AMEP documentation#

The AMEP (Active Matter Evaluation Package) Python library is a powerful tool for analyzing data from molecular-dynamics (MD), Brownian-dynamics (BD), and continuum simulations. It comprises various methods to analyze structural and dynamical properties of condensed matter systems in general and active matter systems in particular. AMEP is exclusively built on Python, and therefore, it is easy to modify and allows to easily add user-defined functions. AMEP provides an efficient data format for saving both simulation data and analysis results based on the HDF5 file format. To be fast and usable on modern HPC (High Performance Computing) hardware, the methods are optimized to run also in parallel.

Description#

The AMEP Python library provides a unified framework for handling both particle-based and continuum simulation data. It is made for the analysis of molecular-dynamics (MD), Brownian-dynamics (BD), and continuum simulation data of condensed matter systems and active matter systems in particular. AMEP provides a huge variety of analysis methods for both data types that allow to evaluate various dynamic and static observables based on the trajectories of the particles or the time evolution of continuum fields. For fast and efficient data handling, AMEP provides a unified framework for loading and storing simulation data and analysis results in a compressed, HDF5-based data format. AMEP is written purely in Python and uses powerful libraries such as NumPy, SciPy, Matplotlib, and scikit-image commonly used in computational physics. Therefore, understanding, modifying, and building up on the provided framework is comparatively easy. All evaluation functions are optimized to run efficiently on HPC hardware to provide fast computations. To plot and visualize simulation data and analysis results, AMEP provides an optimized plotting framework based on the Matplotlib Python library, which allows to easily plot and animate particles, fields, and lines. Compared to other analysis libraries, the huge variety of analysis methods combined with the possibility to handle both most common data types used in soft-matter physics and in the active matter community in particular, enables the analysis of a much broader class of simulation data including not only classical molecular-dynamics or Brownian-dynamics simulations but also any kind of numerical solutions of partial differential equations. The following table gives an overview on the observables provided by AMEP and on their capability of running in parallel and processing particle-based and continuum simulation data.

Observable	Parallel	Particles	Fields
Spatial Correlation Functions:
RDF (radial pair distribution function)	✔	✔	❌
PCF2d (2d pair correlation function)	✔	✔	❌
PCFangle (angular pair correlation function)	✔	✔	❌
SFiso (isotropic static structure factor)	✔	✔	✔
SF2d (2d static structure factor)	✔	✔	✔
SpatialVelCor (spatial velocity correlation function)	✔	✔	❌
PosOrderCor (positional order correlation function)	✔	✔	❌
HexOrderCor (hexagonal order correlation function)	✔	✔	❌
Local Order:
Voronoi tesselation	❌	✔	❌
Local density	❌	✔	✔
Local packing fraction	❌	✔	❌
k-atic bond order parameter	❌	✔	❌
Next/nearest neighbor search	❌	✔	❌
Time Correlation Functions:
MSD (mean square displacement)	❌	✔	❌
VACF (velocity autocorrelation function)	❌	✔	❌
OACF (orientation autocorrelation function)	❌	✔	❌
Cluster Analysis:
Clustersize distribution	❌	✔	✔
Cluster growth	❌	✔	✔
Radius of gyration	❌	✔	✔
Linear extension	❌	✔	✔
Center of mass	❌	✔	✔
Gyration tensor	❌	✔	✔
Inertia tensor	❌	✔	✔
Miscellaneous:
Translational/rotational kinetic energy	❌	✔	❌
Kinetic temperature	❌	✔	❌

Installation#

The AMEP library can be installed via pip, conda, or by manually adding the amep directory to your Python path. Installation via pip or conda is recommended. To use all plot animation features, please additionally install FFmpeg (https://ffmpeg.org/) on your machine (see below).

Installation via pip#

AMEP can be simply installed from PyPI via

pip install amep

Installation via conda#

AMEP can be simply installed from conda-forge via

conda install conda-forge::amep

Manual installation#

Before installing AMEP manually, ensure that your Python environment fulfills the required specifications as published together with each release. If your Python environment is set up, download the latest version from amepproject/amep and extract the zipped file. Then, add the path to your Python path and import amep:

import sys
sys.path.append('/path/to/amep-<version>')
import amep

Alternatively, you can add the path permanently to your Python path by adding the line

export PYTHONPATH="${PYTHONPATH}:/path/to/amep-<version>"

to the .bash_profile file (Linux only). If you use the Anaconda distribution, you can alternatively add the amep directory to Lib/site-packages in the Anaconda installation path.

FFmpeg#

AMEP provides the possibility to animate plots and trajectories. To enable all animation features, FFmpeg must be installed on the device on which you run AMEP. FFmpeg is not automatically installed when you install AMEP. Please visit https://ffmpeg.org/download.html to download FFmpeg and to get further information on how to install FFmpeg on your machine.

Getting started#

The following example briefly demonstrates the AMEP workflow. A typical task is to calculate the average of an observable over several frames of the simulation (time average). In the example below, we first load LAMMPS simulation data stored as individual dump*.txt files for each frame, and second, we calculate and plot the time-averaged radial pair distribution function (RDF).

