Python pipenv tutorial: manage dependencies like a pro

When you're working with Python, you need to install packages and manage dependencies. If you have several projects, dependencies can clash and cause problems. I'll tell you how pipenv solves this issue.

Introduction

When you're working on a Python project and need to install a package, you typically run python3 -m pip install <package>. What happens if two projects depend on different versions of the same package? Problems can arise.

This is where "virtual envs" or "venv" were created, allowing you to have different versions of Python with their respective packages on the same computer.

In this case, pipenv is a quite popular tool that allows you to manage Python virtual environments on your computer, similar to npm in the Node ecosystem.

I'll explain the basic concepts so you can use it too.

Tutorial: how to use pipenv

First, install pipenv with the following command:

python3 -m pip install pipenv

With this, you can use the pipenv command to manage environments. Keep reading!

In my case, it didn't work because the python folder of my user is not in the path. Just run the command python -m site --user-base and add the path it provides and add /bin at the end, to your PATH variables, in my case: export PATH="/Users/alberto/Library/Python/3.9/bin:$PATH".

Now, the basic commands you need to know:

Install dependencies

This command creates a virtual environment if one doesn't exist, so you can launch it directly.

pipenv install <package>

Activate a virtual environment

pipenv shell

Once inside, you can run:

python file.py

Run a file

pipenv run python file.py

Create a virtual environment

pipenv --python <version>

Delete a virtual environment

pipenv --rm

Virtualenv

There's another lower-level tool called virtualenv that allows you to create directories containing all the dependencies to run the project with a specific python version.

Here's a link with all the documentation for the tool.

Summary

In my case, I use pipenv when needed since it's the most convenient. This reminds me of when I started working where we had different projects, all with Python, virtualenv, and docker.