A user-friendly Python package to manipulate input and output files of Large-scale Atomic/Molecular Massively Parallel Simulator LAMMPS molecular dynamics code. One Python class per LAMMPS file type. Include conversion into pandas.DataFrame objects.

Create and activate a virtual Python environment (recommended) with the name easylammps, for example using Anaconda package and environment manager :

conda create --name easylammps
conda activate easylammps

Clone this repository :

git clone https://github.com/kkempfer/easylammps.git

Install easylammps and its dependencies :

cd easylammps
pip install .

Required :

• networkx Data structures for graphs and graph algorithms
• numpy N-dimensional array
• pandas Labeled data analysis

Recommended :

• matplotlib Visualization

Enter to the documentation directory :

cd docs

Create and activate the virtual Python environment easylammps-dev based on the environment-dev.yml file :

conda env create -f environment-dev.yml
conda activate easylammps-dev

Make the documentation using sphinx :

make html

Deactivate your virtual Python environment :

conda deactivate

Access the documentation in the easylammps/docs/build/html directory.

Coupling Python with LAMMPS opens the door to many advanced extensions. Fortunately, the lammps Python library already wraps the LAMMPS C-library interface. Here, we propose a quick installation guide.

Clone the official LAMMPS repository (stable release) :

git clone https://github.com/lammps/lammps.git
cd lammps
git checkout stable
git pull

Use the virtual Python environment where easylammps is installed (recommended) :

conda activate easylammps

Prepare the building directory and run cmake with at least these options (replace pythonX.Y by your python version) :

mkdir build-python
cd build-python
cmake -D BUILD_SHARED_LIBS=ON \
      -D LAMMPS_EXCEPTIONS=ON \
      -D PKG_PYTHON=ON \
      -D CMAKE_INSTALL_PREFIX=$CONDA_PREFIX \
      -D PYTHON_EXECUTABLE=$CONDA_PREFIX/bin/python \
      -D PYTHON_INCLUDE_DIR=$CONDA_PREFIX/include/pythonX.Y \
      -D PYTHON_LIBRARY=$CONDA_PREFIX/lib/libpythonX.Y.so \
      ../cmake

More options to add in cmake are available here. Among them, some useful ones are presented here :

      -D LAMMPS_MACHINE=python # Suffix to append to lmp binary
      -D PKG_MOLECULE=ON # Model molecular systems with fixed covalent bonds
      -D PKG_RIGID=ON    # Rigid constraints on collections of atoms or particles
      -D PKG_KSPACE=ON   # Long-range electrostatic interaction

Once ready, build and install LAMMPS as a shared library with Python :

make
make install

Finally, we need to add the library path which contains the installed liblammps.so to LD_LIBRARY_PATH, but only when our virtual Python environment easylammps is active. Anaconda provides a way to manage environment variables. The procedure is given below.

Enter to the conda environment directory and create these subdirectories and files :

cd $CONDA_PREFIX
mkdir -p ./etc/conda/activate.d
mkdir -p ./etc/conda/deactivate.d
touch ./etc/conda/activate.d/env_vars.sh
touch ./etc/conda/deactivate.d/env_vars.sh

Edit ./etc/conda/activate.d/env_vars.sh as follows :

#!/bin/sh

export LD_LIBRARY_PATH="$CONDA_PREFIX/lib:$LD_LIBRARY_PATH"

Edit ./etc/conda/deactivate.d/env_vars.sh as follows :

#!/bin/sh

# The first, third and fourth lines are there to arrange for
# every component of the search path to be surrounded by `:`,
# to avoid special-casing the first and last component. The
# second line removes the specified component.

LD_LIBRARY_PATH=:$LD_LIBRARY_PATH:
LD_LIBRARY_PATH=${LD_LIBRARY_PATH//:$CONDA_PREFIX\/lib:/:}
LD_LIBRARY_PATH=${LD_LIBRARY_PATH#:}
LD_LIBRARY_PATH=${LD_LIBRARY_PATH%:}

You should now be able to run LAMMPS from the command line and to import lammps module within Python :

lmp_python
python -c "import lammps"

Do not forget to deactivate your virtual Python environment when you are done working :

conda deactivate

NOTE

For now, installing LAMMPS as a shared library with Python is not mandatory to use the easylammps package. In future, we may add some functionalities using the lammps Python library.

Coming soon !

• Add jupyter notebooks tutorials

• Use of flake8 linter and flake8-docstrings extension to check Python code syntax
• Use of black to auto-format Python code
• Use of pytest for testing
• Use of sphinx to auto-build documentation based on Python docstrings
• Add InputTools object to read LAMMPS input file ?
• Add RestartTools object to read LAMMPS restart file ?

EasyLAMMPS is licensed under the GNU Affero General Public License v3 license. See the LICENSE file for a full description.

I kindly thank Julien Devémy who introduced me to the Python programming language. Part of the code used to write easylammps as been taken and modified from his lammps-tools package.

I gratefully acknowledge Alain Dequidt for his inspiring ideas in scientific computing.

Please send bug reports, ideas and questions on GitHub.