Fix symmetry enforcement with sparse matrices in Laplacian filter

[aaditgupta21]

#49

[codecov-commenter]

Codecov Report

Merging #52 (510d455) into main (6071118) will decrease coverage by 0.09%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##             main      #52      +/-   ##
==========================================
- Coverage   18.34%   18.25%   -0.10%     
==========================================
  Files           7        7              
  Lines         398      400       +2     
==========================================
  Hits           73       73              
- Misses        325      327       +2     

Impacted Files	Coverage Δ
torchts/utils/graph.py	0.00% <0.00%> (ø)

---

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 6071118...510d455. Read the comment docs.

[klane]

The line as it was originally written was probably fine. We likely just need an if statement inside the undirected if statement that does the following:

• Check that adj_mx is sparse prior to np.maximum
• Convert adj_mx to dense only if it is sparse

[klane]

The line as it was originally written was probably fine. We likely just need an if statement inside the undirected if statement

The function calls normalized_laplacian, which creates a sparse matrix anyway. We can convert sparse matrices to dense at the top of the function regardless of if it is undirected.