The OMPL Planner Arena code allows you to easily create plots from a benchmark database produced from benchmark log files. See also http://plannerarena.org and the benchmarking tutorial.
Dependencies
- R 3.1.0 or higher
- On Ubuntu you need to run
sudo apt install libv8-dev
. - The following R packages: shinyjs, V8, tidyverse, Hmisc, pool, RSQLite, and markdown. These packages can be installed like so:
R -e "install.packages(c('shinyjs', 'V8', 'tidyverse', 'Hmisc', 'pool', 'RSQLite', 'markdown'), repos='https://cran.r-project.org')"
Running
First, produce a series of benchmark log files. Next, produce a database called benchmark.db with ompl_benchmark_statistics.py
. After that, you need to run the script plannerarena
, which launches Planner Arena (if the script is not in your path, it is in the directory ompl/scripts/plannerarena
). After you upload your database via the web page that should have opened in your browser, it should look similar to http://plannerarena.org.
Docker
PlannerArena is also available as a Docker container. Once you have Docker installed, you can download and run PlannerArena like so:
docker pull mmoll/plannerarena docker run --rm -p 8888:8888 plannerarena
Go to http://127.0.0.1:8888 and upload your benchmark data. If your benchmark database files exceed 50MB, set the environment variable OMPL_PLANNERARENA_MAX_DB_SIZE
to the desired limit in bytes in the Dockerfile for Planner Arena and rebuild the docker container.