This project contains the code of the simulation study for the Automated Systematic Review project. It contains code to run batches of simulation runs in parallel using MPI.
We make use of the SURFSara HPC infrastructure. But with some modifications, the code can be run on other HPC facilities as well.
The Automated Systematic Review project requires Python 3.6+. To run the code you also need an implementation of the MPI standard. The most well known standard is OpenMPI. This is not a python package and should be installed separately.
The simulation project itself can be directly installed with:
pip install --user git+https://github.com/msdslab/automated-systematic-review-simulationsDependencies are automatically installed.
To run a batch of simulations on 4 cores and 12 runs, use the following command:
mpirun -n 4 asreview batch ${DATA_SET} --log_file ${DIR}/results.log --n_runs 12It will create 12 files in the ${DIR} directory, while running on 4 cores in parallel.
-
asreview-visualization Package for visualization of log files.
-
asreview-hyperopt Package for optimizing the parameters for the