SBP-DG

[cmikida2]

Attempts to implement the SBP-DG coupling methods of Kozdon and Wilcox, starting with modification of the advection symbolic operator.

[cmikida2]

@inducer I believe all the tests will pass here, but this is the state of the SBP-DG work in grudge based on the Kozdon and Wilcox paper linked in the PR description. Since I last worked on this in great detail, a number of things in Grudge have changed, and in particular I make no claims about this being the most performant way to implement an SBP-to-DG (and vice-versa) hybrid spatial discretization/interface, but at minimum it may serve as a useful starting point should someone else want to at some point add this capability to grudge and its other models at large.

[alexfikl]

@cmikida2 I can approve the CI if you feel like getting stressed by pylint? 😁

[cmikida2]

@alexfikl go right ahead, I'm sure I'm about to get nuked