Fixes for ComponentArrays 0.15.23

[visr]

In #2027 I had to hold back the latest ComponentArrays version that has this PR: SciML/ComponentArrays.jl#249.

It gave errors like:

ERROR: LoadError: MethodError: Cannot `convert` an object of type ComponentArrays.ComponentIndex{Int64, ComponentArrays.NullAxis} to an object of type Int64

Closest candidates are:
  convert(::Type{T}, !Matched::VectorizationBase.AbstractSIMD) where T<:Union{Bool, Float16, Float32, Float64, Int16, Int32, Int64, Int8, UInt16, UInt32, UInt64, UInt8, SIMDTypes.Bit}
   @ VectorizationBase ~/.julia/packages/VectorizationBase/wHnQd/src/base_defs.jl:201
  convert(::Type{T}, !Matched::SentinelArrays.ChainedVectorIndex) where T<:Union{Signed, Unsigned}
   @ SentinelArrays ~/.julia/packages/SentinelArrays/vt8A6/src/chainedvector.jl:262
  convert(::Type{<:Real}, !Matched::T) where T<:SparseConnectivityTracer.AbstractTracer
   @ SparseConnectivityTracer ~/.julia/packages/SparseConnectivityTracer/zXLIF/src/overloads/conversion.jl:10
  ...

Stacktrace:
  [1] get_state_index(u::ComponentArrays.ComponentVector{Float64, Vector{Float64}, Tuple{ComponentArrays.Axis{(tabulated_rating_curve = ViewAxis(1:0, Shaped1DAxis((0,))), pump = ViewAxis(1:1, Shaped1DAxis((1,))), outlet = ViewAxis(2:4, Shaped1DAxis((3,))), user_demand_inflow = ViewAxis(5:7, Shaped1DAxis((3,))), user_demand_outflow = ViewAxis(8:10, Shaped1DAxis((3,))), linear_resistance = ViewAxis(11:10, Shaped1DAxis((0,))), manning_resistance = ViewAxis(11:10, Shaped1DAxis((0,))), evaporation = ViewAxis(11:13, Shaped1DAxis((3,))), infiltration = ViewAxis(14:16, Shaped1DAxis((3,))), integral = ViewAxis(17:16, Shaped1DAxis((0,))))}}}, link::Tuple{Ribasim.NodeID, Ribasim.NodeID})
    @ Ribasim ~/work/Ribasim/Ribasim/core/src/util.jl:1052

We should look into fixing Ribasim to work with this, assuming this is intended behavior.

[visr]

In discussing with @SouthEndMusic another option may be to try using our own struct of arrays for u and use https://github.com/JuliaArrays/LazyArrays.jl for the operations where u needs to be considered a single vector. Or perhaps something else based on Memory would work, just like https://github.com/JuliaArrays/FixedSizeArrays.jl is based on that.