Glue code for making DiffEqBase work with units.
This defines
-
how to calculate the vector norm 'ODE_DEFAULT_NORM' when the vector is given in units compatible with Unitfu.jl, from registy M8. The differential equation algorithms expects the norm to be unitless, as can be seen in e.g. step size estimators.
-
type-stable and inferrable 'zero', 'value', 'UNITLESS_ABS2', 'similar'
These functions preserve types for mixed-unit vectors. E.g. [1.0kg, 2.0N, 3.0m/s, 4.0m/s].
This package also uses and reexports 'ArrayPartition' from RecursiveArrayTools, which enables type-stable solution of equations with mixed units. It pre-compiles it with use of mixed unit vectors.
An example of inferrable, mixed units calculations with debugging is included in test/test_5.jl.
Note: The way error tolerances are defined is initially confusing, but good to know:
err_scaled = error / (abstol + norm(u) * reltol)
where bold indicates unitful objects.
Some functions are adaptions of corresponding code from DiffEqBase.