Overview

When working on multiple Python projects that require different packages and libraries, we may run into the issue of needing different versions of packages for each project. To avoid upgrading or downgrading versions each time we work on a different project, we can make use of virtual environments. The core idea of a virtual environment is to create an isolated environment that has the packages a project needs contained within it, that can be easily "turned on" and "turned off" as needed. This topic explains how to set up and manage virtual environments using the Python package venvwrapper.

Advantages of virtual environments

Virtual environments allow us to keep certain combinations of package versions required for each project and access them when we work on the respective projects. This way, it is easy to separate project requirements and also ensure that the project scripts will continue to run even if we make changes to some package versions outside the project. Having virtual environments also helps to keep a clear overview of packages from each project and to easily reproduce workflows, even if it is on the same computer or a different one.

Windows Users

The first step we need to take is to install the package. For this, open the Command Prompt and type:

• bash

pip install virtualenvwrapper-win

Next, we navigate to the folder where we want to create our virtual environments, which can be in the same folder as the script of the project or in a separate folder where we keep all environments. To create a virtual environment type the following in the Command Prompt:

• bash

mkvirtualenv name_of_env

To activate the created environment, or switch between environments, we use:

• bash

workon name_of_env

Now, we have this newly created environment open, so we can add packages with pip install name_of_package.

After we install all the packages we want for our project in the environment, we can create a text file which contains all installed packages and their versions. This is useful for keeping track of the configurations of the virtual environments, but also for easily replicating the configuration. To create this list and install packages from it we use the following commands in cmd:

• bash

#create requirements list of installed packages
pip freeze > requirements.txt

#for replicating environments, we can install all packages from requirements.txt
pip install -r requirements.txt

We can see the installed packages by using one of the following commands in the cmd:

• bash

pip list  
#or
lssitepackages

Equivalently, we can see all the created environments by using:

• bash

lsvirtualenv

We can deactivate the environment and return to the 'basic' one, type in cmd:

• bash

deactivate

To delete virtual environments, we use:

• bash

rmvirtualenv name_of_env

Mac and Linux Users

To configure virtual environments on Mac OS, the commands are somehow similar to Windows, but additionally we need to set up path and virtual environment variable. We start by installing virtualenvwrapper by typing in the Terminal:

• bash

pip install virtualenvwrapper

Next, we need to define environment variables that contain the path to the Python installation folder, as well as a path location for where we want to create the virtual environments. We also need to add the path where virtualenvwrapper was installed.

• bash

export WORKON_HOME=~/Envs
mkdir -p $WORKON_HOME
VIRTUALENVWRAPPER_PYTHON=/usr/local/bin/python3 export 
VIRTUALENVWRAPPER_VIRTUALENV=/usr/local/bin/virtualenv
source /usr/local/bin/virtualenvwrapper.sh

Remember to replace the above path with your own installation location.

Then, we can create a virtual environment.

• bash

mkvirtualenv name_of_env

We can add packages to the environment with:

• bash

pip install name_of_package

We can see all the installed packages in the environment with:

• bash

lssitepackages

To see all the created environments we use:

• bash

ls $WORKON_HOME

To create the requirements.txt file, switch environments, deactivate or delete environments, we use the same commands from Windows.

Commands overview

Other virtual environment packages

Besides venvwrapper, there are other packages that create and manage virtual environments: - virtualenv - venv - conda

See also

• Python Virtual Environments - a Primer