Skip to content

GreenfishK/BEAR

master
Switch branches/tags
Code
This branch is 224 commits ahead of rdfostrich:master.
Contribute

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
src
 
 
 
 
 
 
 
 

Reproduce experiment

Follow the instructions below to reproduce this experiment.

Install docker

If you have docker installed already, continue with Build docker images install docker on Ubuntu or other OS get access as a non-root user. Find the commands from that page bellow (07.12.2021).

(sudo groupadd docker)
sudo usermod -aG docker $USER 
newgrp docker
docker run hello-world

Build docker images

Go to BEAR/src/rdfstarArchive and build the java rdfstoreQuery project with docker. The docker file uses a maven image to build and package the project with dependencies:

docker build -t bear-rdfstarstores .

Go to the built jar, which should be in /var/lib/docker/overlay2/<latest_build_id>/diff/target/META-INF/services/org.eclipse.rdf4j.rio.RDFParserFactory and add following line to the file: org.eclipse.rdf4j.rio.nquads.NQuadsParserFactory
The reason is that we have two dependencies - rdf4j-rio-turtle and rdf4j-rio-nquads - which are both implementations of the same interface. That is why the RDFParserFactory gets overriden with whichever is listed first in the dependencies in pom.xml (df4j-rio-turtle in our case).

Troubleshoot

Error1: β€œDocker does not have a release file”

Fix: Edit etc/apt/source.list.d/docker.list and set the release version to an Ubuntu version for which there is a docker release, e.g. β€œfocal”: https://stackoverflow.com/questions/41133455/docker-repository-does-not-have-a-release-file-on-running-apt-get-update-on-ubun

Get data

Create the local data directories for this experiment. Download the datasets & queries and either build the RDF* dataset from the BEAR-B ICs and change sets or use our two pre-computed RDF* datasets. See here.

Run the evaluation script

Run the built jar file via docker using our bash script to evaluate data ingestion and query performance with Jena and GraphDB for the three timestamp-based archiving policies (tb (=named graphs), tb_rdf_star_h, tb_rdf_star_f) and materialization queries.

Plot performance measurements

Use the python script to plot the performance across all versions for different timestamp-based archiving policies, query categories and query sets.

Contact

filip.kovacevic@tuwien.ac.at

Original OSTRICH README

see README_orig.txt

About

πŸ‹ Fork of the BEAR benchmark with additional evaluations: https://github.com/webdata/BEAR

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Jupyter Notebook 55.5%
  • Java 33.9%
  • C++ 6.5%
  • Python 2.6%
  • Shell 1.5%