple query on your new endpoint: select * from { ?s ?p ?o }

Publication

Knowledge Base Shipping to the Linked Open Data Cloud by Natanael Arndt, Markus Ackermann, Martin Brümmer und Thomas Riechert in 11th International Conference on Semantic Systems Proceedings [Bibsonomy]