Class 9 - extension modules

Python code at the end of class: class9_inclass.zip.

1. We wrote in C a simple extension module that calculates n-th Fibonacci number and compiled it with setuptools.
2. We covered ways to speed up Python code by using Cython, nanobind, nanobind with Eigen3 library, as compared to Python with NumPy and pure C code. Other C++ libraries providing multidimensional array object could also be made to work with nanobind relatively easily, since it provides a data pointer and shape, stride information necessary to wrap it in an object.

To compile nanobind modules we used cmake -S . -B build and cmake --build build commands and then copied the resulting dynamic library from build directory to src/c_for_speed directory. To compile cython modules we used python setup.py build_ext --inplace command and then copied the resulting library to src/c_for_speed directory. To make use of these modules, we needed to install the package once in editable mode with pip install -e ..