This project aims to provide a framework for sampling molecular dynamical systems. The package offers different dynamical systems which can be simulated with different time evolution methods (see also below for more info).
In the environments folder different dynmical models can be found as well as different time integration methods. The gradient of the different resulting potential is computed by using atomatic differentiation.
This folder contains different time integration methods which can be combined with a dynamical system which can be found in the model folder. At the moment one can find
- Verlet
- Euler-Maruyama
In this folder one can find different initializations for differen dynamical systems. At the moment there are
- clone the repo
$ git clone git@github.com:jannes002A/molecules.git
- set python version (>=3.7)
$ pyenv local 3.9.7
- create virtual environment and install required packages
$ make venv
- activate venv
$ source venv/bin/activate
You can run demos for the different models by running a specific script in the "/src/molecules/demos/" folder for e.g. in the SDE case
$ python src/molecules/demos/sde/run_sde_eng.py
The SDE environment samples a double well potential in a choosen dimension. Here is a 10d examples
This model runs an anisotropic oscillator model.
With this env you can run butane trajectories.
