Implement Stacey boundary condition

[SouzaEM]

Implement Stacey boundary condition for isotropic elastic wave propagator

[SouzaEM]

With the implementation of Stacey boundary condition, spyro will have three options of boundary conditions (correct me if I'm missing other implementations that should be addressed):

• PML
• Clayton-Engquist A1
• Stacey

I think we need to define how the absorbing boundary condition should be structured to allow configuring each of the boundary conditions in a clear way for the user. My initial proposal is to introduce a nested dictionary structure to separate parameters that are specific to a absorbing boundary condition method. For example:

d = {
    "absorving_boundary_conditions":  {
        "damping_type": "PML", # or local (we may introduce other options later, such as hydrid)
        "PML": {
            "exponent": 2,
            "cmax": 4.7,
            "R": 1e-6,
            "pad_length": 0.25,
        },
        "local": {
            "type": "Stacey", # or CE_A1 (Clayton-Engquist A1) - in the future we may also add Higdon, Lysmer-Kuhlemeyer, etc for completeness
            "dt_scheme": "backward", # or "central" (finite-difference scheme for velocity calculation)
        },
}

Do you have any thoughts?

PS: the key "absorving_boundary_conditions" is misspelled in the code. I can fix it during the development of this issue.

[Olender]

I agree with your nested dictionary approach for structuring boundary condition parameters; this layout seems clean and flexible, particularly for introducing new boundary options in the future.

Thanks for catching the spelling issue as well!