PyDMFT: A library and a reproducible science case study

PyDMFT is a scientific library written in Python with Cython extensions and use of Numba for speed. It has been developed as a simple toolset for research on interacting electronic systems. It keeps data abstractions to the minimum thus making it easily compatible with NumPy arrays, easy to serialize and store data. Beyond the scope of the library, PyDMFT offers a large collection of python scripts that by using the library are capable of reproducing results presented on peer reviewed journal articles. This collections of research recipes are executed, organized, and displayed in a web gallery using the package Sphinx-Gallery. This large archive allows to find the specific implementation of a research work, makes it easy to recall how to execute a script as all the runtime data and setup need to be available for the gallery to be generated. As the library evolves, new builds of the gallery make it clear if examples break due to change in the library.