Python Virtual Environments


title: Virtual Environments

Virtual Environments

Virtual environments can be described as isolated installation directories. This isolation allows you to localize the installation of your project’s dependencies, without forcing you to install them system-wide.

Imagine you have two applications App1 and App2. Both require the package Pak, but each requires a different version. If you install Pak version 2.3 for App1, you would not be able to run App2, because it requires version 3.1. Here is when virtual environments come in handy.

Benefits:

  • You can have multiple environments, with multiple sets of packages, without conflicts among them. This way, different projects’ requirements can be satisfied at the same time.
  • You can easily release your project with its own dependent modules.

Here are three ways you can create Python virtual environments.

Tools used for creating Python virtual environments

Libraries

1. Virtualenv

virtualenv is a tool used to create isolated Python environments. It creates a folder which contains all the necessary executables to use the packages that a Python project would need.

You can install it with pip:

pip install virtualenv

Verify the installation with the following command:

virtualenv --version

Create an Environment

To create a virtual environment use:

virtualenv --no-site-packages my-env

This creates a folder in the current directory with the name of the environment (my-env/). This folder contains the directories for installing modules and Python executables.

You can also specify the Python version you want to work with. Just use the argument --python=/path/to/python/version. For instance, python2.7:

virtualenv --python=/usr/bin/python2.7 my-env

You can list the available environments with:

ls virtualenv

Activate an Environment:

Note : On Windows, activating a virtual environment requires the user to have the permission to run scripts.

Before you can start using the environment you need to activate it:

For Mac OS or Linux systems :

source my-env/bin/activate

And for Windows:

.\my-env\Scripts\activate.bat

This ensures that only packages under my-env/ are used.
You will notice that the name of the environment is shown on the left of the prompt. This way you can see which is the active environment.

You can install packages one by one, or by setting a requirements.txt file for your project.

pip install some-package pip install -r requirements.txt

If you want to create a requirements.txt file from the already installed packages, run the following command:

pip freeze > requirements.txt

The file will contain the list of all the packages installed in the current environment, and their respective versions. This will help you release your project with its own dependent modules.

If you are done working with the virtual environment you can deactivate it with:

deactivate

This puts you back to the system’s default Python interpreter with all its installed libraries.

To delete an Environment simply delete the environment folder.

2. venv

venv is available by default in Python 3.3 and later.

To create a virtual environment use:

python3 -m venv venv

3. pyenv

pyenv simple Python version management based on rbenv.
Used together with pyenv-virtualenv plugin
Create an Environemnt:

To create a virtual environment use:

pyenv virtualenv venv34

4. pyvenv

Deprecated in Python 3.6.

Dependency managers

1. Pipenv

Pipenv manages dependencies on a per-project basis. It is like Node.js’ npm or Ruby’s bundler.

pipenv automatically creates and track packges in your virtual environments. It combines pip and virtualenv together so you do need to use them separately.

You can install it with pip:

pip install pipenv

Verify the installation with the following command:

pipenv --version

Create an Environment

To create a virtual environment, use pipenv and specify which version of Python you want to use:

pipenv --python 3.6

Install Packages

Virtual environments created by pipenv do not need to be activated. Each project is managed independently. You simply go inside the project directory and install the packages you want:

pipenv install some-package

Pipenv will install the packge and create a Pipfile inside the directory. The Pipfile tracks all packages and dependencies that your project needs. Pipfile is the replacement of requirements.txt when you use pipenv.

Uninstall Packages

To uninstall a package and remove it from Pipfile use:

pipenv uninstall some-package

If you want to uninstall all packages use:

pipenv uninstall --all

Import requirements.txt to Pipfile

You can import the packages and dependencies from requirements.txt to your Pipfile using:

pipenv install -r requirements.txt

Executing Python Scripts

To run Python scripts use:

pipenv run python main.py

pipenv run makes sure that packages installed with pipenv will be available to your program. You do not need to activate your virtual environment to do so.

Remove an Environment

If you want to remove the virtual environment from pipenv use:

pipenv --rm

You can also delete the environment by deleting the project directory.

Separate software stacks

1. Conda

Conda is a package, dependency and environment management for many languages, including Python.

To install Conda, follow these instructions.

To create a virtual environment use:

conda create --name my-env

Conda will create the corresponding folder inside the Conda installation directory.

You can also specify which version of Python you want to work with:

conda create --name my-env python=3.6

You can list all the available environments with:

conda info --envs

Before you can start using the environment you need to activate it:

source activate my-env

Install packages like with virtualenv.

If you are done working with the virtual environment you can deactivate it with:

source deactivate

If you want to remove an environment from Conda use:

conda remove --name my-env

Additional Resources:

This article needs improvement. You can help improve this article. You can also write similar articles and help the community.