Remve dim parameter from AbstractDim

[lijas]

It is a bit cleaner like this I think :)

[codecov-commenter]

Codecov Report

Merging #41 (6c7278d) into main (57ceea5) will decrease coverage by 0.33%.
The diff coverage is 63.63%.

@@            Coverage Diff             @@
##             main      #41      +/-   ##
==========================================
- Coverage   92.07%   91.74%   -0.34%     
==========================================
  Files          11       12       +1     
  Lines         404      436      +32     
==========================================
+ Hits          372      400      +28     
- Misses         32       36       +4     

Impacted Files	Coverage Δ
src/wrappers.jl	78.78% <63.63%> (-2.88%)	⬇️
src/CohesiveMaterials/XuNeedleman.jl	95.23% <0.00%> (ø)
src/CrystalViscoPlastic/CrystalViscoPlasticRed.jl	98.63% <0.00%> (+0.01%)	⬆️
src/CrystalViscoPlastic/CrystalViscoPlastic.jl	98.55% <0.00%> (+0.02%)	⬆️
src/Plastic.jl	98.33% <0.00%> (+0.02%)	⬆️
src/FiniteStrain/yeoh.jl	95.65% <0.00%> (+0.19%)	⬆️
src/FiniteStrain/stvenant.jl	95.23% <0.00%> (+0.23%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 57ceea5...6c7278d. Read the comment docs.

[lijas]

Seems to be type stable:

mat = LinearElastic(E = 1.0, ν = 0.3)
ε = rand(SymmetricTensor{2,2,Float64})
state = initial_material_state(mat)

@code_warntype material_response(PlaneStrain(), mat, ε, state)

Static Parameters
  d = 2
  T = Float64
Arguments
  #self#::Core.Const(MaterialModels.material_response)
  dimstate::Core.Const(PlaneStress())
  m::LinearElastic
  Δε::SymmetricTensor{2, 2, Float64, 3}
  state::Core.Const(LinearElasticState())
Body::Tuple{SymmetricTensor{2, 2, Float64, 3}, SymmetricTensor{4, 2, Float64, 9}, LinearElasticState}
1 ─ %1 = (#self#)(dimstate, m, Δε, state, MaterialModels.nothing)::Tuple{SymmetricTensor{2, 2, Float64, 3}, SymmetricTensor{4, 2, Float64, 9}, LinearElasticState}
└──      return %1

[kimauth]

Though this would mean that the dimension is not available at compile time anymore I think.

Perhaps fixable by defining

getdim(::Type{UniaxialStrain}) = 1

and so on?

[lijas]

Hmm, if I remember correctly it everything was type-stable. Maybe because it is a singleton? I can double check

[kimauth]

What would I use for generic 2D cases that are neither plane strain nor plane stress now (like 2D cohesive elements)?