# import the amep library
import amep

# load simulation data (creates an h5amep file and returns a trajectory object)
traj = amep.load.traj('./examples/data/lammps')

# calculate the radial pair-distribution function, skip the first half of the
# trajectory, and average over 10 frames with equal distance in time
rdf = amep.evaluate.RDF(traj, skip=0.5, nav=10, nbins=1000)

# save result in file
rdf.save('./rdf.h5')

# plot the result
fig, axs = amep.plot.new()
axs.plot(rdf.r, rdf.avg)
axs.set_title(amep.plot.to_latex(rdf.name))
axs.set_xlim(0,10)
axs.set_xlabel(r'$r$')
axs.set_ylabel(r'$g(r)$')
amep.plot.set_locators(axs, which='both', major=1, minor=0.2)
fig.savefig(rdf.name + '.png')
fig.savefig(rdf.name + '.pdf')

For more detailed examples, check the examples directory.

Module descriptions#

In the following, we provide a list of all AMEP modules together with a short description.

Module:	Description:
base.py	base classes (backend)
cluster.py	cluster analysis for particle-based data
continuum.py	coarse-graining and continuum field analysis
evaluate.py	trajectory analysis
functions.py	mathematical functions and fitting
load.py	loading simulation data and analysis results
order.py	spatial order analysis
pbc.py	handling of periodic boundary conditions
plot.py	visualization and animation
reader.py	simulation data reader (backend)
spatialcor.py	spatial correlation functions
statistics.py	statistical analysis
thermo.py	thermodynamic observables
trajectory.py	trajectory classes (backend)
utils.py	collection of utility functions

Data Formats#

AMEP is compatible with multiple data formats. The current version can load particle-based simulation data obtained from LAMMPS (https://www.lammps.org) and continuum simulation data with the following format: The main directory should contain one file with data that stays constant throughout the entire simulation such as the boundaries of the simulation box, the shape of the underlying grid and the grid coordinates. It’s standard name is grid.txt and it should have the following form:

BOX:
<X_min> <X_max>
<Y_min> <Y_max>
<Z_min> <Z_max>
SHAPE:
<nx> <ny> <nz>
COORDINATES: X Y Z
<X_0> <Y_0> <Z_0>
<X_1> <Y_1> <Z_1>
...

All data that varies in time is to be put into files named dump<index>.txt. The index should increase with time, i.e., the file dump1000.txt should contain the data of the continuum simulation at timestep 1000, and the prefix dump is user-defined and can be changed (if it is changed, the new naming convention has to be specified with the keyword dumps in amep.load.traj, e.g., for files named field_100.txt, field_200.txt, …, use dumps='field_*.txt'). The data files should have the following form:

TIMESTEP:
<Simulation timestep>
TIME:
<Physical time>
DATA: <fieldname 0> <fieldname 1> <fieldname 2> <fieldname 3>
<field 0 0> <field 1 0> <field 2 0> <field 3 0>
<field 0 1> <field 1 1> <field 2 1> <field 3 1>
<field 0 2> <field 1 2> <field 2 2> <field 3 2>
...

Support#

If you need support for using AMEP, we recommend to use our GitHub discussions page. If you find a bug, please create an issue.

Creating issues#

To create an issue, go to amepproject/amep#issues and click on New issue. Then, continue with the following steps:

Add a short and clear title.
Write a precise description of the bug which you found. If you got an error message, add it to the description together with a short code snippet with which you can reproduce the error.
If it is already known how the bug can be fixed, please add a short to-do list to the description.

When creating issues, text is written as markdown, which allows formatting text, code, or tables for example. A useful guide can be found here.

Roadmap#

Planned new features for future releases are listed as issues in the issue list.

Contributing#

If you want to contribute to this project, please check the file CONTRIBUTING.md.

Contributors/Authors#

The following people contributed to AMEP:

Lukas Hecht (creator and lead developer)
Kay-Robert Dormann (developer)
Kai Luca Spanheimer (developer)
Aritra Mukhopadhyay (developer)
Mahdieh Ebrahimi (developer)
Suvendu Mandal (developer)
Benno Liebchen (planning)
Lukas Walter (former developer)
Malte Cordts (former developer)

Acknowledgments#

Many thanks to the whole group of Benno Liebchen at the Institute for Condensed Matter Physics at Technical University of Darmstadt for testing and supporting AMEP, for fruitful discussions, and for very helpful feedback. Additionally, the authors gratefully acknowledge the computing time provided to them at the NHR Center NHR4CES at TU Darmstadt (project number p0020259). This is funded by the Federal Ministry of Education and Research, and the state governments participating on the basis of the resolutions of the GWK for national high performance computing at universities (https://www.nhr-verein.de/unsere-partner).

License#

The AMEP library is published under the GNU General Public License, version 3 or any later version. Please see the file LICENSE for more information.