Skip to content

Latest commit

 

History

History
142 lines (106 loc) · 6.32 KB

README.md

File metadata and controls

142 lines (106 loc) · 6.32 KB

asyncmd

Synopsis

asyncmd is a library to write concurrent code to run and analyze molecular dynamics simulations using pythons async/await synthax.

Motivation

Molecular dynamics simulations are fun and we can learn a lot about the simulated system. Running many molecular dynamics simulations of the same system concurrently is tedious, error-prone and boring but we can learn even more about the simulated system and are more efficient in doing so. This library addresses the tedious, error-prone and boring part of setting up many similar simulations, but it leaves you with the fun part of understanding the simulated system.

Code Example

Run N gromacs engines concurently from configurations randomly picked up along a trajectory (traj.trr) for n_steps integration steps each, drawing random Maxwell-Boltzmann velocities for each configuration on the way. Finally turn the python function func (which acts on Trajectory objects) into an asyncronous and cached function by wrapping it and apply it on all generated trajectories concurrently:

import asyncio
import numpy as np
import asyncmd
import asyncmd.gromacs as asyncgmx

in_traj = asyncmd.Trajectory(trajectory_files="traj.trr", structure_file="conf.gro")
# get a random number generator and draw N random frames (with replacement)
rng = np.default_rng()
frame_idxs = rng.choice(len(in_traj), size=N)
# use the RandomVelocitiesFrameExtractor to directly get the frames with MB-vels
extractor = asyncmd.trajectory.convert.RandomVelocitiesFrameExtractor(T=303)
mdps = [asyncgmx.MDP("config.mdp") for _ in range(N)]
# MDConfig objects (like MDP) behave like dictionaries and are easy to modify
for i, mdp in enumerate(mdps):
    # here we just modify the output frequency for every engine separately
    # but you can set any mdp option like this
    # Note how the values are in the correct types? I.e. that they are ints?
    mdp["nstxout"] *= (i + 1)
    mdp["nstvout"] *= (i + 1)
# create N gromacs engines
engines = [asyncgmx.GmxEngine(mdp=mdp, gro_file="conf.gro", top_file="topol.top",
                              # optional (can be omited or None), however naturally without an index file
                              # you can not reference custom groups in the .mdp-file or MDP object
                              ndx_file="index.ndx",
                              )
           for mdp in mdps]
# extract starting configurations with MB-vels and save them to current directory
start_confs = await asyncio.gather(*(extractor.extract_async(
                                          outfile=f"start_conf{i}.trr",
                                          traj_in=in_traj, idx=idx)
                                     for i, idx in enumerate(frame_idxs)))
# prepare the MD (for gromacs this is essentially a `grompp` call)
await asyncio.gather(*(e.prepare(starting_configuration=conf,
                                 workdir=".", deffnm=f"engine{i}")
                       for i, (conf, e) in enumerate(zip(start_confs, engines))
                       )
                     )
# and run the molecular dynamics
out_trajs = await asyncio.gather(*(e.run_steps(nsteps=n_steps) for e in engines))
# wrapp `func` and apply it on all output trajectories concurrently
wrapped_func = asyncmd.trajectory.PyTrajectoryFunctionWrapper(function=func)
cv_vals = await asyncio.gather(*(wrapped_func(traj) for traj in out_trajs))

Note that running via the SLURM queueing system is as easy as replacing the GmxEngine with a SlurmGmxEngine and the PyTrajectoryFunctionWrapper with a SlurmTrajectoryFunctionWrapper (and suppling them both with sbatch script skeletons).

For an in-depth introduction see also the examples folder in this repository which contains jupyter notebooks on various topics.

Installation

pip install directly from the repository

Please note that you need to have git-lfs (an open source git extension) setup to get all input files needed to run the notebooks in the examples folder. However, no git-lfs is needed to get a working version of the library.

git clone https://github.com/bio-phys/asyncmd.git
cd asyncmd
pip install .

API Reference

The documentation can be build with sphinx, use e.g. the following to build it in html format:

cd asyncmd  # Need to be at the top folder of the repository for the next line to work
sphinx-build -b html docs/source docs/build/html

Use pip install .\[docs\] to install the requirements needed to build the documentation.

Tests

Tests use pytest. To run them just install asycmd with the test requirements

git clone https://github.com/bio-phys/asyncmd.git
cd asyncmd
pip install .\[tests\]

And then run the tests (against the installed version) as

pytest

Contribute

If you discover any issues or want to propose a new feature please feel free to open an issue or a pull request!

Developer install

For the developer install I recommend:

git clone https://github.com/bio-phys/asyncmd.git
cd asyncmd
pip install -e .\[dev\]

This will in addition to the requirements to run the tests and to build the documentation install flake8 and some of its plugins, such that you get yelled at to write nicely formated code. It will also install coverage and its pytest-cov plugin such that you have an idea of the test coverage for your newly added code. To get a nice html coverage report you can run the tests as

pytest --cov=asyncmd --cov-report=html

Contributors

This project was originally conceived and started by Hendrik Jung in 2021/2022. You can also see the file __about__.py or check the string asyncmd.__author__. When you contribute code dont forget to add your name there to claim the credit for your work!

License

asyncmd is under the terms of the GNU general public license version 3 or later, i.e. SPDX identifier "GPL-3.0-or-later".


This README.md is printed from 100% recycled electrons.