Attractors is a package for simulation and visualization of strange attractors.

Overview

attractors

Build status PyPI version PyPI license PyPI Downloads codecov

Attractors is a package for simulation and visualization of strange attractors.

Installation

The simplest way to install the module is via PyPi using pip

pip install attractors

Alternatively, the package can be installed via github as follows

git clone https://github.com/Vignesh-Desmond/attractors
cd attractors
python -m pip install .

To set up the package for development and debugging, it is recommended to use Poetry. Just install with poetry install and let Poetry manage the environment and dependencies.

Prerequisites

To generate video output, the package uses ffmpeg. Download and install from here according to your os and distribution and set PATH accordingly. Note that this is only required for generating video output.

Basic Usage

A simple code snippet using the attractors package

from attractors import Attractor

obj = Attractor("lorenz").rk3(0, 100, 10000) #rk3(starttime, endtime, simpoints)

In the above snippet, obj is an generator instance yielding an attractor instance with X, Y, and Z attributes. The generator reaches StopIteration after iterating simpoints (number of points used for the simulation) times.

The parameters of each attractor can be given as kwargs as follows:

attr = Attractor("lorenz", sigma = 5, rho = 28.5, init_coord = [0.2,0.1,0.1])

When parameters are not given, the default parameters are loaded for each attractor. In the above example, since beta is not given, the default value of 2.66667 is loaded.

To obtain the 3D coordinates of an attractor, we need to solve (usually) 3 non-linear ODE, one for each dimension. The solution can be derived via approximation using the Runge-Kutta methods. Currently, this package consists of the following iterative explicit RK methods:

  • Euler
  • RK2 (Heun, Ralston, Improved Polygon)
  • RK3
  • RK4
  • RK5

For 2nd order Runge-Kutta, the method can be specified via the positional argument rk2_method

obj = attr.rk3(0, 100, 10000, rk2_method="heun")  #methods = "heun", "ralston", "imp_poly"

A list of attractors and ODE solvers can be obtained via the static methods list_attractors() and list_des() respectively.

Plotting and Animation

The attractors package also comes with plotting and animation functions using Matplotlib. There are 2 plotting types, Multipoint and Gradient.

Plot

Multipoint plot can be used to visualize multiple attractor objects which can be used to demonstrate the chaotic nature based on perturbances in initial conditions and parameters

The following sample code shows the usage of plot_multipoint()

from attractors import Attractor
import numpy as np

n = 3
a = "rossler"
simtime = 100
simpoints = simtime * 100

# Create a list of n attractor instances
attrs = [Attractor(a) for _ in range(n)]

# Change the initial coordinates randomly for n-1 objects
for attr in attrs[1:]:
    attr.coord += np.random.normal(0, 0.01, size=3)

# Solve the ODE equations and store the generators
objs = []
for a in attrs:
    func = getattr(a, "rk3")
    objs.append(func(0, simtime, simpoints))

# Use plot_multipoint with necessary kwargs
ax = Attractor.plot_multipoint(
    simpoints - 1,
    *objs,
    dpi=240,
    bgcolor="#FFFFFF",
    palette=["#616161", "#7a7a7a", "#2e2e2e", "#1c1c1c"],
    linekwargs={"linewidth": 0.5, "alpha": 0.7},
    pointkwargs={"markersize": 1}
)

The output figure generated for the code snippet

plot_multipoint() is a class method that requires 2 arguments:

  • index : timestep of the attractor objects on plot
  • *objs : generator list

Additionally, it also takes in multiple kwargs that

  • set the figure parameters: width, height, dpi
  • set the axes limits: xlim, ylim, zlim
  • set line and point parameters via linekwargs, pointkwargs (pass to matplotlib kwargs)
  • set color
    • by theme
    • by manually by specifying bgcolor (single hexcode) and palette (list of hexcodes). Overrides theme settings if given.

The figure parameters, axes limits and theme can also be set via set_figure(), set_limits() and set_theme() methods respectively

plot_gradient() is similar to plot_multipoint(), however it can only take one generator as input. And it also takes an extra kwarg: gradientaxis to specify the axis along which the gradient is applied. (X, Y or Z).

Both plot_gradient() and plot_multipoint() returns an Matplotlib.axes object which can be used to display or save the figure and also change axes parameters after plotting.

