At the moment whole equations are recognized as belonging to a particular boundary, or being a type of interface between regions.
We want a method for seeing whether they are dirichlet, robin, neumann, periodic or interface as many subpackages will rely on these primatives.
This should be robust to different forms, i.e. u(0, t) - 1 ~ 0 and 0 ~ 1-u(0, t) should be recognized as dirichlet BCs, not robin, and equivalent to u(0, t) ~ 1
Should have the answer stored in a few types of struct such as:
struct DirichletBC <: AbstractAtomicBC
"""
The dependent variable of the BC (e.g. u(0, t))
"""
u
"""
The independent variable of the BC (e.g. x), where this is the one held constant
"""
x
"""
Where this bc is defined, unused for now but will be important for exotically shaped domains
"""
support
"""
The RHS
"""
expr
end
At the moment whole equations are recognized as belonging to a particular boundary, or being a type of interface between regions.
We want a method for seeing whether they are dirichlet, robin, neumann, periodic or interface as many subpackages will rely on these primatives.
This should be robust to different forms, i.e.
u(0, t) - 1 ~ 0and0 ~ 1-u(0, t)should be recognized as dirichlet BCs, not robin, and equivalent tou(0, t) ~ 1Should have the answer stored in a few types of struct such as: