Python Virtual Environment Management

virtualenv: create isolated python environments

Usually you don't want to mess your system python with project dependencies, you can use this to create an isolated python environment, the new environment has its own bin, lib, include, it also has its own pip and setuptools packages, all other packages will be installed inside its own folders.

Create and Activate virtualenv

Run virtualenv YOUR_NEW_ENV_DIR to create a new environment. To use or work with such isolated environments, usually you need to tweak your PATH environment variable, it give it precedence over system default python. But you don't need to do it manually, each virtual environment has a bash script inside its own bin directory, with all settings necessary to activate it, all you need to do is to run source bin/activate.

Deactivate and Remove virtualenv

Run deactivate to deactivate current virtualenv. Run rm -r /path/to/your_virtualenv to delete the virtualenv folder.

To make it easier to use, you could use some tools like virtualenvwrapper.

virtualenvwrapper: virtualenv extension and utility

Basically, virtualenvwrapper provides utilities to do virtualenv and project management, the most important ones include: - workon: switch between different virtual environments, setup path and environment variables - mkvirtualenv: create a new python environment - mkproject: create a new python project directory, link to a virtualenv

It also has tools like: cdvirtualenv, cdsitepackages, setvirtualenvproject, cdproject, lsvirtualenv and many others. It create all your python einvironment in the same place like ~/.virtualenv(this is customizable). It also manage all your python projects inside a project root directory(customizable). Since I always use the latest python3.x for my own projects, I don't use this tool a lot.

pyenv: Pure shell utility to manage python versions

pyenv is something different, it is pure shell script, no python dependencies. It can install different versions and flavors of python, which could be grouped like this:

python 2.x
python 3.x
pypy2
pypy3
stackless
miniconda
anaconda
ironpython
jython

it will download the specified python installation files from official websites, and install them to ~/.pyenv(this is customizable). To use a specific python version or flavor for your specific project, you simply either put a .python-version file inside the project root folder, when you enter this project folder, pyenv will get the version specified in this .python-version file, and switch to use the right binaries and packages, if you .python-version file contains 3.0.1, then python will be using python3.0.1, it is really handy. You could even override system default python by setting PYENV_VERSION in your .bashrc or put a .python-version file in the root the filesystem.

pipenv: Sacred Marriage of Pipfile, Pip, & Virtualenv

Another new virtual environment management tool, written in pure python, so support for Windows should be better.

tox: A virtualenv manager and a test utility

tox aims to automate and standardize testing in python. It does basically only three things: - create virtualenv - install dependencies like pytest, and your package under test - run the tests it is extremely useful if you want to test your code against many different versions or flavours of python.