Add a scipy.sparse numba extension

[flying-sheep]

OK, I now made the benchmarks bigger. This way only some is_constant tests for sparse matrices are still <1ms (and therefore show up in the regressions). (This is because when reducing along the major axis, many rows/cols without entries can be skipped)

[codspeed-hq]

CodSpeed Performance Report

Merging #73 will degrade performances by 37.09%

_{Comparing pa/nbext (59b9fc3) with main (104bf1c)}

Summary

❌ 4 (👁 4) regressions
✅ 84 untouched benchmarks
🆕 36 new benchmarks

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
🆕	test_stats_benchmark[0-numpy.ndarray-int32-is_constant]	N/A	3.7 ms	N/A
🆕	test_stats_benchmark[0-numpy.ndarray-int32-mean]	N/A	4.7 ms	N/A
🆕	test_stats_benchmark[0-numpy.ndarray-int32-mean_var]	N/A	9.5 ms	N/A
🆕	test_stats_benchmark[0-numpy.ndarray-int32-sum]	N/A	4.7 ms	N/A
👁	test_stats_benchmark[0-scipy.sparse.csc_array-float32-is_constant]	482.4 µs	747.2 µs	-35.43%
👁	test_stats_benchmark[0-scipy.sparse.csc_array-float64-is_constant]	479.6 µs	743.2 µs	-35.47%
🆕	test_stats_benchmark[0-scipy.sparse.csc_array-int32-is_constant]	N/A	746 µs	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csc_array-int32-mean]	N/A	11.8 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csc_array-int32-mean_var]	N/A	6.7 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csc_array-int32-sum]	N/A	11.7 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csr_array-int32-is_constant]	N/A	22.9 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csr_array-int32-mean]	N/A	16.7 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csr_array-int32-mean_var]	N/A	19.6 ms	N/A
🆕	test_stats_benchmark[0-scipy.sparse.csr_array-int32-sum]	N/A	16.6 ms	N/A
🆕	test_stats_benchmark[1-numpy.ndarray-int32-is_constant]	N/A	3.8 ms	N/A
🆕	test_stats_benchmark[1-numpy.ndarray-int32-mean]	N/A	4.2 ms	N/A
🆕	test_stats_benchmark[1-numpy.ndarray-int32-mean_var]	N/A	9.2 ms	N/A
🆕	test_stats_benchmark[1-numpy.ndarray-int32-sum]	N/A	4.1 ms	N/A
🆕	test_stats_benchmark[1-scipy.sparse.csc_array-int32-is_constant]	N/A	22.9 ms	N/A
🆕	test_stats_benchmark[1-scipy.sparse.csc_array-int32-mean]	N/A	16.6 ms	N/A
...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.91%. Comparing base (104bf1c) to head (59b9fc3).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #73      +/-   ##
==========================================
+ Coverage   98.49%   98.91%   +0.42%     
==========================================
  Files          15       17       +2     
  Lines         265      370     +105     
==========================================
+ Hits          261      366     +105     
  Misses          4        4              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[ilan-gold]

I do wonder if some of these pyi files could be contributed back to the packages for which they are made

[flying-sheep]

I’m certainly planning to do that!