Use a naive sparse histogram.

[delucchi-cmu]

We really just need the non-zero indexes and non-zero values, and the complexity of the general scipy sparse array is adding a lot of compute overhead. This simplifies the API.

This is a breaking change for LSDB and hats-import, and PRs are coming with fixes for those repos.

[github-actions]

Before [e07916b]	After [41c274d]	Ratio	Benchmark (Parameter)
75.3±0.3ms	76.1±0.8ms	1.01	benchmarks.MetadataSuite.time_load_partition_join_info
361±3ms	364±2ms	1.01	benchmarks.Suite.time_outer_pixel_alignment
19.1±0.2ms	19.2±0.3ms	1	benchmarks.MetadataSuite.time_load_partition_info_order6
13.4±0.4ms	13.4±0.3ms	1	benchmarks.Suite.time_inner_pixel_alignment
113±0.3ms	113±0.4ms	1	benchmarks.time_test_alignment_even_sky
1.02±0.02ms	1.02±0.01ms	1	benchmarks.time_test_cone_filter_multiple_order
76.1±0.3ms	75.6±0.4ms	0.99	benchmarks.MetadataSuite.time_load_partition_info_order7
41.4±0.6ms	41.0±0.7ms	0.99	benchmarks.Suite.time_pixel_tree_creation

Click here to view all benchmarks.

[codecov]

Codecov Report

Attention: Patch coverage is 93.10345% with 2 lines in your changes missing coverage. Please review.

Project coverage is 93.14%. Comparing base (e07916b) to head (6876408).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/hats/pixel_math/sparse_histogram.py	93.10%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #446      +/-   ##
==========================================
- Coverage   93.24%   93.14%   -0.10%     
==========================================
  Files          47       47              
  Lines        2012     2014       +2     
==========================================
  Hits         1876     1876              
- Misses        136      138       +2     

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

[camposandro]

Looks great, thanks! Added just a minor comment.