I've been getting more and more questions about using Python's virtualenv with python27 and python33 collections from RHSCL, so I decided to write a very short tutorial about this topic. The "tl;dr" version is: everything works perfectly fine as long as you remember to enable the collection first.

Update 2018: An updated article has been published, See How to install Python 3, pip, venv, virtualenv, and pipenv on Red Hat Enterprise Linux.

What is Virtualenv

Citing Virtualenv official documentation: "virtualenv is a tool to create isolated Python environments". In short, Virtualenv allows you to setup multiple runtime environments with different sets of Python extension packages on a single machine. Unlike Ruby's RVM (Ruby Virtual Machine), it can't install the language interpreter itself - just the extension libraries.

When you create a new virtual environment "foo", a few things happen:

  • The "foo" directory is created with a few subdirectories: bin, lib, lib64 and include.
  • The bin directory contains python, pythonX and pythonX.Y executables. These are basically aliases for the system Python interpreter executable. This directory also contains activate script (in few variants for different shells) - this is used to activate the environment in the current shell session.
  • Extension packages are installed into the lib directory, lib64 is a symlink that points to lib.
  • Python header files are located in include/pythonX.Y, which is a symlink that points to the include directory of system Python installation.

Creating a Virtual Environment

Creating a virtual environment is easy and it works in the same way for both python27 and python33 collections. Both of these collections contain python-virtualenv RPM, so the only thing you need to do is install the desired collection with yum: yum install python27 or yum install python33. I'm going to show an example using the python33 collection:

# run scl-enabled shell and create the virtual environment
scl enable python33 bash
virtualenv foo
cd foo
source bin/activate

# test your virtualenv by installing Django and printing its version
pip install django
python -c "import django; print(django.__file__)"

# now just run "deactivate" to deactivate the environment
# in current shell session
deactivate
# or just "exit" the current shell, which both terminates
# the virtual environment and SCL-enabled shell
exit

The first four instructions above are all that you need to do to create and activate your virtual environment - the rest of the lines just demonstrate that the environment works properly by installing Django and printing the location from where it was imported. If you have ever worked with Virtualenv before, you've probably already noticed that the only difference is that an SCL-enabled bash was run first, all other steps stay the same.

Wrap Up

The only thing you need to remember is to run "scl enable pythonXY bash" before activating the virtual environment. This is the only difference from working with non-SCL Virtualenv. Another nice thing is, that exactly the same commands work for both python27 and python33 collections from RHSCL. I also recommend creating virtual environments with --system-site-packages option, which will allow you to import RPM packaged modules from RHSCL collection.

And that's all you need to know to work with RHSCL Virtualenv.

 

Update 2018: An updated article has been published, See How to install Python 3, pip, venv, virtualenv, and pipenv on Red Hat Enterprise Linux.

Last updated: February 22, 2024