Skip to main content

The triple-random constrainable curve generator for smooth function perturbation

Project description


The triple-random smooth curve generator for function perturbation

Smurves is a tool for random smooth curve generation that allows for several constraints to be put on the generation process. If offers a more constrainable alternative to using, for example, Gaussian processes for that purpose. The method is based on Newtonian projectile motion and takes its inspiration from Brandon Sanderson's book series The Stormlight Archive and the books' magic system called "surgebinding". More specifically, this approach makes note of the parts that deal with the change in direction and magnitude of gravity for an in-flight object, which provided the initial concept.

With this idea in mind, Smurves generates smooth curves by randomly sampling the gravitational force, as well as locations and the number of gradient sign changes, for one projectile per curve path and making sure that all constraints are adhered to, with velocity and flight angle being retained after gravitational direction switches.

In essence, the code's inner workings can be imagined as firing a bullet at either zero angle or a random angle, from a specified or random point at either the left or the right side of the x-axis interval, depending on the user preferences. At a random number of random points, gravity gets turned upside down with a new random magnitude while the bullet continues its flight, all subject to the constraints set by the user for the properties of the required curves.

The motivation for this approach was to find a novel way* of generating utterly random curves with certain constraints to smoothly perturbate functions, for example the matter power spectrum in cosmology.

* Because it's from a novel. This documentation comes fully equipped with bad jokes.


Smurves can be installed via PyPI, with a single command in the terminal:

pip install smurves

Alternatively, the file can be downloaded from the folder smurves in this repository and used locally by placing the file into the working directory for a given project. An installation via the terminal is, however, highly recommended, as the installation process will check for the package requirements and automatically update or install any missing dependencies, thus sparing the user the effort of troubleshooting and installing them themselves.

Quickstart guide

The descriptions and example usage below provide a quick tutorial on Smurves. In addition, the examples.ipynb Jupyter Notebook in the examples folder in this repository show the use of the tool for various constraints and with explanations for each parameter set, and with the code necessary to plot the curves.

In addition to the number of curves and interval constraints for both x-axis and y-axis beyond which the curves shouldn't stray, the tool requires the number of measurement points amd the maximum number of directional changes per curve.

Six optional parameters include the placement of a point in which the curves should converge, the choice of a logarithmic scale for the x-axis, the choice to launch the curve trajectories at random instead of zero angles, the choice to let the curves converge on the right instead of the left side if a convergence point is provided, the percentiles along the x-axis before and which no directional gravity changes should be implemented, and the placing of a threshold point before which no deviation from the convergence point's x-axis value should take place. These parameters are described in the table below.

Variables Explanations Default
n_curves The number of curves you want to generate
x_interval The allowed x-axis interval for the curves
y_interval The allowed y-axis interval for the curves
n_measure The number of equally-spaced measurement points
direction_maximum The maximum number of allowed gradient changes
convergence_point (optional) The left-side point of convergence for curves
log_scale (optional) Whether measurements should be on a log-scale False
random_launch (optional) Whether the first launch angle should be random False
right_convergence (optional) Whether convergence should be on the right side False
change_range (optional) The x-axis percentiles before and after which no
gradient changes should take place for curves
[0.1, 0.9]
change_spacing (optional) The minimum space in measurement points between
the different points of a gravitational force change
change_ratio (optional) The multiplier for the last partial trajectory to get
the upper limit for the next partial trajectory's force
start_force (optional) The point of the first deviation from unity None

After the installation via PyPI, or using the file locally, the usage looks like this:

from smurves import surgebinder

curves = surgebinder(n_curves = 10,
                     x_interval = [0.001, 10.0],
                     y_interval = [0.0, 5.0],
                     n_measure = 100,
                     direction_maximum = 3,
                     convergence_point = [0.001, 1.0],
                     log_scale = True,
                     change_range = [0.2, 0.8],
                     start_force = 0.01)

Note that if we want a logarithmic scale, the x-axis interval, as well as the start_force parameter to enforce no deviation before that value, have to provide powers of ten, e.g. 0.1, 10 or 1000.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

smurves-1.0.0.tar.gz (15.8 kB view hashes)

Uploaded source

Built Distribution

smurves-1.0.0-py3-none-any.whl (15.4 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page