Attractors
A small module that provides functions and classes for very efficient simulation and rendering of iterated function systems; dynamical systems, roulette curves, (strange) attractors, and so on.
Installation
Clone this repository and install with pip or another package manager. Alternatively, just clone/download the repo and use a relative import to include the scripts in your project.
Dependencies
- Numba
- NumPy
- Matplotlib
- SciPy (optional, only needed for image postprocessing)
- nbdev (if building from source/developing)
Usage
attractors tries to conform to the principle of least astonishment wherever possible (and variable names, classes, parameters etc. aim to be readable), so using the tools should be fairly intuitive.
If we want to make a new RouletteCurve, for instance, the following will initialize one with the default parameters (including randomized arm lengths/rotation speeds):
R = RouletteCurve(num_sections=2)
Then, we can run simulate and render; function chaining is usually available since most class methods return the class instance ("self"):
R.simulate_accelerated(steps=10000).render(mode='hist', hist_args=dict(bins=150))
<matplotlib.image.AxesImage at 0x227b753a910>
Other rendering modes are available; line will trace between each generated point.
RouletteCurve(num_sections=2).simulate_accelerated(steps=200).render(mode='line')
<matplotlib.image.AxesImage at 0x227b76a13d0>
A softer render can be achieved using dist (and an optional falloff value that corresponds to the norm order when generating the brush).
RouletteCurve(num_sections=3).simulate_accelerated(steps=10000).render(mode='dist', falloff=3)
[[0.31748021 0.37475618 0.39893899 0.39893899 0.37475618]
[0.37475618 0.52913368 0.65863376 0.65863376 0.52913368]
[0.39893899 0.65863376 1.58740105 1.58740105 0.65863376]
[0.39893899 0.65863376 1.58740105 1.58740105 0.65863376]
[0.37475618 0.52913368 0.65863376 0.65863376 0.52913368]]
<matplotlib.image.AxesImage at 0x227bb8fd2e0>
