allow functionoperator (batch = false) to accept/return vec'd arrays

[vpuri3]

@ChrisRackauckas

with this branch,

julia> f(u, p, t) = 2u; f(v, u, p, t) = mul!(v, 2, u);                                                                                                               [30/1927]

julia> F = FunctionOperator(f, ones(4, 4));

julia> F * ones(4, 4)
4×4 Matrix{Float64}:
 2.0  2.0  2.0  2.0
 2.0  2.0  2.0  2.0
 2.0  2.0  2.0  2.0
 2.0  2.0  2.0  2.0                        

julia> F * vec(ones(4, 4))
16-element Vector{Float64}:
 2.0
 2.0
 2.0
 2.0
 2.0
...
julia> mul!(zeros(16), F, ones(16))
16-element Vector{Float64}:
 2.0
 2.0
 2.0
 2.0
 2.0
...

[codecov[bot]]

Codecov Report

Merging #210 (734205b) into master (f169b36) will increase coverage by 0.80%. The diff coverage is 73.41%.

@@            Coverage Diff             @@
##           master     #210      +/-   ##
==========================================
+ Coverage   70.31%   71.12%   +0.80%     
==========================================
  Files          10       10              
  Lines        1533     1562      +29     
==========================================
+ Hits         1078     1111      +33     
+ Misses        455      451       -4     

Impacted Files	Coverage Δ
src/func.jl	84.55% <73.41%> (+3.48%)	:arrow_up:

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more

[ChrisRackauckas]

Should we just remove the vec in that spot of OrdinaryDiffEq if it's hitting an operator?

[vpuri3]

We will remove the entire WOperator object evenetually.

The idea behind this PR is that we should support vec'd arrays anyway because one or the other downstream packages would use it. I've added warnings, docs etc all over the place as well. So after this PR, I believe we will be hitting all use cases and this package will be as versatile as it gets.

[vpuri3]

We should also move preconditioner tests to interfaceII in ODE.jl

[vpuri3]

BTW the failing test in preconditioner.jl is passing now