Animate

The Animate functions set_animate_multipoint() and set_animate_gradient() are similar to their plot function counterparts. By default, the visualization output will be saved in an MPEG4 encoded video. An example for gradient animation is as follows

from attractors import Attractor

obj = Attractor("dequan_li").rk3(0, 10, 10000)

Attractor.set_animate_gradient(obj,
    width=10,
    height=10,
    theme="nord").animate(outf="example.mp4")

The above code generates a video example.mp4 in the directory that it was run from. animate is a class method acting on the Attractor class instance. It has no required argmunents and it takes the following kwargs

  • live: boolean arg to show the animated plot in a window interactively or save as output video.
  • fps: frames per second of animation
  • outf: filename of output video if generated
  • show: boolean arg to disable plt.show() and return the Matplotlib.FuncAnimation instance (only when live is True)

Both set_animate_gradient() and set_animate_multipoint() have 2 addititonal parameters: elevationrate and azimuthrate which control the rate of change of eleveation and azimuth angle for the duration of the animation respectively.

Output animation (converted to gif and sliced for README)

CLI

The attractors package also comes with its own command-line parser as a legacy interface (from v1.0.0). Simply type attractors -h to display the help message. The parser wraps the Attractor class and only supports animation.

The simplest way to visualize an Lorenz attractor is

attractors -p 100000 -s 100 -t multipoint lorenz

Full help:

$ attractors -h
usage: attractors [-v] [-h] -t {multipoint,gradient}
                  [--des {rk2,rk3,euler,rk5,rk4}] [--width WIDTH]
                  [--height HEIGHT] [--dpi DPI] [--theme THEME] -s SIMTIME -p
                  SIMPOINTS [--bgcolor BGCOLOR] [--cmap CMAP] [--fps FPS]
                  [--n N] [--rk2 {heun,imp_poly,ralston}] [--outf OUTF]
                  [--live]
                  ATTRACTOR ...

optional arguments:
  -v, --version         show program's version number and exit
  -h, --help            show this help message and exit

required arguments:
  -t {multipoint,gradient}, --type {multipoint,gradient}
                        choose simulation type
  -s SIMTIME, --simtime SIMTIME
                        set the simulation time
  -p SIMPOINTS, --simpoints SIMPOINTS
                        set the number of points to be used for the simulation

other arguments:
  --des {rk2,rk3,euler,rk5,rk4}
                        choose the Differential Equation Solver. Default: rk4
  --width WIDTH         set width of the figure Default: 16
  --height HEIGHT       set height of the figure Default: 9
  --dpi DPI             set DPI of the figure Default: 120
  --theme THEME         choose theme (color palette) to be used
  --bgcolor BGCOLOR     background color for figure in hex. Overrides theme
                        settings if specified Default: #000000
  --cmap CMAP           matplotlib cmap for palette. Overrides theme settings
                        if specified Default: jet
  --fps FPS             set FPS for animated video (or interactive plot)
                        Default: 60
  --n N                 number of initial points for Multipoint animation
                        Default: 3
  --rk2 {heun,imp_poly,ralston}
                        method for 2nd order Runge-Kutta if specified to be
                        used. Default: heun
  --outf OUTF           output video filename Default: output.mp4
  --live                live plotting instead of generating video.

Attractor settings:
  Choose one of the attractors and specify its parameters

  ATTRACTOR
    lorenz              Lorenz attractor
    rabinovich_fabrikant
                        Rabinovich Fabrikant attractor
    lotka_volterra      Lotka Volterra attractor
    rossler             Rossler attractor
    wang_sun            Wang Sun attractor
    rikitake            Rikitake attractor
    nose_hoover         Nose Hoover attractor
    aizawa              Aizawa attractor
    three_cell_cnn      Three Cell CNN attractor
    bouali_type_1       Bouali Type 1 attractor
    bouali_type_2       Bouali Type 2 attractor
    bouali_type_3       Bouali Type 3 attractor
    finance             Finance attractor
    burke_shaw          Burke Shaw attractor
    moore_spiegel       Moore Spiegel attractor
    sakarya             Sakarya attractor
    dadras              Dadras attractor
    halvorsen           Halvorsen attractor
    hadley              Hadley attractor
    chen                Chen attractor
    chen_lee            Chen Lee attractor
    chen_celikovsky     Chen Celikovsky attractor
    thomas_cyclically_symmetric
                        Thomas Cyclically Symmetric attractor
    dequan_li           Dequan Li attractor
    yu_wang             Yu Wang attractor

