1. What is it?

Virtual environment (virtualenv in short) is a common way of managing installed packages in your Python project.

2. Why do I need it?

Using virtual environments allows you to separate your projects from each other. Packages installed in one environment are not visible in other projects. Therefore, you can handle using different versions of various libraries in your projects. Additionally the environment is lighter if it contains only the packages that are used.

3. How to use it?

First and foremost – in a terminal/command prompt.

Installation

If you don’t have virtualenv package installed, open your terminal/command prompt and run:

pip install virtualenv

Initialization

In your directory of choice (execute: cd path/to/directory) create the environment. Depending on how you installed virtualenv, you can try one of the following:

1
2

virtualenv my_venv
python -m venv my_venv

Note that a conventional name for a virtual environment is venv, not my_venv

Activation

After creating a new virtualenv it needs to be activated first. To activate it, run:

1
2
3

source my_venv/bin/activate  # for Linux or Mac

.\my_venv\Scripts\activate  # for Windows

After activation you’ll see name of the environment in brackets at the beginning of line in terminal:

1

(my_venv) your_user:path/to/directory$

If you want to deactivate the environment, just run deactivate.

Installing packages

Only after activating the environment you can install packages to it. Do it using pip install, e.g.:

1

pip install numpy

All libs you install are downloaded to your virtualenv in my_venv/lib directory.

Freezing virtualenv

If you want to share the project, don’t send my_venv folder to anyone. Instead, you need to freeze your environment by saving its state to a .txt file, conventionally named requirements.txt:

pip freeze > requirements.txt

Now anyone can recreate the environment by creating a new one first and then installing everything from requirements.txt:

pip install -r requirements.txt

4. Summary

The workflow of virtual environments is the following:

1
2
3
4
5
6
7
8
9
10

virtualenv venv  # create a virtual environment named "venv"

source venv/bin/activate  # activate venv on Linux or Mac
./venv/Scripts/activate  # activate venv on Windows

pip install <package_name>  # install required packages

pip freeze > requirements.txt  # dump state of your environment to a .txt file

pip install -r requirements.txt  # recreate the environment from file (after creating a new one)

Alternatives

There are some alternatives for using virtualenv which are worth checking:

  • Anaconda – used mainly in data science, allows you to store all you need in a one big environment which can be separated from the pure Python interpreter on your machine
  • Poetry
  • Pipenv