[NDTensors] [BUG] DiagBlockSparse * DiagBlockSparse contraction broken

[mtfishman]

Description of bug

julia> s = [QN(0) => 1, QN(1) => 1];

julia> i, j, k = Index.((s, s, s));

julia> t1 = δ(i, dag(j));

julia> t2 = δ(j, dag(k));

julia> t1 * t2
ERROR: UndefVarError: IndsR not defined
Stacktrace:
  [1] contraction_output_type(TensorT1::Type{NDTensors.DiagBlockSparseTensor{Float64, 2, Tuple{Index{Vector{Pair{QN, Int64}}}, Index{Vector{Pair{QN, Int64}}}}, NDTensors.DiagBlockSparse{Float64, Float64, 2}}}, TensorT2::Type{NDTensors.DiagBlockSparseTensor{Float64, 2, Tuple{Index{Vector{Pair{QN, Int64}}}, Index{Vector{Pair{QN, Int64}}}}, NDTensors.DiagBlockSparse{Float64, Float64, 2}}}, indsR::Tuple{Index{Vector{Pair{QN, Int64}}}, Index{Vector{Pair{QN, Int64}}}})
    @ NDTensors ~/.julia/dev/ITensors/NDTensors/src/blocksparse/diagblocksparse.jl:278
[...]

This is with NDTensors v0.1.49 and is a regression caused by #1051, which means that this operation wasn't be tested properly (and should be tested in NDTensors). @kmp5VT I can look into this one.

[mtfishman]

This is a bit more involved then I expected, and ultimately would be made much easier by making BlockSparse and DiagBlockSparse more generic by sharing more code. I think we should be able to combine them into a single parametrized type, parametrized by the storage type of the blocks. So I'll leave this one open, for now people can use:

julia> @show denseblocks(t1) * denseblocks(t2);
denseblocks(t1) * denseblocks(t2) = ITensor ord=2
Dim 1: (dim=2|id=584) <Out>
 1: QN(0) => 1
 2: QN(1) => 1
Dim 2: (dim=2|id=9) <In>
 1: QN(0) => 1
 2: QN(1) => 1
NDTensors.BlockSparse{Float64, Vector{Float64}, 2}
 2×2
Block(1, 1)
 [1:1, 1:1]
 1.0

Block(2, 2)
 [2:2, 2:2]
 1.0