38 some tensor conversions

[srosenbu]

Adds

• function to convert $u$ to $\varepsilon$ in Mandel form for ufl expressions
• function to convert an array of $\nabla u$ to $\varepsilon$ in Mandel form

@pdiercks , I am not sure about the dimensions of some of the strain vectors. For plane strain, the 4-element vector is really useful, but I am not sure if this is the same for plane stress

[Edit] Also adds basic testing to the github workflow.

[srosenbu]

I just remembered that the nabla operator is ambiguous for vector valued expressions: see https://jsdokken.com/dolfinx-tutorial/chapter2/linearelasticity_code.html

For the strain it does not matter, but for the spin tensor $W=0.5\cdot(\nabla u - \nabla u^\top)$ it matters if we used grad or nabla_grad. Which should we use?

[Edit] The book Kontinuums-mechanik by Altenbach uses the nabla_grad convention. I think, I will change it in the code, even though it does not matter in this specific case.

[pdiercks]

@pdiercks , I am not sure about the dimensions of some of the strain vectors. For plane strain, the 4-element vector is really useful, but I am not sure if this is the same for plane stress

we discussed this already, but for the purpose of documentation: for plane stress I would use the same strain input as for plane stress. The user is responsible to compute the stress such that $\sigma_{zz}=0$, e.g. setting $\varepsilonzz$ as a function of $\varepsilon{xx}$ and $\varepsilon_{yy}$ in the case of linear elasticity.

I agree with the nabla_grad convention.