Levi Velázquez

Virtualenv and Virtualenvwrapper tutorial

Virtualenv is a helpful tool to create isolated Python environments. So, inside those environments you can create your own projects and install it’s python packages and dependencies without affecting your system’s site-packages. Also, you can control packages versions for each project and much more.

We are going to need pip , if you are on Mac OS X, you can installing via easy_install:

$ sudo easy_install pip

Installing virtualenv and virtualenvwrapper

Install virtualenv via pip:

$ sudo  pip install virtualenv

We are going to use also virtualenvwrapper, a tool to create/delete virtualenvs in an easier way. This because virtualenv itself is too verbose and a bit complicated.

Install virtualenvwrapper via pip:

$ sudo pip install virtualenvwrapper

Let’s create where our virtual environments would live. NOTE: our project folder does not need live in this folder too. We can set it anywhere.

$ mkdir ~/.virtualenvs

You can name it as you want, I rather prefer to use virtualenvs.

If you are using standard shell, open your ~/.bashrc or ~/.zshrc if you use oh-my-zsh. Add this two lines

 
export WORKON_HOME=$HOME/.virtualenvs  
source /usr/local/bin/virtualenvwrapper.sh

Starting a new virtual environment

Let’s create a virtualenv called myenv

 
$ mkvirtualenv myenv
$ > New python executable in myenv/bin/python
$ > Installing setuptools, pip, wheel...done.
(myenv)$

You will notice a change in your command prompt, it will contain the virtualenv name that indicates that it has been activated. Now, all the python packages that your install without using sudo and having your env activated will install inside virtualenv own site-packages.

Our new env is located at ~/.virtualenvs/myenv

Now, we can create our project folder, It does not matter where is located, we just need to have our virtualenv activated and no more.

 
(myenv)$ mkdir ~/ourproject
(myenv)$ cd ~/ourproject

Specifying a python version

By default, python virtualenv executable is took from python system version, if you want to specify your own version, add --python=your_python_path. If you don’t know your python path, use which command:

 
$ which python3
$ > usr/local/bin/python3
$ mkvirtualenv --python=usr/local/bin/python3 myenv

For more information, read virtualenvwrapper docs

Shutting down an enviroment

In order to deactivate your virtualenv:

(myenv)$ deactivate

Activating an environment

To activate an existing virtualenv, use command workon:

$ workon myenv
(myenv)$

Binding our virtualenv to a project path

If we want to bind our virtualenv to specified folder in order to each time that we activate it, that folder will become in our current directory, let’s use setvirtualenvproject [virtualenv_path project_path] command.

We need to have our virtualenv activated, go to our project folder and set it.

 
(myenv)$ cd ~/ourproject
(myenv)~/ourproject/$ setvirtualenvproject $VIRTUAL_ENV $(pwd)

Now each time we activate our virtualenv using workon, ~/ourproject would be our current directory.

 
$ workon myenv
(myenv)~/ourproject/$

Testing our virtualenv

In order to test our virtualenv, let’s install Django.

(myenv)$ pip install django
> Collecting django
>   Using cached Django-1.10-py2.py3-none-any.whl
> Installing collected packages: django
> Successfully installed django-1.10

Run a Django command, in this case, let’s try to create a new project.

 
(myenv)$ django-admin.py startproject testproject
(myenv)$ ls
> testproject

Our command just worked as expected. Now we are going to deactivate our virtualenv and try to create a new project.

 
(myenv)$ deactivate
$django-admin.py startproject testproject
> command not found: django-admin.py

We can be sure that our Django library is only available inside our new virtualenv.

Please, share if you found this post useful :)

Check out my others posts here…