Note: All of these instructions are intended for use on a macOS system (or other Linux-based system).
The following packages (and associated versions) have been verified to work with the codebase. It is recommended to use pip to install all Python packages within the roborobo conda environment.
- conda 4.11.0
- roborobo 4.0
- python 3.9.7
- numpy 1.21.5
- scipy 1.7.3
- deap 1.3.3
- qdpy 0.1.2.1
- torch 1.12.1
- scikit-learn 1.1.3
- seaborn 0.12.2
- matplotlib 3.6.2
Once you have set up your roborobo conda environment and installed the necessary direct package dependencies, you can verify that environment contains all the correct package versions by referring to list below. This is obtained by running the first two commands, i.e., conda activate roborobo followed by conda list.
% conda activate roborobo
(roborobo) % conda list
# packages in environment at /Users/scott/opt/anaconda3/envs/roborobo:
#
# Name Version Build Channel
alabaster 0.7.12 pyhd3eb1b0_0
babel 2.9.1 pyhd3eb1b0_0
blas 1.0 openblas
brotlipy 0.7.0 py39h9ed2024_1003
ca-certificates 2022.10.11 hecd8cb5_0
certifi 2022.9.24 py39hecd8cb5_0
cffi 1.14.6 py39h2125817_0
charset-normalizer 2.0.4 pyhd3eb1b0_0
colorama 0.4.4 pyhd3eb1b0_0
commonmark 0.9.1 pyhd3eb1b0_0
contourpy 1.0.6 pypi_0 pypi
cryptography 36.0.0 py39hf6deb26_0
cycler 0.11.0 pypi_0 pypi
deap 1.3.3 pypi_0 pypi
docutils 0.17.1 py39hecd8cb5_1
fonttools 4.38.0 pypi_0 pypi
future 0.18.2 py39hecd8cb5_1
greenlet 2.0.1 pypi_0 pypi
idna 3.3 pyhd3eb1b0_0
imagesize 1.3.0 pyhd3eb1b0_0
intel-openmp 2021.4.0 hecd8cb5_3538
jinja2 3.0.2 pyhd3eb1b0_0
joblib 1.1.1 py39hecd8cb5_0
kiwisolver 1.4.4 pypi_0 pypi
libcxx 12.0.0 h2f01273_0
libffi 3.3 hb1e8313_2
libgfortran 3.0.1 h93005f0_2 anaconda
libopenblas 0.3.20 h9a5756b_0
llvm-openmp 14.0.6 h0dcd299_0
markupsafe 2.0.1 py39h9ed2024_0
matplotlib 3.6.2 pypi_0 pypi
mkl 2021.4.0 hecd8cb5_637
mkl-service 2.4.0 py39h9ed2024_0
ncurses 6.3 hca72f7f_2
nomkl 3.0 0
numpy 1.21.5 py39h0f1bd0b_3
numpy-base 1.21.5 py39hbda7086_3
numpydoc 1.1.0 pyhd3eb1b0_1
openssl 1.1.1s hca72f7f_0
packaging 21.3 pyhd3eb1b0_0
pandas 1.5.2 pypi_0 pypi
pillow 9.2.0 pypi_0 pypi
pip 21.2.4 py39hecd8cb5_0
psutil 5.9.4 pypi_0 pypi
pybind11 2.8.1 py39hf018cea_1 conda-forge
pybind11-global 2.8.1 py39hf018cea_1 conda-forge
pycparser 2.21 pyhd3eb1b0_0
pygments 2.10.0 pyhd3eb1b0_0
pyopenssl 21.0.0 pyhd3eb1b0_1
pyparsing 3.0.4 pyhd3eb1b0_0
pysocks 1.7.1 py39hecd8cb5_0
python 3.9.7 h88f2d9e_1
python-dateutil 2.8.2 pypi_0 pypi
python_abi 3.9 2_cp39 conda-forge
pytz 2021.3 pyhd3eb1b0_0
pyyaml 6.0 pypi_0 pypi
pyzmq 24.0.1 pypi_0 pypi
qdpy 0.1.2.1 pypi_0 pypi
readline 8.1 h9ed2024_0
recommonmark 0.6.0 pyhd3eb1b0_0
requests 2.26.0 pyhd3eb1b0_0
roborobo 4.0.0 pypi_0 pypi
scikit-learn 1.1.3 py39he9d5cce_0
scipy 1.7.3 py39hfb86763_0
scoop 0.7.2.0 pypi_0 pypi
seaborn 0.12.2 pypi_0 pypi
setuptools 58.0.4 py39hecd8cb5_0
six 1.16.0 pyhd3eb1b0_0
sklearn 0.0.post1 pypi_0 pypi
snowballstemmer 2.2.0 pyhd3eb1b0_0
sphinx 4.2.0 pyhd3eb1b0_1
sphinx_rtd_theme 0.4.3 pyhd3eb1b0_0
sphinxcontrib-applehelp 1.0.2 pyhd3eb1b0_0
sphinxcontrib-devhelp 1.0.2 pyhd3eb1b0_0
sphinxcontrib-htmlhelp 2.0.0 pyhd3eb1b0_0
sphinxcontrib-jsmath 1.0.1 pyhd3eb1b0_0
sphinxcontrib-qthelp 1.0.3 pyhd3eb1b0_0
sphinxcontrib-serializinghtml 1.1.5 pyhd3eb1b0_0
sqlite 3.37.0 h707629a_0
threadpoolctl 2.2.0 pyh0d69192_0
tk 8.6.11 h7bc2e8c_0
torch 1.12.1 pypi_0 pypi
torchaudio 0.12.1 pypi_0 pypi
torchvision 0.13.1 pypi_0 pypi
