2d PDE example

[jw3126]

I tried to do a 2d poisson equation example, but I run into multiple issues. Here is an MWE:

using DiffEqOperators

nx, ny = (3, 3)
hx, hy = 1.0 ./ (1 .+ (nx, ny))
yknots = range(hy, step=hy, length=ny)

qx = [DirichletBC(randn(2)...) for y in yknots]
Qx = MultiDimBC{1}(qx)
D = CenteredDifference{1}(2,2, hx, nx)

D*(Qx*randn(nx, ny)) # no error
D*Qx # no error
(D*Qx)*randn(nx, ny) # error

MethodError: no method matching SparseArrays.SparseMatrixCSC(::MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1})
Closest candidates are:
  SparseArrays.SparseMatrixCSC(::MultiDimDirectionalBC{T,B,D,N,K}, !Matched::Any) where {T, B, D, N, K} at /home/jan/.julia/dev/DiffEqOperators/src/derivative_operators/concretization.jl:188
  SparseArrays.SparseMatrixCSC(!Matched::LinearAlgebra.SymTridiagonal{Tv,V} where V<:AbstractArray{Tv,1}) where Tv at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.3/SparseArrays/src/sparsematrix.jl:436
  SparseArrays.SparseMatrixCSC(!Matched::LinearAlgebra.Tridiagonal{Tv,V} where V<:AbstractArray{Tv,1}) where Tv at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.3/SparseArrays/src/sparsematrix.jl:407
  ...

Stacktrace:
 [1] convert(::Type{AbstractArray{T,2} where T}, ::MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1}) at /home/jan/.julia/dev/DiffEqOperators/src/derivative_operators/concretization.jl:162
 [2] *(::MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1}, ::SubArray{Float64,1,Array{Float64,2},Tuple{Base.Slice{Base.OneTo{Int64}},Int64},true}) at /home/jan/.julia/packages/DiffEqBase/4V8I6/src/operators/common_defaults.jl:14
 [3] mul!(::SubArray{Float64,1,Array{Float64,2},Tuple{Base.Slice{Base.OneTo{Int64}},Int64},true}, ::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1}}, ::SubArray{Float64,1,Array{Float64,2},Tuple{Base.Slice{Base.OneTo{Int64}},Int64},true}) at /home/jan/.julia/dev/DiffEqOperators/src/derivative_operators/ghost_derivative_operator.jl:12
 [4] mul!(::Array{Float64,2}, ::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1}}, ::Array{Float64,2}) at /home/jan/.julia/dev/DiffEqOperators/src/derivative_operators/ghost_derivative_operator.jl:25
 [5] *(::GhostDerivativeOperator{Float64,DerivativeOperator{Float64,1,false,Float64,StaticArrays.SArray{Tuple{3},Float64,1,3},StaticArrays.SArray{Tuple{0},StaticArrays.SArray{Tuple{4},Float64,1,4},1,0},Nothing,Nothing},MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1}}, ::Array{Float64,2}) at /home/jan/.julia/dev/DiffEqOperators/src/derivative_operators/ghost_derivative_operator.jl:32
 [6] top-level scope at In[124]:14

[ChrisRackauckas]

MethodError: no method matching SparseArrays.SparseMatrixCSC(::MultiDimDirectionalBC{Float64,RobinBC{Float64,Array{Float64,1}},1,2,1})

This needs a length to be given, but in the use case here it should be able to be inferred from the problem vector sizes?