Function reference
Plain reducers
NumPy-like reducers where NaN and infinity participate as ordinary IEEE values.
| mean | Return the arithmetic mean with NaN/inf propagation. |
| average | Return the weighted average with NaN/inf propagation. |
| sum | Return the sum with NaN/inf propagation. |
| min | Return the minimum with NaN propagation. |
| max | Return the maximum with NaN propagation. |
| median | Return the median with NaN propagation. |
| var | Return the variance with NaN/inf propagation. |
| std | Return the standard deviation with NaN/inf propagation. |
| percentile | Return linear-interpolation percentiles with NaN propagation. |
| quantile | Return linear-interpolation quantiles with NaN propagation. |
NaN-aware reducers
NumPy nan* parity by default, with ignore_inf=True for finite-only reductions.
| nanmean | Return the arithmetic mean while skipping NaN values. |
| nanaverage | Return the weighted average while skipping NaN values. |
| nansum | Return the sum while skipping NaN values. |
| nanmin | Return the minimum while skipping NaN values. |
| nanmax | Return the maximum while skipping NaN values. |
| nanminmax | Return the minimum and maximum while skipping NaN values. |
| nanmedian | Return the median while skipping NaN values. |
| nanvar | Return the variance while skipping NaN values. |
| nanstd | Return the standard deviation while skipping NaN values. |
| nanpercentile | Return linear-interpolation percentiles while skipping NaN values. |
| nanquantile | Return linear-interpolation quantiles while skipping NaN values. |
Extra reducers
Reducers-specific helpers.
| lmedian | Return the lower value-selecting median while skipping NaN values. |
| minmax | Return the minimum and maximum with NaN propagation. |
| count_finite | Return the number of finite values. |
Sigma clipping
Finite-only clipping with reusable settings and fused mean/median outputs.
| sigclip | Sigma-clipping rejection. |
| sigclip_1d | Sigma-clipping rejection for a 1-D value vector. |
| sigclip_mask | Return only the sigma-clipping rejection mask. |
| sigclip_mask_1d | Return only the sigma-clipping rejection mask for a 1-D value vector. |
| sigclip_combine | Return an output-only sigma-clipped mean or median. |
| sigclip_combine_1d | Return a 1-D sigma-clipped mean or median. |
| sigclip_restored_flags | Return per-sample restored-candidate flags for sigma clipping. |
| grow_mask | Dilate a stack mask over spatial axes by an exact Euclidean radius. |
| SigClip | Reusable, immutable settings for axis-0 sigma clipping. |