Respect IOContext while displaying a SparseMatrixCSC

[jishnub]

Fix #399

On main:

julia> S = sprand(Int8, 20, 20, 0.3);

julia> [S]
1-element Vector{SparseMatrixCSC{Int8, Int64}}:
 sparse([2, 3, 4, 8, 9, 13, 14, 16, 18, 2  …  8, 13, 16, 17, 19, 4, 6, 10, 18, 19], [1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 10, 10, 10, 10, 11, 11, 11, 11, 11, 12, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 15, 15, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 17, 17, 17, 17, 17, 18, 18, 18, 18, 18, 19, 19, 19, 19, 19, 20, 20, 20, 20, 20], Int8[108, 6, -86, 17, 48, 51, 23, 88, 82, -24  …  -27, -3, 24, -56, 109, 73, 70, 101, -100, 46], 20, 20)

This PR

julia> [S]
1-element Vector{SparseMatrixCSC{Int8, Int64}}:
 sparse([2, 3, 4, 8, 9, 13, 14, 16, 18, 2  …  8, 13, 16, 17, 19, 4, 6, 10, 18, 19], [1, 1, 1, 1, 1, 1, 1, 1, 1, 2  …  19, 19, 19, 19, 19, 20, 20, 20, 20, 20], Int8[108, 6, -86, 17, 48, 51, 23, 88, 82, -24  …  -27, -3, 24, -56, 109, 73, 70, 101, -100, 46], 20, 20)

The main change is to use an AbstractVector for the column indices instead of printing them one by one to the display, which ensures that the IOContext is automatically respected in the same way as the row indices and the values.

Indexing into the AbstractVector is O(log n) as it uses searchsortedlast, but since this is used only in display, the performance should not matter much.

[codecov]

Codecov Report

Merging #423 (1b7e1b1) into main (3d1eda9) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #423      +/-   ##
==========================================
+ Coverage   92.41%   92.42%   +0.01%     
==========================================
  Files          12       12              
  Lines        7669     7666       -3     
==========================================
- Hits         7087     7085       -2     
+ Misses        582      581       -1     

Files Changed	Coverage Δ
src/sparsematrix.jl	95.68% <100.00%> (+0.03%)	⬆️

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

[rayegun]

Can we do this without introducing a new struct? I'm not sure I see the benefit of a new type for a single function.

[jishnub]

What approach would you recommend? One may duplicate the code for printing an AbstractVector, but that seems to only increase redundancy. What's the issue with introducing a struct, especially one that doesn't feature in the core functions?

[rayegun]

After thinking SGTM. You're right this is probably the least overall code.

If you wanted to do it as an iterator that would work as well.

If you could add a docstring (what it's for and state that it's not part of the public interface) I'll merge.

Fixed