Skip to content

Latest commit

 

History

History
95 lines (59 loc) · 2.46 KB

README.md

File metadata and controls

95 lines (59 loc) · 2.46 KB
Raw

BOUNCER

Privacy-aware Query Processing Over Federations of RDF Datasets

Installing BOUNCER

BOUNCER runs on Debian GNU/Linux and OS X and Python 3.x

  1. Download BOUNCER Clone using git:

    $ git clone https://github.com/SDM-TIB/BOUNCER.git

  2. Go to BOUNCER folder:

    $ cd BOUNCER

  3. Run:

    pip install -r requirements.txt

  4. Install BOUNCER:

    python setup.py install

Configure BOUNCER

  1. Create endpoints list endpoints.txt

    Example:

 http://biotea.linkeddata.es/sparql
 http://colil.dbcls.jp/sparql

  1. Run RDF molecule template extractor in scripts folder:

    scripts$ python3.5 create_rdfmts.py -s endpoints.txt -o 'templates/mytemplates.json'

  2. Create configuration file, config.json in config folder:

    Example:

  {
  "MoleculeTemplates": [
    {
       "type": "filepath",
       "path":"templates/mytemplates.json"
         }
      ]
   }
  1. Now BOUNCER is ready to "investigate" :)

About supported endpoints

BOUNCER currently supports endpoints that answer queries either on JSON. Expect hard failures if you intend to use BOUNCER on endpoints that answer in any other format.

Running BOUNCER

Once you installed BOUNCER and the Molecule Templates are ready with config.json, you can start running BOUNCER using the following script:

$ python3.5 test_bouncer.py -p <planonly> -q <query> -c <path/to/config.json> -s <isstring>

where:

  • <query>: - SPARQL QUERY
  • <path/to/config.json>: - path to configuration file
  • <isstring>: - (Optional) set if is sent as string: available values 1 or -1. -1 is default, meaning query is from file
  • <planonly>: - (Optional) if set True, then only execution plan is generated and showed. If False (default), then the generated plan will be executed, too.

Running experiments:

$./runQueries.sh <path/to/queries-dir> <path/to/config.json> <path/to/results-folder> errors.txt <planonlyTorF> &

OR

$ python3.5 start_experiment.py -c <path/to/config.json> -q <query-file> -r <path/to/results-folder> -t 'MULDER' -s True -p <planonly>

References

Endris, Kemele & Almhithawi, Zuhair & Lytra, Ioanna & Vidal, Maria-Esther & Auer, Sören. "BOUNCER: Privacy-aware Query Processing Over Federations of RDF Datasets", (To appear)In International Conference on Database and Expert Systems Applications, DEXA 2018.