Fix: regression in identityoperator()+sparse_matrix due to the move to Eye #109

[SatyaBade12]

No description provided.

[codecov]

Codecov Report

Merging #110 (b7e25fe) into master (94ac58b) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #110      +/-   ##
==========================================
+ Coverage   93.34%   93.37%   +0.02%     
==========================================
  Files          24       24              
  Lines        2884     2897      +13     
==========================================
+ Hits         2692     2705      +13     
  Misses        192      192              

Impacted Files	Coverage Δ
src/operators_sparse.jl	96.42% <100.00%> (+1.07%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[Krastanov]

Thanks, @SatyaBade12 , the contribution is really appreciated! I added a few more methods through the webeditor, lets see if I broke anything! This should be mergeable shortly and I will then make an issue about possible further optimizations.

[amilsted]

Thanks @SatyaBade12 and @Krastanov. Should have tested these more thoroughly! I'm surprised sums and products of Eyes with numbers give us dense matrices. Arguably this should be done better in FillArrays, so perhaps we should file an upstream issue too.

Edit: JuliaArrays/FillArrays.jl#265

[amilsted]

It seems I probably missed this because SquareEye behaves differently to Eye under these operations. The former behaves correctly in sums and krons with sparse matrices, the latter does not.

identityoperator(some_basis) ought to be producing a SquareEye, but apparently it isn't. In any case, the general Eye case definitely needed fixing!

Edit: #111

[Krastanov]

@amilsted , I added tests for non-square operators. If this is green on CI I will merge it and we can then follow up with #111 (which would now have some broken tests)