How to Get the Bits of Python You Need Where You Want Them

Following up my recent short video on installing Anaconda I now introduce miniconda, which is a distribution that contains only Python and the conda software on which Anaconda is based. The remainder of the Python packages you need can be installed in virtual environments managed by conda.

When you are building Python systems with mundane requirements this is a great way to put your environments together. From my experience so far it's professionally built and unusually well engineered. Kudos to Continuum Analytics for getting this stuff out.

How to Get Almost All the Python You Might Need

I am often asked what is the easiest way to build Python environments. For development purposes it's most convenient to have a "batteries included" environment that can be pared down once development is over. Continuum Analytics have for a while offered an open source project called Anaconda to provide easy access to both Python and the batteries.

If you are a Python professional and you get questions from people who would like to try the language I would suggest that the graphical installer would be most appropriate. For anyone familiar with the command line, the command-line installer is not only slightly smaller but allows the scripted creation of disposable Python environments which you can activate just by adding their bin subdirectory at the start of your path.

The video should explain this in detail.

In the video I create a disposable environment under /tmp. Anaconda environments are far heavier than a standard virtual environment, but they offer a huge variety of Python libraries without ever having to understand the complexities of package installation.

[For a more selective installation utility see How to Get the Bits of Python You Need Where You Want Them]