Using Python virtualenv

Why We Need virtualenv

In many cases the Python environments used by projects differ — not only between Python2 and Python3, but also between library versions; for example BeautifulSoup has versions 3 and 4. After installing Python on our machine, sometimes we don’t want our Python environment polluted, and we also want the local environment to match the server environment, including the Python version and library versions.

Different languages also provide different “container” or “package management” forms. Package management: Python can use requirements.txt, Node has npm’s package.json, Java has Maven and Gradle — these package managers make environment setup especially simple.

Containers: Python has virtualenv, Node has npm’s node_modules, and there is also the recently popular Docker — these containers isolate the local environment from the development environment.

Whether in “container” or “package management” form, the goal is ultimately to unify code development or environments to some degree. From a slightly higher perspective, these tools exist to simplify development. Don’t put the cart before the horse — code quality is what really matters.

Installing and Basic Usage of virtualenv

• There are many ways to install virtualenv. I use Archlinux, where it can be installed simply via pacman or Python’s pip. Other Linux distributions have their own methods; worst case you can install from the source package.

• Using virtualenv

virtualenv2 venv    # create a Python2 environment; virtualenv3 creates a Python3 environment

After running this command, a venv folder appears in the current directory; check it with ls -al.

• Activate the environment

source venv/bin/activate

This command updates the current Python environment. After running it,

you can see the (venv) marker in front of PS1, showing that we are now using the newly created environment. At this point, run the python command to see that the interactive shell environment is now different.

Using virtualenv Together with requirements.txt

If virtualenv provides the ability to quickly create a clean Python environment, then requirements.txt provides the ability to quickly import a development environment.

• Using requirements.txt

  We use requirements.txt to let pip handle importing libraries. After getting a project’s source code,

if it contains a requirements.txt, we can import the libraries like this:

pip install -r requirements.txt

Of course, preferably inside an environment already created with virtualenv.

• Creating requirements.txt

When we start a new project we always use some libraries. In a virtualenv environment, the libraries in our environment are exactly what we need, so export them like this:

pip freeze > requirements.txt

After exporting, a requirements.txt file is created in the current directory, similar to:

appdirs==1.4.0
packaging==16.8
pyparsing==2.1.10
six==1.10.0