ferray-ma

Masked arrays with mask propagation for the ferray scientific computing library.

What's in this crate

• MaskedArray<T, D> — array with boolean mask for missing/invalid data
• Mask propagation: arithmetic and ufuncs automatically propagate masks
• Full reductions: sum, prod, mean, std, var, cumsum, cumprod, min, max, argmin, argmax, median, ptp, average, anom
• Bitwise / comparison ufuncs: and/or/xor/shift, equal/less/greater family, logical_*, floor_divide/fmod/power/remainder/true_divide/hypot/negative/sqrt/square
• Array creation: arange, zeros, ones, empty, identity, masked_all, masked_array, masked_values, masked_object, masked_outside, masked_less/_equal, masked_not_equal, frombuffer, fromfunction, fromflex
• Manipulation: concatenate/stack family, reshape, ravel, squeeze, swapaxes, transpose, expand_dims, atleast_*, take, put, putmask, repeat, resize, append, diag, diagflat, diagonal, choose, clip
• Mask manipulation: harden_mask/soften_mask, mask_or, make_mask*, compress_rows/cols/rowcols/nd, mask_rows/cols/rowcols, clump_*masked, *notmasked_contiguous/edges, flatten_mask, flatten_structured_array
• Linalg / set / corr: dot, inner, outer, innerproduct, outerproduct, trace, correlate, convolve, corrcoef, cov, unique, in1d, isin, intersect1d, setdiff1d, setxor1d, union1d, polyfit
• apply_along_axis / apply_over_axes / vander for axis-aware mapping
• Fill-value protocol: default_fill_value, maximum_fill_value, minimum_fill_value, set_fill_value, common_fill_value
• Class helpers: mvoid, nomask, masked_singleton, masked_print_option, mr_, isMaskedArray/isMA/isarray, getmaskarray, ids

Implements numpy.ma for Rust.

Usage

use ferray_ma::MaskedArray;
use ferray_core::prelude::*;

let data = Array1::<f64>::from_vec(Ix1::new([4]), vec![1.0, 2.0, 3.0, 4.0])?;
let mask = vec![false, false, true, false]; // mask out index 2
let ma = MaskedArray::new(data, mask)?;

This crate is re-exported through the main ferray crate with the ma feature (enabled by default).

License

MIT OR Apache-2.0