Reduce variability in manipulation of small to medium sized collections across all my node.js projects. Can justify spending time on ad-hoc pipelines and stuff when collections become very big. However, until that actually happens, this should suffice.
Here is a simple model of how things are organised:
Keep it fully tested so it can be used across production projects as well as prototypes.