typing-extensions 4.3.0 pypi_0 pypi
tzdata 2021e hda174b7_0
urllib3 1.26.7 pyhd3eb1b0_0
wheel 0.37.0 pyhd3eb1b0_1
xz 5.2.5 h1de35cc_0
zlib 1.2.11 h4dc903c_4
Once you have activated the roborobo conda environment (by running conda activate roborobo) and are in the root directory of this repository, there are number of relevant commands for running the extended simulator:
| Action | Command |
|---|---|
| Start Evolution | python run.py -s <config file> <run id> |
| Resume Evolution | python run.py -r <checkpoint file> |
| Export Results | python run.py -e <checkpoint file> |
| Aggregate Archives | python run.py -a <aggregate prefix> <generation> |
| Plot Figures | python run.py -p <graph type> [variant options] |
| View Simulation | python run.py -v <checkpoint file> |
Note: The command to plot figures expects a specific naming format for run IDs since the aggregate prefixes have been hardcoded (e.g. "shom-e", "shet-m", "mhom-d", "ashet-e", etc.). If you want to use a different naming format, please update the relevant process/plot_*.py files.
Note: It is recommended that you set pDynamicProgressOutput=true in your corresponding config/*.properties while running evolution locally. This option dynamically outputs the progress for each generation and gives an estimate of time remaining. However, you will need to set this option to false again when running in a cluster environment.
Note: All of these instructions are intended for use on a Linux system that uses the PBS queuing system.
Start by logging into the relevant cluster account being used to run simulations. Here is an example of how to SSH onto the CHPC cluster:
ssh username@lengau.chpc.ac.za
First remove any conda config from .bashrc (normally at end of file).
The installation process will probably differ between cluster envirnoments, but here is how it is done on the CHPC:
module purge
module load chpc/BIOMODULES anaconda/3
Verify that both Python and Conda are installed correctly:
python --version
conda list
Refresh the shell environment by executing the following command and then exiting and logging in again:
conda init bash
If roborobo4 is not installed, you need to activate the necessary module. This is how it is done on the CHPC:
module purge
module load chpc/BIOMODULES anaconda/3
module load chpc/BIOMODULES roborobo4
Run the following command if in PBS:
eval "$(conda shell.bash hook)"
Verify that the conda, python and roborobo modules are all installed correctly:
which conda # expected: /apps/chpc/bio/anaconda3-2020.02/bin/conda
conda activate roborobo
which python # expected: /apps/chpc/bio/anaconda3-2020.02/envs/roborobo/bin/python
python -c 'import pyroborobo'
Make sure the roborobo conda environment is activated and then install the necessary Python packages as follows:
pip install deap
pip install qdpy
pip install torch
pip install scikit-learn
pip install seaborn
From a new terminal session (i.e. a session that is not logged into the cluster via SSH), upload the source code as follows:
scp -r "/Users/user/Local/GitHub Projects/username/msc-project" username@lengau.chpc.ac.za:~/lustre/msc-project
After having defined the necessary job script (see under jobs/ for examples), you will need to submit a request to add your simulation job to the queue.
Make sure you are SSH'ed into the cluster enviroment and located at the root directory of this repository, then run the following command if in PBS:
qsub jobs/default.job