Installation¶

Using the conda package¶

foamalgo¶

conda install -c conda-forge foamalgo

pyfoamalgo¶

conda install -c conda-forge pyfoamalgo

From source¶

foamalgo requires a modern C++ compiler which supports C++17.

conda env update -f environment-dev.yml
git clone https://github.com/zhujun98/foamalgo.git
cd foamalgo

foamalgo¶

mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=your_install_prefix
make install

pyfoamalgo¶

export CC=gcc-10 CXX=g++-10  # Specify the compiler in your system which supports C++20
export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
pip install . -v

Bugs and workarounds¶

Install h5py on MacOS with m1 chip.

brew install hdf5
export HDF5_DIR=/opt/homebrew/Cellar/hdf5/<version>
pip install --no-binary=h5py h5py

On the Maxwell cluster¶

module load anaconda3

# It is highly recommended to create an independent environment.
conda create -n foamalgo python=3.9.12
conda activate foamalgo

conda install -c conda-forge pyfoamalgo
conda install jupyter
python -m ipykernel install --user --name foamalgo --display-name "foamalgo"

Now, you should be able to start a Jupyter notebook via https://max-jhub.desy.de/ and select the newly created kernel.