Defining your own modules#

Every Python file defines a module. We’ve seen modules already by way of the import statement, which tells Python that we’d like to use someone else’s code. There’s a lot to learn about modules— check out the Python documentation. We’ll learn two things:

  • how to define your own modules

  • where Python looks for modules

Defining a module is as simple as defining a file—if you create one file, then you can run import foo in another file to load in foo’s definitions.

So, if you have a file

# this is

def fact(n):
    res = 1
    while n >= 2:
        res = res * n
        n = n - 1
return res

If you create a file in the same directory, you can reference util in a few ways. You can use import and then reference util’s definitions using the . operator:

# this is (version 1)
import util


our_fact = util.fact # you can bind modules' definitions locally


Or you can directly import a definition from foo:

# this is (version 2)
from util import fact


Module resolution#

We saw that a file named resolved to a module util. But we’ve used other modules—like math or random that don’t have files in the same directory. What’s going on?

Python resolves modules in one of three ways:

  • Relative to the directory of the initial Python script (or the directory where python was run).

  • Relative to colon-separated directories in the PYTHONPATH environment variable.

  • Relative to the global “site packages” and user site packages directories, which is globally specified for a given installation.

(There are also some sneaky, built-in packages that are part of Python itself.)

You can read more about these details in the Python documentation.