tensor_product macro

[KnutAM]

Make it easy to define new unrolled tensor expressions based on an index expression,
xref: #205 (comment)

Examples

Tensors.@tensor_product(@inline @inbounds function dcontract(A::Tensor{4,3}, B::Tensor{4,3})
    C[i,j,k,l] = A[i,j,m,n]*B[m,n,k,l]
end, muladd)

Tensors.@tensor_product(@inline @inbounds function otimes(A::Tensor{2,3}, B::Tensor{2,3})
    C[i,j,k,l] = A[i,j]*B[k,l]
end)

TODO

• Support SymmetricTensor output (need to automatically figure out when based on the input and expression)
• Replace current functions with the generated functions where that makes sense (e.g. otimes already use index expression, and that approach seems faster than manually unrolling)
• Document the macro such that users can create their special functions (not public API yet)

[codecov-commenter]

Codecov Report

Attention: 16 lines in your changes are missing coverage. Please review.

Comparison is base (ce37b01) 96.91% compared to head (62d0fc7) 93.69%.
Report is 3 commits behind head on master.

Files	Patch %	Lines
src/utilities.jl	92.48%	16 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #211      +/-   ##
==========================================
- Coverage   96.91%   93.69%   -3.23%     
==========================================
  Files          17       17              
  Lines        1490     1522      +32     
==========================================
- Hits         1444     1426      -18     
- Misses         46       96      +50     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[termi-official]

I think this can also really help users getting #188 functional for their corner cases

[KnutAM]

@fredrikekre or @KristofferC - would any of you have time to review?

Feels like there is too much code for "parsing" the user input, but didn't find a way to reduce it...