Our models follow a design pattern where the model calls a dynamics function either once or twice, depending on whether forecasting is enabled. A ton of parameters are passed into dynamics either as keywords or in a big struct called params. But most of the parameters don't change. I think this would be cleaner if dynamics were a nested function and only parameters that changed between the main period and the forecasting period were passed in.
Our models follow a design pattern where the model calls a
dynamicsfunction either once or twice, depending on whether forecasting is enabled. A ton of parameters are passed intodynamicseither as keywords or in a big struct calledparams. But most of the parameters don't change. I think this would be cleaner ifdynamicswere a nested function and only parameters that changed between the main period and the forecasting period were passed in.