PySavvy

You may have seen the line (usually at the end of a Python script):

if __name__ == '__main__':
    # some other code

And wondered what it does, and why it is there. After reading this you will never wonder why again!

What does it do?

In short, if __name__ == '__main__': only runs the body of the if block if it the file that was directly called (e.g. python my_script.py), not if it is imported (e.g. import my_script).

When would I use this?

Let’s say you have a script that processes some large datasets. You have the parts broken up into separate functions for a clean design, it may look something like this (let’s call it my_model.py):

import numpy as np
import pandas as pd


def read_dataset(dataset_path):
    return pd.read_csv(dataset_path)


def train_model(dataset):
    # training code here...
    return model


def test_model(model):
    # test the model


dataset = read_dataset('my_data.csv')
model = train_model(dataset)
test_model(model)

This works great when it is all in a single file, but let’s say you want to use the read_dataset function in a different file. In your new file you write from my_model import read_dataset and your new file takes a long time to run… this is because when you import the function it is also reading the dataset, training the model, and testing the model!

You can modify my_model.py so that this doesn’t happen, like this:

import numpy as np
import pandas as pd


def read_dataset(dataset_path):
    return pd.read_csv(dataset_path)


def train_model(dataset):
    # training code here...
    return model


def test_model(model):
    # test the model

if __name__ == '__main__':  # added this line and put the following 3 lines in the block
    dataset = read_dataset('my_data.csv')
    model = train_model(dataset)
    test_model(model)

See the difference? Now, when you run python my_model.py the script will read the dataset, train the model, and test the model. But, when you import this file these functions won’t run; however, you can still use the functions!

Why does it work this way?

Every module (i.e. file) in Python has the __name__ attribute set. If the module is the entry point (what you typed in the command prompt) then the __name__ is going to be '__main__'. Otherwise, it is going to be the name of the file and any packages that it is a part of.

For example, if you have the following directory structure:

my_python_package/
    __init__.py
    cool_subpackage/
        __init__.py
        super_cool_module.py

And you have import my_python_package.cool_subpackage.super_cool_module the __name__ inside super_cool_module will be my_python_package.cool_subpackage.super_cool_module.