region_plot

A function for plotting (two-dimensional) regions of points that satisfy a given predicate.

Installation

Currently, the easiest approach for installing region_plot is to clone it from github and run pip install locally.

git clone https://github.com/cetinkaya/region_plot.git
pip install --user region_plot/

Use

To plot the region {(x,y) : sin(x^2)+3y^2 <= 1} we can write:

import numpy as np
import matplotlib.pyplot as pl
import region_plot as rp

def example0():
    def pred(x, y):
        return np.sin(x*x) + 3*y*y <= 1
    pl.figure(figsize=(5, 5))
    rp.region_plot(pred, (-2, 2), (-2, 2))
    pl.xlabel("$x$", fontsize=14)
    pl.ylabel("$y$", fontsize=14)
    pl.title("$\sin(x^2)+3y^2 \leq 1$", fontsize=14)
    pl.tight_layout()
    pl.savefig("example0.png")

example0()

which results in the following image:

Here we have plotted the region within the rectangle {(x,y) : -2 <= x,y <= 2} by passing the arguments (-2, 2), (-2, 2).

Subdivisions

region_plot works by evaluating the given predicate (sin(x^2)+3y^2 <= 1) at different points. We can partially control the evaluation procedure through the parameters discovery and nof_subdivisions.

Initially, region_plot evaluates the predicate at the corner points of the initial rectangle. Then the initial rectangle is divided into smaller rectangles and those smaller rectangles can be divided further. Here the parameter nof_subdivisions controls how many times in total the rectangles are divided. If the corners of a small enough rectangle all satisfy the predicate, then the predicate is considered to be true in all points in the rectangle. This can introduce errors. To improve the approximation of the region, more points may need to be evaluated. The function region_plot ranks all current rectangles to choose which rectangle to subdivide. Typically more priority is given to subdividing the rectangles for which only some of the corners satisfy the predicate. But this can be changed. Increasing the discovery parameter increases the subdivision ranking of rectangles whose corners all satisfy (or all fail to satisfy) the predicate. In addition to discovery and nof_subdivisions, other arguments can also be passed to change the style of the plot. For instance, color and transparency of the region can be changed by using facecolor and alpha parameters.

import numpy as np
import matplotlib.pyplot as pl
import region_plot as rp

def example1():
    def pred(x, y):
        return np.sin(x*x) + 3*y*y <= 1
    pl.figure(figsize=(5, 5))
    rp.region_plot(pred, (-2, 2), (-2, 2), discovery=0.01, nof_subdivisions=5000, facecolor="#9090FF", alpha=0.5)
    pl.xlabel("$x$", fontsize=14)
    pl.ylabel("$y$", fontsize=14)
    pl.title("$\sin(x^2)+3y^2 \leq 1$", fontsize=14)
    pl.tight_layout()
    pl.savefig("example1.png")

example1()

In this case we obtain:

Effect of discovery

To see the effect of discovery parameter, consider the following example where the animation shows the evaluated rectangles for discovery values 0.01 and 0.1. To obtain the animation we use the following code:

import numpy as np
import matplotlib.pyplot as pl
import matplotlib.animation as animation
import region_plot as rp

def animate(v=0):
    def pred(x, y):
        return np.sin(x*x) + 3*y*y <= 1
    sp = pl.subplot(1, 2, 1)
    pl.title("$\sin(x^2)+3y^2 \leq 1\,\,(\\mathrm{discovery}=0.01)$", fontsize=14)
    rp.region_plot(pred, (-2, 2), (-2, 2), discovery = 0.01, nof_subdivisions=v, show_all=True, edgecolor="#000000", facecolor="#9090FF", alpha=0.5)
    pl.xlabel("$x$", fontsize=14)
    pl.ylabel("$y$", fontsize=14)
    sp = pl.subplot(1, 2, 2)
    rp.region_plot(pred, (-2, 2), (-2, 2), discovery = 0.1, nof_subdivisions=v, show_all=True, edgecolor="#000000", facecolor="#9090FF", alpha=0.5)
    pl.xlabel("$x$", fontsize=14)
    pl.ylabel("$y$", fontsize=14)
    pl.title("$\sin(x^2)+3y^2 \leq 1\,\,(\\mathrm{discovery}=0.1)$", fontsize=14)

fig = pl.figure(figsize=(12, 6))
ani = animation.FuncAnimation(fig, animate, frames=200)
ani.save("example2.mp4")

The saved animation is as follows.