Each attractor also has its own parameters to set. The settings for each attractor can be obtained by the help command: attractors ATTRACTOR -h

Attractor help

$ attractors finance -h
usage: attractors finance [-h] [--a A] [--b B] [--c C]
                          [--initcoord INITCOORD INITCOORD INITCOORD]
                          [--xlim XLIM XLIM] [--ylim YLIM YLIM]
                          [--zlim ZLIM ZLIM]

optional arguments:
  -h, --help            show this help message and exit

Finance attractor parameters:
  --a A                 Parameter for Finance attractor Default: 1e-05
  --b B                 Parameter for Finance attractor Default: 0.1
  --c C                 Parameter for Finance attractor Default: 1.0
  --initcoord INITCOORD INITCOORD INITCOORD
                        Initial coordinate for Finance attractor. Input
                        format: "x y z" Default: [0.0, -10.0, 0.1]
  --xlim XLIM XLIM      x axis limits for figure. Input format: "xmin xmax"
                        Default: [-3.0, 3.0]
  --ylim YLIM YLIM      y axis limits for figure. Input format: "ymin ymax"
                        Default: [-5.0, -15.0]
  --zlim ZLIM ZLIM      z axis limits for figure. Input format: "zmin zmax"
                        Default: [-1.5, 1.5]

Changelog

See changelog for previous versions

Development

This package is under early stages of development it's open to any constructive suggestions. Please send bug reports and feature requests through issue trackers and pull requests.

License

This package is licensed under the MIT License

You might also like...
Exploratory analysis and data visualization of aircraft accidents and incidents in Brazil.
Exploratory analysis and data visualization of aircraft accidents and incidents in Brazil.

Exploring aircraft accidents in Brazil Occurrencies with aircraft in Brazil are investigated by the Center for Investigation and Prevention of Aircraf

Advanced_Data_Visualization_Tools - The present hands-on lab mainly uses Immigration to Canada dataset and employs advanced visualization tools such as word cloud, and waffle plot to display relations between features within the dataset.
Fast data visualization and GUI tools for scientific / engineering applications

PyQtGraph A pure-Python graphics library for PyQt5/PyQt6/PySide2/PySide6 Copyright 2020 Luke Campagnola, University of North Carolina at Chapel Hill h

Apache Superset is a Data Visualization and Data Exploration Platform
Apache Superset is a Data Visualization and Data Exploration Platform

Superset A modern, enterprise-ready business intelligence web application. Why Superset? | Supported Databases | Installation and Configuration | Rele

Fast data visualization and GUI tools for scientific / engineering applications

PyQtGraph A pure-Python graphics library for PyQt5/PyQt6/PySide2/PySide6 Copyright 2020 Luke Campagnola, University of North Carolina at Chapel Hill h

3D plotting and mesh analysis through a streamlined interface for the Visualization Toolkit (VTK)
3D plotting and mesh analysis through a streamlined interface for the Visualization Toolkit (VTK)

PyVista Deployment Build Status Metrics Citation License Community 3D plotting and mesh analysis through a streamlined interface for the Visualization

Fast data visualization and GUI tools for scientific / engineering applications

PyQtGraph A pure-Python graphics library for PyQt5/PyQt6/PySide2/PySide6 Copyright 2020 Luke Campagnola, University of North Carolina at Chapel Hill h

3D plotting and mesh analysis through a streamlined interface for the Visualization Toolkit (VTK)
3D plotting and mesh analysis through a streamlined interface for the Visualization Toolkit (VTK)

PyVista Deployment Build Status Metrics Citation License Community 3D plotting and mesh analysis through a streamlined interface for the Visualization

Apache Superset is a Data Visualization and Data Exploration Platform
Apache Superset is a Data Visualization and Data Exploration Platform

Apache Superset is a Data Visualization and Data Exploration Platform

Releases(1.4.1)
  • 1.4.1(Aug 1, 2021)

  • 1.4.0(Jul 31, 2021)

    Announcements

    • Minor release introducing documentation and added features

    Changes

    • Added 3 new attractors: Newton-Leipnik, Rucklidge and Shimizu-Morioka (Thanks @Sarath-24)
    • index is now a kwarg for plot methods
    • Added docstrings
    • Added type hints
    • Set-up documentation with Sphinx-Napolean (Google Style) and ReadTheDocs
    • Renamed attractors: Aizawa -> Langford, Chen Celikovsky -> Chen LU, Thomas Cyclically Symmetric -> Thomas
    • Added tests for video generation with FFMPEG and colormaps
    Source code(tar.gz)
    Source code(zip)
  • 1.3.0(Jul 20, 2021)

