Finding Modules and Packages

Python most times is simple but sometimes it appears complex and even complicated when your modules or packages just won’t import.

Usually you will get an error:

ImportError: No module named 'xxx'

Before we get over this hurdle let’s take a step nback and understand some of the fundementals

Fundmentals of Modules and Packages

  • A module is a single python file
  • A package is a folder containing python files along with a that tells python it is a package to import files from

Remember all these files have .py extensions but importing them with the .py is not needed

Where does python look for modules and packages

Python looks in the environment variable: $PYTHONPATH


Printing out the environment variable does not work for me:


To find out the directories where python is looking for modules and packages use:

import sys

How do I add a directory to the PYTHONPATH

One way is in your file:

import sys
sys.path.insert(0, "/path/to/your/package_or_module")

Another way is to add it to your environemtn variables permanently:

Add the following line to your ~/.profile or ~/.bashrc file.

export PYTHONPATH=$PYTHONPATH:/path/you/want/to/add

But this method is not the best because usually you will be in a virtual environment for a specific project and don’t want to globally set your environment

Finding where a module is coming from

The module __file__ attribute

Say you have imported a module but you are unsure where it is coming from you can use the:


which will output the location of the directory that module is coming from

Note: The file attribute is not present for C modules that are statically linked into the interpreter

Recent changes to PYTHONPATH

The current working directory is not added if PYTHONPATH is empty

An empty PYTHONPATH was equivalent to setting it to . previously.

How to add the current working directory in the python path without modifying the package

SAy you are getting import errors and need to add the current working directory to the PYTHONPATH but because this is a package that others are going to use you do not want to use the methods above.

You can set the python path in your virtual environemnt:

vim /env/bin/activate

and add:

export PYTHON_PATH='/my-module/working/dir'

Note this only works with an activated pipenv shell and not when riunning pipenv run