Generates random simple polygons. This can be useful to test computational geometry algorithms.

$ pip install polygenerator

There are 3 functions and each returns a list of (x, y) tuples:

• random_convex_polygon(num_points)
• random_polygon(num_points)
• random_star_shaped_polygon(num_points)

All polygons are generated to be counterclockwise. You can reverse the order outside if you need the points in clockwise order.

The generated polygon is made to fit the bounding box (0.0, 0.0) ... (1.0, 1.0) and you can then scale and translate it to where you need it.

from polygenerator import (
    random_polygon,
    random_star_shaped_polygon,
    random_convex_polygon,
)

from plot import plot_polygon

# this is just so that you can reproduce the same results
import random
random.seed(5)


polygon = random_polygon(num_points=20)

print(polygon)
# [(0.752691110661913, 0.948158571633034), (0.7790276993942304, 0.05437135270534656), ..., (0.633385213909564, 0.7365967958574935)]

plot_polygon(polygon, "random_polygon.png")

polygon = random_star_shaped_polygon(num_points=20)
plot_polygon(polygon, "random_star_shaped_polygon.png")

polygon = random_convex_polygon(num_points=20)
plot_polygon(polygon, "random_convex_polygon.png")

• For the generation of a concave/general polygon, algorithms with better scaling exist but this was good enough for me since for testing I did not need polygons with more than 100 points. Improvements welcome.