Implement KaHyPar optimizer

[jofrevalles]

This PR addresses issue #7 by implementing the Hypergraph partitioning optimizer, inspired by the work in cotengra and its paper.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (20b36e8) 77.41% compared to head (3ed894e) 78.70%.
Report is 16 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #35      +/-   ##
==========================================
+ Coverage   77.41%   78.70%   +1.28%     
==========================================
  Files          10       11       +1     
  Lines         217      277      +60     
==========================================
+ Hits          168      218      +50     
- Misses         49       59      +10     

Files	Coverage Δ
src/EinExprs.jl	100.00% <ø> (ø)
src/Optimizers/KaHyPar.jl	100.00% <100.00%> (ø)
src/Optimizers/Optimizers.jl	50.00% <ø> (ø)

... and 2 files with indirect coverage changes

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

[mofeing]

Implementation finished!

To do

• Write tests
• Optimize incidence matrix construction by using subviews instead of recreating sub incidence matrices
• Check KaHyPar configuration path
• Check what node_weights,edge_weights should be

[mofeing]

Runtime logs by the KaHyPar library can only be disabled via the quiet option in the configuration files. kahypar/KaHyPar.jl#18

[mofeing]

@jofrevalles I'm merging because this branch is based on an old commit and some bugs are resurfacing when I try to use it. It's already usable but we need to understand some things before doing a release. Specifically managing KaHyPar profiles and checking that edge_weigths and node_weights are correctly defined.