aMOF is a python package consisting in a collection of tools to analyze Molecular Dynamics (MD) trajectories of amorphous Metal-Organic Frameworks (MOFs).
This package brings together a number of analyses that can be performed on every MD trajectory (not necessarily MOFs), heavily using both other python packages and non-python codes. It can compute the following properties:
- Radial Distribution Functions (RDF), based on ASAP
- Bond-Angle Distributions
- Coordination numbers
- Mean Squared Displacement (MSD)
- Elastic constants from cell properties, and mechanical properties from elastic constants by calling ELATE
- Pore analysis by wrapping Zeo++, reusing code from pysimm
- Ring statistics by wrapping the RINGS code
The backend for manipulating trajectories is ASE, which are ASE trajectory objects.
A module called coordination allows the identification of the different building blocks of amorphous MOFs with an ad hoc approach per MOF system.
This allows the computation of rings statistics of the metal-ligand network. Only Zn-based ZIFs are natively supported, although the code can be quickly adapted for similar structures.
This code is designed to be compatible with molsys, and can be used to generate input files in mfpx format.
The detailed algorithm for ZIF-4 is presented in the supporting information of the paper Challenges in Molecular Dynamics of Amorphous ZIFs using Reactive Force Fields.
To use the pore and ring modules, Zeo++ and the RINGS code need to be installed and accessible in the system path.
First download and follow the installation instructions on their respective websites (here for Zeo++ and here for RINGS).
Then for Zeo++, create the following variable on your system:
export ZEOpp_EXEC=/path/to/zeo++-0.3/network
For RINGS, ensure that the rings binary is in your path:
export PATH=$PATH:/path/to/rings/bin
aMOF can be installed from PyPI:
pip install amof
or directly from source:
git clone https://github.com/coudertlab/amof.git
cd amof
pip install .
Special care should be taken with Asap, which can only be installed if numpy is already installed and is thus not a default dependency.
To solve this, either install asap independently (following their installation guide), or first install numpy then install amof with the rdf extra:
pip install numpy
pip install amof[rdf]
By default, graphical dependencies (used in the plot module) are not installed. To install them use the graphics extra:
pip install amof[graphics]
Support for installing aMOF with conda is not included at the moment.
However, conda can be used to first install all dependencies before installing amof with pip.
conda install -c conda-forge ase=3.20.1 asap3 pandas numpy xarray=0.19.0 dask netCDF4 bottleneck ase asap3 joblib pymatgen hvplot atomman jupyter jupyterlab jupytext matplotlib networkx scipy plotly seaborn cairosvg pyarrow selenium
A documentation is can easily be created with pdoc.
Simply run (with pdoc installed):
pdoc --html --output-dir path\to\docs path\to\amof
Examples analyses can be found in the examples folder of this repository.
An example workflow combining MD runs and their subsequent analyses can be found on the amof-workflow repository.
If you use the python package in published results (paper, conference, etc.), please cite the first paper for which it was developed: Challenges in Molecular Dynamics of Amorphous ZIFs using Reactive Force Fields.