Modules and Imports

Python has a simple package structure. Any directory with a file named __init__.py can be considered a Python module.

Note: a __init__.py file is no longer required for Python 3 modules, but it’s still supported and can be useful.

For example, let’s make a basic function and start a new module to house it:

def add_numbers(x, y):
    return x + y

Let’s put our function in a file called __init__.py and place it in a folder called my_math_functions. Now, as long as the folder my_math_functions is somewhere in our PYTHONPATH, we can import my_math_functions and reach add_numbers(). If we start our REPL from the folder that contains my_math_functions, we can import it:

>>> import my_math_functions
>>> my_math_functions.add_numbers(1, 2)
3

Best Practices

There a few different ways to import modules or even just specific objects from modules. You can import everything from a module into the local namespace using *:

>>> from my_math_functions import *
>>> add_numbers(1, 2)
3

This isn’t a good practice, because it’s hard to tell where a specific function is coming from without the namespace context. Also, function names can conflict, and this can make things very difficult to debug.

Better is to import functions specifically:

>>> from my_math_functions import add_numbers
>>> add_numbers(1, 2)
3

This make things a little clearer, as we can look at the top and see where the add_number() function came from. However, an even better way is to just import the module and use it in calls to maintain the namespace context:

>>> import my_math_functions
>>> my_math_functions.add_numbers(1, 2)
3

This can be slightly more verbose, but unless it makes your function calls ridiculously long, it generally makes things much easier to debug.

You can use the as keyword to make things a little easier on yourself.

>>> import my_math_functions as mmf
>>> mmf.add_numbers(1, 2)
3

PYTHONPATH

What is the PYTHONPATH we mentioned earlier? This is a list of paths on your system where Python will look for packages. Python will always look first in the working directory (the folder you’re in when you start the REPL or run your program), so if your module folder is there, you can import it. You can also install your modules just like any other external modules, using a setup.py file. It’s also possible to change or add paths to your PYTHONPATH list if you need to store modules elsewhere, but this isn’t a very portable solution.