TODO

*   improve testing and documentation (always ...), profile changes, identify optimization opportunities
*   add production lookup functionality to Tidal and Wind (lookup tables not quite the same thing as passing a production time series ... certainly less data to read in!)
*   will probably need to de-couple normalized performance data from the assets themselves (or each reset will trigger a new disk read ... which could get expensive fast)
*   implement model predictive control in Controller (means getting linear programming functionality in here ... need to find a good solution for that!)
*   implement Python Controller API (to allow users to define their own control algorithms); this will hurt runtime performance, but greatly increase modelling flexibility
*   implement some kind of pumped hydro storage functionality, either as part of Noncombustion :: Hydro, or by adding a new child class Nomcombustion :: Hydro :: PumpedHydro