fix: prune nan idxs in output when omitting nans #83

Performance Regression: -55.89%

⚡ 13 improvements
❌ 48 regressions
✅ 617 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

	Benchmark	`BASE`	`HEAD`	Change
❌	`test_lttb_no_x[False-float32-5,000-100,000]`	1.2 ms	1.3 ms	-12.9%
❌	`test_lttb_no_x[False-float64-5,000-100,000]`	1.3 ms	1.5 ms	-12.59%
❌	`test_lttb_no_x[False-int32-5,000-100,000]`	1.3 ms	1.5 ms	-11.97%
❌	`test_lttb_no_x[False-int64-5,000-100,000]`	1.5 ms	1.7 ms	-10.86%
❌	`test_lttb_no_x[True-float32-5,000-100,000]`	1.2 ms	1.4 ms	-12.57%
❌	`test_lttb_no_x[True-float64-5,000-100,000]`	1.3 ms	1.5 ms	-12.33%
❌	`test_lttb_no_x[True-int32-5,000-100,000]`	1.3 ms	1.5 ms	-11.9%
❌	`test_lttb_no_x[True-int64-5,000-100,000]`	1.5 ms	1.7 ms	-10.78%
❌	`test_m4_no_x[False-float32-1,000-100,000]`	404.6 µs	466.6 µs	-13.3%
❌	`test_m4_no_x[False-float32-5,000-100,000]`	530.4 µs	705.3 µs	-24.8%
❌	`test_m4_no_x[False-float64-5,000-100,000]`	760 µs	941.8 µs	-19.3%
❌	`test_m4_no_x[False-int32-1,000-100,000]`	531.3 µs	590.8 µs	-10.07%
❌	`test_m4_no_x[False-int32-5,000-100,000]`	652.1 µs	820 µs	-20.47%
❌	`test_m4_no_x[False-int64-5,000-100,000]`	867.3 µs	1,047.1 µs	-17.17%
❌	`test_m4_no_x[True-float32-5,000-100,000]`	591.7 µs	737.5 µs	-19.76%
❌	`test_m4_no_x[True-float64-5,000-100,000]`	789 µs	986.4 µs	-20.02%
❌	`test_m4_no_x[True-int32-1,000-100,000]`	575.3 µs	651.6 µs	-11.72%
❌	`test_m4_no_x[True-int32-5,000-100,000]`	702.9 µs	867 µs	-18.93%
❌	`test_m4_with_x[False-float32-5,000-100,000]`	903.5 µs	1,075 µs	-15.95%
❌	`test_m4_with_x[False-float64-5,000-100,000]`	1.1 ms	1.3 ms	-14.12%
...	...	...	...	...

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