Add SummationByPartsOperators.jl package extension (#152)

* adding initial extension for SummationByPartsOperators

* adding missing package imports

* add SBP extension tests

* rename multidimensional SBP tests

update name of multidim SBP tests

d

* update docs

* bump Julia compat to 1.9

* revert back to Julia compat 1.7, add Requires

* add forgotten part

* add [extras] to Project.toml

* try to fix tests to work with Requires for pre 1.9 julia

* fix Requires by moving ext into `init` function

* add dropped import

* formatting

* add mroe tests

Project.toml

@@ -23,6 +23,12 @@ StaticArrays = "90137ffa-7385-5640-81b9-e52037218182"
 Triangulate = "f7e6ffb2-c36d-4f8f-a77e-16e897189344"
 WriteVTK = "64499a7a-5c06-52f2-abe2-ccb03c286192"
 
+[weakdeps]
+SummationByPartsOperators = "9f78cca6-572e-554e-b819-917d2f1cf240"
+
+[extensions]
+StartUpDGSummationByPartsOperatorsExt = "SummationByPartsOperators"
+
 [compat]
 ConstructionBase = "1"
 FillArrays = "0.13, 1"
@@ -37,6 +43,10 @@ Requires = "1"
 Setfield = "1"
 SimpleUnPack = "1"
 StaticArrays = "1"
+SummationByPartsOperators = "0.5"
 Triangulate = "2"
 WriteVTK = "1"
 julia = "1.7"
+
+[extras]
+SummationByPartsOperators = "9f78cca6-572e-554e-b819-917d2f1cf240"

docs/src/RefElemData.md

@@ -72,9 +72,22 @@ By default, `RefElemData` is constructed for a nodal basis (in order to facilita
 
 ## `RefElemData` based on SBP finite differences
 
-It is also possible to construct a [`RefElemData`](@ref) based on [multi-dimensional SBP finite difference operators](https://doi.org/10.1137/15M1038360). These utilize nodes constructed by [Tianheng Chen and Chi-Wang Shu](https://doi.org/10.1016/j.jcp.2017.05.025), [Ethan Kubatko](https://sites.google.com/site/chilatosu/ethan-bio), and [Jason Hicken](https://doi.org/10.1007/s10915-020-01154-8).
+It is also possible to construct a [`RefElemData`](@ref) based on both traditional finite difference SBP operators and [multi-dimensional SBP finite difference operators](https://doi.org/10.1137/15M1038360). The traditional finite difference SBP operators are built using [SummationByPartsOperators.jl](https://github.com/ranocha/SummationByPartsOperators.jl). The multi-dimensional SBP operators utilize nodes constructed by [Tianheng Chen and Chi-Wang Shu](https://doi.org/10.1016/j.jcp.2017.05.025), [Ethan Kubatko](https://sites.google.com/site/chilatosu/ethan-bio), and [Jason Hicken](https://doi.org/10.1007/s10915-020-01154-8). 
 
-Some examples:
+Some examples of traditional finite difference SBP operators on tensor product domains:
+```julia
+using StartUpDG
+using SummationByPartsOperators # this package must be loaded to enable extension
+
+D = derivative_operator(MattssonNordström2004(), 
+                        derivative_order=1, accuracy_order=N+1,
+                        xmin=-1.0, xmax=1.0, N=20)
+
+rd = RefElemData(Line(), D) 
+rd = RefElemData(Quad(), D)
+rd = RefElemData(Hex(), D)
+```
+Some examples of multi-dimensional SBP operators:
 ```julia
 N = 3
 rd = RefElemData(Quad(), SBP(), N) # defaults to SBP{TensorProductLobatto}