tpellet
commented
1 year ago Bootstrap: docker
From: ubuntu:20.04
%post
# Set noninteractive frontend for package installation
export DEBIAN_FRONTEND=noninteractive
# Set default timezone to avoid getting stuck during tzdata installation
echo "Etc/UTC" > /etc/timezone
# Update package lists and upgrade installed packages
apt-get update
# Install necessary tools and libraries
apt-get install -y wget build-essential gcc-10 g++-10 gfortran-10 \
libopenmpi-dev python3.8 python3-pip nano openjdk-17-jdk
# Set the default compiler to GCC 10
update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-10 100 --slave /usr/bin/g++ g++ /usr/bin/g++-10
# Download and extract OpenMPI 4.1.1
wget https://download.open-mpi.org/release/open-mpi/v4.1/openmpi-4.1.1.tar.gz
tar -xvzf openmpi-4.1.1.tar.gz
cd openmpi-4.1.1
# Configure and build OpenMPI with GCC 10.2.0
./configure --prefix=/usr
make -j$(nproc)
make install
# Clean up temporary files
cd ..
rm -r openmpi-4.1.1 openmpi-4.1.1.tar.gz
# Install NetLogo
wget https://ccl.northwestern.edu/netlogo/6.3.0/NetLogo-6.3.0-64.tgz
tar -xvzf NetLogo-6.3.0-64.tgz
mv NetLogo\ 6.3.0 /opt/NetLogo
rm NetLogo-6.3.0-64.tgz
# Install Python packages
pip3 install numpy mpi4py pynetlogo scikit-learn pandas statsmodels matplotlib distributed scipy dask futures
# Set python3.8 as the default python
update-alternatives --install /usr/bin/python python /usr/bin/python3.8 1
# Create a scripts directory inside the container
mkdir -p /opt/scripts
# Create mpi_test.py file in the scripts directory
cat > /opt/scripts/mpi_test.py << 'EOM'
#!/usr/bin/env python3
from mpi4py import MPI
import sys
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
size = comm.Get_size()
if rank == 0:
sys.stdout.write(f'Hello from the master (process {rank} of {size})\n')
else:
sys.stdout.write(f'Hello from a worker (process {rank} of {size})\n')
EOM
# Make the MPI test script executable and copy it to /usr/bin
chmod +x /opt/scripts/mpi_test.py
cp /opt/scripts/mpi_test.py /usr/bin/
%environment
export LD_LIBRARY_PATH="/usr/lib:$LD_LIBRARY_PATH"
export PATH="/opt:$PATH"
export PATH="/opt/NetLogo:$PATH"
export JAVA_HOME="/usr/lib/jvm/java-17-openjdk-amd64"
quaquel
Hi!
I have been trying to use PyNetLogo inside a container because my university HPC does not have NetLogo 6.3 installed. I have not been able to load a model successfully so far.
I am on a Ubuntu 20.04 container and whether I use the java jvm in the NetLogo 6.3 runtime folder or I install a separate java (13/17) using apt-get install, I still get a problem when loading the model.
Here is the error message (I have attached the singularity build for reference):
I have been trying to debug it by installing different versions of Java (13/17) without success. I have tried changing the jvm path to the NetLogo folder '/opt/NetLogo/lib/runtime/lib/client/' without success. Same with setting JAVA_Home to the NetLogo runtime as opposed to the openJDK version.
When I do it on my windows computer simply setting JAVA_HOME to my NetLogo runtime folder it does work.