Please note that unfortunately - like the provided ATL solution - the solutions cannot be started directly using MS Windows, since they use cd ../dir && ./gradlew
. If only Windows is available, these commands have to be adapted.
The solution can be found at solutions/RelationalRAGSolution. All variants described in the paper are started from separate subfolders with a solution-specific suffix.
If there are problems, please submit an issue or write to johannes.mey@tu-dresden.de
The solution can be found at solutions/RSYNC. All variants described in the paper are started from separate subfolders with a solution-specific suffix.
If there are problems, please submit an issue or write to christopher.werner@tu-dresden.de
The relational RAG solutions use Gradle as build tool. Thus there are no extra prerequisites.
However, in order to run the RSYNC solutions, the following tools are necessary:
The scripts
directory contains the run.py
script.
At a first glance, invoke it without any arguments so that the solution will be built, benchmarked, running times visualized and the results compared to the reference solution's.
One might fine tune the script for the following purposes:
run.py -b
-- builds the projectsrun.py -b -s
-- builds the projects without testingrun.py -g
-- generates the instance modelsrun.py -m
-- run the benchmark without buildingrun.py -v
-- visualizes the results of the latest benchmarkrun.py -e
-- compare results to the reference output. The benchmark shall already been executed using -m
.run.py -m -e
-- run benchmark without building, then extract and compare results to the reference outputrun.py -t
-- build the project and run tests (usually unit tests as defined for the given solution)The config
directory contains the configuration for the scripts:
config.json
-- configuration for the model generation and the benchmark
-m
for run.py
), so it is not applied to e.g. the build phase (see -b
for run.py
).The script runs the benchmark for the given number of runs, for the specified tools and input models.
The benchmark results are stored in a CSV file. The header for the CSV file is stored in the output/header.csv
file.
To implement a tool, you need to create a new directory in the solutions directory and give it a suitable name.