Streamlining an IRAF data reduction process with AstroPy and NDMapper

In the course of re-writing my typical top-level GMOS-IFU data reduction sequence in Python for a research project, I have developed a small module that helps express the scientific process in a relatively intuitive way as a Pythonic series of operations on NDData collections, mapped to files, with existing IRAF steps integrated almost seamlessly (pending their eventual replacement). For scientific end-user purposes, this experiment aims to obviate a need for pipeline machinery, favouring simple control flow in the main script and retaining a smooth transition from high-level process description to lower-level libraries by encapsulating necessary bookeeping within the data representation and simple wrappers. The I/O abstraction should make support for file formats other than FITS (eg. ASDF) straightforward to add. This work-in-progress can be found at https://github.com/jehturner/ndmapper and I intend to split its functionality involving IRAF or instrument processing into a separate "ndprocess" module as the prototype nears completion, leaving a core "ndmapper" package, without any special dependencies, as a general add-on for nddata.NDMapper is "a Python module for constructing intelligible astronomical data reduction processes, based on an abstraction that maps files on disk to collections of AstroPy NDData objects". This module currently includes limited functionality for processing data from the Gemini Multi-Object Spectrographs