The examples on this site are compatible with Mac OS X, Linux, and Windows. To run these examples you will need various C/C++ and Python software installed. The instructions here cover the prerequisites necessary to run the Cython, SWIG, and CFFI examples.
To run most of the examples you need a C compiler toolchain and for some of them you also need a C++ compiler toolchain. How to get this installed depends on which operating system (OS) you are running.
If you are using any of Linux OSes which use apt for a package manager (Debian, Ubuntu, or Mint), then you can do the following to get all of the C and C++ tools you need:
sudo apt-get install build-essentialFor other Linux package managers, the fundamental tools you want to install are gcc and g++.
Install the free XCode developer tools from the App Store. Make sure you run it at least once and say "yes" if it asks you if you want to install the command-line tools.
There are a number of options for Windows, but by far the simplest is to use the same compiler which Python itself was compiled with. For recent versions of Python released within the past couple years, this is the Visual C++ compiler which comes with Visual Studio 2017.
It is relatively easy to use both SWIG and Cython on Windows because all C/C++ compilation can coordinated within the setup.py file and handled either by the setuptools or distutils Python module. In this case, Python will automatically know where all of the correct libraries to link against are.
Using CFFI can be a little bit more finicky since it is directly calls into pre-compiled dynamic libraries and makes some assumptions regarding the C ABI. So you really need to make sure you are using the exact compiler that Python itself was compiled with.
See the WindowsCompliers section of the Python Wiki for more information.
If you try to use the examples on Windows and run into difficulties, you may wish to setup a Linux virtual machine (VM) for evaluation purposes since C/C++ compilers are a lot easier to setup on Linux. We recommend using VMware Player or Virtualbox to install a Linux Virtual Machine (VM).
Pipenv is an easy way to install compatible versions of cffi, cython, and
pybind11. Once pipenv in installed, these other packages can be installed in a virtual environment using:
pipenv installThen this Python virtual environment can be entered using:
pipenv shellAnother easy way to get all of the tools installed and setup is by using the Anaconda Python
distribution. This is a free distribution of Python available for Windows, Mac OS X, and Linux which
comes with a couple hundred of the most common and useful Python modules pre-installed including Cython
and CFFI. It also comes with the conda package manager which can be used to install additional
tools such as SWIG.
Download the latest version of Anaconda. We recommend Python 3.6 or newer.
Install Anaconda per the instructions and let it modify your .bashrc, .bash_profile, or path accordingly.
NOTE: Do not use sudo to install Anaconda, just install it as a normal user.
This also installs Cython and CFFI.
If Anaconda is installed, you can use the included conda package manager to install SWIG. Do the following in a command-line terminal:
conda install swigAlternatively, on macOS you can use the Homebrew package manager to install SWIG:
brew install swigPyPy is best installed with a package manager such as brew on Mac OS X or apt-get* on many Linux distros.
Use the Homebrew package manager to install
brew install pypyUse your the package manager which comes your distro to install pypy. For example, on Debian, Ubuntu, or Mint distros:
sudo apt-get install pypyDownload a Windows binary from the PyPy website: http://pypy.org/download.html