Fix/simple direct summand

[kalmarek]

This reverts the "fix" I applied some time ago ;/

I realized that e.g. for abelian groups there aren't enough projections in the group ring (all of them are central), so it's (mathematically) not possible to realize minimal rank projection as an element of the group ring ;/ life is hard and sometimes math kicks :D

so, when χ is a complex irreducible character, for real affordable ψ = χ + conj(χ) we return semisimple projection. It should be possible to find the rank one projection numerically though, by noticing that image_basis(..., ψ) consists a multiple of 2-dimensional mutually orthogonal subspaces and we need to project on exactly one of them (or their sum, or whatever). I tried to do so numerically, but failed ;)

@blegat

[codecov]

Codecov Report

Merging #40 (6e16157) into master (dbbb167) will decrease coverage by 1.84%.
The diff coverage is 66.17%.

@@            Coverage Diff             @@
##           master      #40      +/-   ##
==========================================
- Coverage   87.89%   86.04%   -1.85%     
==========================================
  Files          15       16       +1     
  Lines         958      982      +24     
==========================================
+ Hits          842      845       +3     
- Misses        116      137      +21     

Flag	Coverage Δ
unittests	86.04% <66.17%> (-1.85%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/matrix_projections.jl	91.91% <0.00%> (-3.88%)	⬇️
src/wedderburn_decomposition.jl	69.07% <61.76%> (-5.09%)	⬇️
src/actions.jl	83.33% <64.28%> (-5.96%)	⬇️
src/minimal_projections.jl	89.33% <85.71%> (-0.15%)	⬇️
src/sa_basis.jl	98.91% <100.00%> (+0.03%)	⬆️
src/util.jl	100.00% <100.00%> (ø)
src/direct_summands.jl	53.84% <0.00%> (-46.16%)	⬇️

---

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 dbbb167...6e16157. Read the comment docs.