skipi 0.1.8

Intuitive package to easily work with mathematical functions

skipi

skipi is a library to easily define mathematical functions and apply various transforms on it.

A function always consists of a domain and a map. Usually the domain is ommited since it's clear for the human what the domain is, however, not for the computer.

This library aims to combine the domain and the map into one Function object and offer multiple convenient operations on it.

Examples

Algebraic operations

Supported features are: Addition, Subtraction, Multiplication, Division, Exponentiation, Composition

import numpy as np
from skipi.function import Function

f = Function(np.linspace(0, 10, 100), lambda x: 2+x)
g = Function(f.get_domain(), lambda x: np.sin(x))
h1, h2, h3, h4, h5, h6 = f+g, f-g, f*g, g/f, f.composeWith(g), f**g

Plotting

A function is plotted using matplotlib calling plot(). If you want to plot multiple functions into one graph, simply use

g.plot()  # does not draw the graph yet
f.plot(show=True) # draws it

Remeshing

If you want to re-mesh a function on a different domain/grid, you can use remesh or vremesh. The method remesh assigns a new mesh, independent of the previous one.

f = Function(np.linspace(0, 10, 10), lambda x: np.sin(x))
f.remesh(np.linspace(0, 20, 1000))

However, if you want to restrict the domain, you can use vremesh which has a similar syntax as slice except that instead of indices we use values and it allows multiple slicing:

f = Function(np.linspace(0, 10, 1000), lambda x: np.sin(x))
f.vremesh((np.pi, 2*np.pi)) # domain is now restricted to [pi, 2pi]
f.vremesh((None, 2*np.pi)) # domain is now restricted to [0, 2pi]
f.vremesh((np.pi, None)) # domain is now restricted to [pi, 10]
f.vremesh((0.5, 1.5), (2.0, 2.5)) # domain is now restricted to [0.5, 1.5] union [2.0, 2.5]

Creating functions from data

If you don't have an analytical formulation of y = f(x), but rather have y_i and x_i values, then you can create a function by interpolation. By default, linear interpolation is used.

x_i = np.linspace(0, 10, 100)
y_i = np.sin(x_i)

f = Function.to_function(x_i, y_i)
print(f(0.1234)) # linearly interpolated, not sin(0.1234)!

Integration

Calculate the integral function of f(x) = 5x

import numpy as np
from skipi.function import Function, Integral

f = Function(np.linspace(0, 10, 100), lambda x: 5*x)
F = Integral.from_function(f) # Integral function
F.plot(show=True)

Fourier transform

Calculate the fourier transform (analytical fourier transform, not fft) of f(x) = exp(-x^2)

from skipi.fourier import FourierTransform, InverseFourierTransform

t_space, freq_space = np.linspace(-5, 5, 100), np.linspace(-10, 10, 100)
f = Function(t_space, lambda x: np.exp(-x**2))
F = FourierTransform.from_function(freq_space, f)
f2 = InverseFourierTransform.from_function(t_space, F)

# f2 should be equal to f
(f-f2).plot(show=True)