1. brew install pyenv
2. Open ~/.bashrc (or .bash_profile, on Linux) and add the following between the stuff for Rbenv and Git:

# Pyenv
export PYENV_ROOT=/usr/local/var/pyenv
eval "$(pyenv init -)"

1. pyenv install 3.5.1 (or other version)
2. pyenv global 3.5.1 (this can be combined with the previous command as pyenv install 3.5.1 && global 3.5.1)
3. Python doesn't ship with the most up to date version of package manager pip, so upgrade pip : pip install -U pip
4. Install pyenv-virtualenv, a tool that creates a named directory of executable dependencies and makes that directory available at the global level. Think of it this way:

• With Ruby and Bundler, bundle install install all gems listed in the Gemfile into a global bucket (technically, subdivided by version of Ruby) which is then accessible to all projects; running bundle exec ensures that you execute the right versions of those dependencies for any given project.
• With Node and NPM, npm install installs all Node moules listed in package.json into a local node_modules directory; since there's a separate node_modules folder inside each project, all executables are implicitly project-specific. However, this has the downside of downloading lots of copies of the same Node modules.
• With pyeyv-virtualenv, you have a single global directory (bounded by Python version), just like Ruby; however, inside that global directory, you can create custom bundles of dependencies (i.e. 'virtual environments') and load a single bundle of dependencies inside multiple different projects. Kind of like lots of little node_modules directories, but with different names, and capable of being referenced by name and manually loaded.

1. Install mamba and expects : pip install mamba expects

When you start a new python project, load up the appropriate requirements file (e.g. requirements-test.txt) by running pip install -r requirements-test.txt

To actually run your mamba tests, go to the root of your project repo and run mamba; it will look for a specs directory.