I propose concept where we will not require to create dimensions objects on TrajectoryData creation. The dimension constructors will be evaluated on access. Also it allow to use Any raw object with expected fields (like Named tuple, Cython or C -struct).
So it can increase productivity for the purposes when you need no unit conversions after integration process or you need no some data to be evaluated.
Also we can add caching for fields that was evaluated once
I propose concept where we will not require to create dimensions objects on TrajectoryData creation. The dimension constructors will be evaluated on access. Also it allow to use Any raw object with expected fields (like Named tuple, Cython or C -struct).
So it can increase productivity for the purposes when you need no unit conversions after integration process or you need no some data to be evaluated.
Also we can add caching for fields that was evaluated once