Allow different precisions for spline calculations

[EmilyBourne]

Currently spline calculations are hard-coded to use double precision. It would be nice to make this a template parameter so lower precision can be tested

[tpadioleau]

Do you think if we use the build option DDC_BUILD_DOUBLE_PRECISION along with the type ddc::Real, that could be enough ? Or do you need a finer control ? This option is meant to configure the floating point type of DDC discrete dimensions such as UniformPointSampling...

[EmilyBourne]

This could be a way to test, but it is possible that we may wish to use different precisions in different operators. That said there are not many operators in DDC so maybe this is sufficient 🤔

[EmilyBourne]

Doubles are used for the coordinates though and what you suggest would be a cleaner way to mess with this