Build a local Python package

Data engineering

Author

Kedar Dabhadkar

Published

May 23, 2021

Build a Local Python Package

How do you ensure that you reduce redundancy when writing Python code? Think about that Matplotlib code that you once wrote to add a cool new element to your charts, or the number of times you had to copy-paste the server name of your database. Did you know that you can avoid that repetition by installing a locally built Python package across your environment?

Let me demonstrate how you can do that using just the most basic code required.

Let’s build a Python package named ‘toolbox’ and include a function to find the roots of a quadratic equation.

Step 1. Build this directory structure

toolbox/
- setup.py
- toolbox/
  - __init__.py
  - roots.py

Step 2. Write your Python function

Populate roots.py with your repetitive code. In my case, the function to find out roots of a quadratic

def get_roots_quadratic(a, b, c):
    """
    Returns the roots of a quadratic equation using the coefficients.
    """
    root1 = (-b + ((b ** 2) - (4 * a * c)) ** 0.5) / 2 * a
    root2 = (-b - ((b ** 2) - (4 * a * c)) ** 0.5) / 2 * a
    
    return root1, root2

Step 4. Populate setup.py

From the root directory, edit setup.py

import setuptools

setuptools.setup(name='toolbox',
                version='0.0.1',
                descrption='Save frequently used scripts',
                packages=setuptools.find_packages(),
                zip_safe=False)

Step 5. Install your new package

Go to the root directory and install your package

pip install .

Use your package

from toolbox import roots

get_roots_quadratic(1, -2, 1)

(1.0, 1.0)