- The code is written using Java 21.
- Install dependencies and build the project with
mvn install
You can configure the execution using a JSON file.
There are example configurations (in the configs directory) for the three datasets used in the paper.
The most important configuration options are:
filePrefix: the directory where the dataset is located (remember trailing/)dependencies: the dependencies that should be repaired (this is also the order of dependencies forPIRANHA)matchingType: algorithm used to do the repair (one ofEQUALITY, GROUND_TRUTH, PIRANHA, CHROMIDE, GOBIES)distanceMeasure: the distance measure used to create the general metric space (these must be precomputed)computeDistances: boolean to precompute the distanceslinkage: linkage function for the hierarchical clustering to generate the tree metric spacenumThreads: the number of threads used for parallelized parts of the codevisualizeResult: show changes done by the repairverbose: boolean to turn on verbose output