Python virtual environments

The following is a quick survival guide for disambiguating Python virtual environments, the tools used to create them and terms such as:

• Distribution
• Installation of a distribution
• Virtual environment
• python and pip
• virtualenv, venv, virtualenv-wrapper, pipenv, pyenv.

Distributions

There exist several Python distributions:

• CPython (official one downloadable from python.org)
• Anaconda
• PyPy
• ActivePython

Each distribution provides a Python interpreter that can execute Python programs (i.e. a distribution provides the python command).

Distribution version

Each distribution supports multiple versions of the Python programming language e.g., Python 3.8 and Python 3.9.

Installation of a distribution

You can have one or more Python distributions installed at the computer at the same time. E.g. you can have CPython 3.7, 3.8, 3.9. Often, your operating system also already ships a default Python installation.

What's important to understand is that each installation of a distribution provides an isolated Python environment. Technically this translates down to having multiple python commands in different paths (e.g. /usr/bin/python or /usr/bin/local/python3.8/python3).

Pay attention to symlinks and aliases, e.g. python might be an alias for python3.

Python distribution tools

The most prime example of a Python distribution tool is pip, which you use to install Python packages from the Python Package Index (PyPI).

The command pip is always tied to a specific installation of a Python distribution. So when you call pip install ..., it will install the Python package into that specific installation of a distribution.

Thus in addition to having multiple python commands, you can have multiple pip commands that belongs to a distribution.

Virtual environments

A Python virtual environment is created from a distribution installation. It isolates the Python packages and commands installed.

For instance, if you have two Python programs A and B, and Python program A depends on version 1 of package X, and program B depends on version 2 of the same package X. Both version 1 and version 2 cannot be installed at the same time in a single Python distribution.

Virtual environments allow you to have a virtual environment for program A and another for program B each with a different version installed of package X.

This means that a virtual environment has its own python and pip commands.

Tools

• pyenv is a tool to manage and install multiple Python distributions.
• venv and virtualenv are tools to create virtual environments for a specific installation of a distribution.
• pipenv, poetry, virtualenv-wrapper are tools used to manage venv and/or virtualenv.
• brew, yum, apt are package manages that can be used to install e.g. multiple Python distributions.
• invenio-cli uses pipenv on your local machine or in containers to create and manage a virtual environment.