    Announcements

    • Minor release with breaking changes (not backwards compatible)

    Changes

    • DES methods now return a generator instead of an Attractor instance
    • X, Y, Z attributes of the Attractor are now single-valued floats instead of lists
    • Modified plotting and animation methods to support generators
    • Removed slice method
    • (Dev) Coverage tests
    Source code(tar.gz)
    Source code(zip)
  • 1.2.0(Jul 15, 2021)

    Announcements

    • Minor version bump adding 2 new attractors

    Changes

    • Added 2 new attractors : Dequan Li and Yu Wang
    • Removed legacy animate functions animate_gradient() and animate_sim()
    • Added kwargs to control gradient axis, elevation and azimuth rate.
    • Added line and point kwargs to pass to matplotlib
    • Tweaked default params for multiple attractors
    • Minor bugfixes
    Source code(tar.gz)
    Source code(zip)
  • 1.1.1(Jul 9, 2021)

  • 1.1.0(Jul 7, 2021)

    Announcements

    • Update of attractors package with various new features and bugfixes

    Changes

    • Completely overhauled attractor class for modularity
    • Pooling update_func() for figure with pathos
    • New methods for setting various params independently
    • Included plotting methods
    • Bugfixes and minor QoL changes
    Source code(tar.gz)
    Source code(zip)
  • 1.0.0(Jul 3, 2021)

Owner
Vignesh M
Student @ SSN, Currently pursuing Electronics and Communication Engineering
Vignesh M
This is a super simple visualization toolbox (script) for transformer attention visualization ✌

Trans_attention_vis This is a super simple visualization toolbox (script) for transformer attention visualization ✌ 1. How to prepare your attention m

Mingyu Wang 3 Jul 9, 2022
Python package for hypergraph analysis and visualization.

The HyperNetX library provides classes and methods for the analysis and visualization of complex network data. HyperNetX uses data structures designed to represent set systems containing nested data and/or multi-way relationships. The library generalizes traditional graph metrics to hypergraphs.

Pacific Northwest National Laboratory 304 Dec 27, 2022
A Python package that provides evaluation and visualization tools for the DexYCB dataset

DexYCB Toolkit DexYCB Toolkit is a Python package that provides evaluation and visualization tools for the DexYCB dataset. The dataset and results wer

NVIDIA Research Projects 107 Dec 26, 2022
Python Package for CanvasXpress JS Visualization Tools

CanvasXpress Python Library About CanvasXpress for Python CanvasXpress was developed as the core visualization component for bioinformatics and system

Dr. Todd C. Brett 5 Nov 7, 2022
Smoking Simulation is an app to simulate the spreading of smokers and non-smokers, their interactions and population during certain amount of time.

Smoking Simulation is an app to simulate the spreading of smokers and non-smokers, their interactions and population during certain

Bohdan Ruban 5 Nov 8, 2022
A simple Monte Carlo simulation using Python and matplotlib library

Monte Carlo python simulation Install linux dependencies sudo apt update sudo apt install build-essential \ software-properties-commo

Samuel Terra 2 Dec 13, 2021
A Scheil-Gulliver simulation tool using pycalphad.

scheil A Scheil-Gulliver simulation tool using pycalphad. import matplotlib.pyplot as plt from pycalphad import Database, variables as v from scheil i

pycalphad 6 Dec 10, 2021
This is a Boids Simulation, written in Python with Pygame.

PyNBoids A Python Boids Simulation This is a Boids simulation, written in Python3, with Pygame2 and NumPy. To use: Save the pynboids_sp.py file (and n

Nik 17 Dec 18, 2022
Debugging, monitoring and visualization for Python Machine Learning and Data Science

Welcome to TensorWatch TensorWatch is a debugging and visualization tool designed for data science, deep learning and reinforcement learning from Micr

Microsoft 3.3k Dec 27, 2022
China and India Population and GDP Visualization

China and India Population and GDP Visualization Historical Population Comparison between India and China This graph shows the population data of Indi

Nicolas De Mello 10 Oct 27, 2021