porchlight

porchlight is a function management suite that handles shared inputs and outputs of methods and/or functions which evolve over the lifetime of a program.

This package's original intent was to be a part of a modular scientific package yet to be released. Rather than isolating this method to a single model, the already-developed work has been modified to stand alone as a package.

porchlight does not have any dependencies outside of the standard CPython library. Please note that porchlight requires Python 3.9+, and that examples may require external libraries such as numpy and matplotlib.

Installation

You can install porchlight using pip:

pip install porchlight

Usage

The main object used in porchlight is the porchlight.Neighborhood object. This groups all functions together and keeps track of call order and parameters.

import porchlight


# To add a function, we simply define it and pass it to porchlight.
def increase_x(x: int, y: int) -> int:
    x = x * y
    return x

# Type annotations are optional, as with normal python.
def string_x(x):
    x_string = f"{x = }"
    return x_string

def increment_y(y=0):
    y = y + 1
    return y

# Generating a complete, coupled model between these functions is as simple as
# adding all these functions to a Neighborhood object.
neighborhood = Neighborhood([increment_y, increase_x, string_x])

# The neighborhood object inspects the function, finding input and output
# variables if present. These are added to the collections of functions and
# parameters.
print(neighborhood)

# We initialize any variables we need to (in this case, just x), and then
# executing the model is a single method call.
neighborhood.set_param('x', 2)

neighborhood.run_step()

# Print out information.
for name, param in neighborhood.params.items():
    print(f"{name} = {param}")

Documentation

Documentation for porchlight can be found on Read the Docs here: https://porchlight.readthedocs.io/en/latest/

Other info

• You can find slides from presentations about porchlight within the docs folder, under docs/slides.