VIG is a data scaler specifically designed for benchmarks of Ontology-based Data Access (OBDA) systems such as Ontop. VIG takes as input a source database instance and a scale factor, and produces a scaled database instance that satisfies the schema constraints and that is “similar” to the source instance according to certain ad-hoc similarity measures. The produced database instance is in form of csv files that can be effectively imported into any relational database management system (RDBMS).
VIG is currently the official data scaler of the NPD benchmark.
The version in the develop branch is very close to be released. This version solves several critical issues that are currently present in the latest release (master branch). We strongly suggest to use the version in develop rather than the latest release. In case of problems, please contact the authors.
VIG is bundled as a maven project, therefore it can be built using the standard maven commands. We provide a bash script to save time:
$ bash build.sh
The jar (with dependencies) containing the application will be generated under the vig-distribution/target folder.
Before running, a few things need to be configured. Please refer to the documentation (http://ontop.github.io/vig/ or docs folder) for more information.
The following list contains a few publications describing VIG. We suggest to skim through them, so as to save time and understand exactly how VIG works and what data it can generate for you.
- Longest Technical Description and Evaluation of VIG (Submission to SWJ)
- Long Technical Description and Evaluation of VIG (@BLINK '16)
- Short Technical Description of VIG
- Evaluation of VIG with the BSBM Benchmark (@ISWC Posters '16)
Checkout to the evaluations/results branch. Evaluations will be in the "evaluations" folder.