ndarray 0.17.2

An n-dimensional array for general elements and for numerics. Lightweight array views and slicing; views support chunking and splitting.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147

//! Parallelization features for ndarray.
//!
//! Parallelization features are based on the crate [rayon] and its parallel
//! iterators. Ndarray implements the parallel iterable traits for arrays
//! and array views, for some of its iterators and for [Zip].
//! There are also directly parallelized methods on arrays and on [Zip].
//!
//! This requires the crate feature `rayon` to be enabled.
//!
//! The following types implement parallel iterators, accessed using these
//! methods:
//!
//! - [`Array`], [`ArcArray`] `.par_iter()` and `.par_iter_mut()`
//! - [`ArrayView`] `.into_par_iter()`
//! - [`ArrayViewMut`] `.into_par_iter()`
//! - [`AxisIter`], [`AxisIterMut`] `.into_par_iter()`
//! - [`AxisChunksIter`], [`AxisChunksIterMut`] `.into_par_iter()`
//! - [`Zip`] `.into_par_iter()`
//!
//! The following other parallelized methods exist:
//!
//! - [`ArrayRef::par_map_inplace()`](crate::ArrayRef::par_map_inplace)
//! - [`ArrayRef::par_mapv_inplace()`](crate::ArrayRef::par_mapv_inplace)
//! - [`Zip::par_for_each()`] (all arities)
//! - [`Zip::par_map_collect()`] (all arities)
//! - [`Zip::par_map_assign_into()`] (all arities)
//!
//! Note that you can use the parallel iterator for [Zip] to access all other
//! rayon parallel iterator methods.
//!
//! Only the axis iterators are indexed parallel iterators, the rest are all
//! “unindexed”. Use ndarray’s [Zip] for lock step parallel iteration of
//! multiple arrays or producers at a time.
//!
//! For the unindexed parallel iterators, an inherent method [`with_min_len`](Parallel::with_min_len)
//! is provided to limit the number of elements each parallel task processes in way that is
//! similar to Rayon's [`IndexedParallelIterator::with_min_len`](rayon::prelude::IndexedParallelIterator::with_min_len).
//!
//! # Examples
//!
//! ## Arrays and array views
//!
//! Compute the exponential of each element in an array, parallelized.
//!
//! ```
//! use ndarray::Array2;
//! use ndarray::parallel::prelude::*;
//!
//! let mut a = Array2::<f64>::zeros((128, 128));
//!
//! // Parallel versions of regular array methods
//! a.par_map_inplace(|x| *x = x.exp());
//! a.par_mapv_inplace(f64::exp);
//!
//! // You can also use the parallel iterator directly
//! a.par_iter_mut().for_each(|x| *x = x.exp());
//! ```
//!
//! ## Axis iterators
//!
//! Use the parallel `.axis_iter()` to compute the sum of each row.
//!
//! ```
//! use ndarray::Array;
//! use ndarray::Axis;
//! use ndarray::parallel::prelude::*;
//!
//! let a = Array::linspace(0., 63., 64).into_shape_with_order((4, 16)).unwrap();
//! let mut sums = Vec::new();
//! a.axis_iter(Axis(0))
//!  .into_par_iter()
//!  .map(|row| row.sum())
//!  .collect_into_vec(&mut sums);
//!
//! assert_eq!(sums, [120., 376., 632., 888.]);
//! ```
//!
//! ## Axis chunks iterators
//!
//! Use the parallel `.axis_chunks_iter()` to process your data in chunks.
//!
//! ```
//! use ndarray::Array;
//! use ndarray::Axis;
//! use ndarray::parallel::prelude::*;
//!
//! let a = Array::linspace(0., 63., 64).into_shape_with_order((4, 16)).unwrap();
//! let mut shapes = Vec::new();
//! a.axis_chunks_iter(Axis(0), 3)
//!  .into_par_iter()
//!  .map(|chunk| chunk.shape().to_owned())
//!  .collect_into_vec(&mut shapes);
//!
//! assert_eq!(shapes, [vec![3, 16], vec![1, 16]]);
//! ```
//!
//! ## Zip
//!
//! Use zip for lock step function application across several arrays
//!
//! ```
//! use ndarray::Array3;
//! use ndarray::Zip;
//!
//! type Array3f64 = Array3<f64>;
//!
//! const N: usize = 128;
//! let a = Array3f64::from_elem((N, N, N), 1.);
//! let b = Array3f64::from_elem(a.dim(), 2.);
//! let mut c = Array3f64::zeros(a.dim());
//!
//! Zip::from(&mut c)
//!     .and(&a)
//!     .and(&b)
//!     .par_for_each(|c, &a, &b| {
//!         *c += a - b;
//!     });
//! ```

#[allow(unused_imports)] // used by rustdoc links
use crate::iter::{AxisChunksIter, AxisChunksIterMut, AxisIter, AxisIterMut};
#[allow(unused_imports)] // used by rustdoc links
use crate::{ArcArray, Array, ArrayBase, ArrayView, ArrayViewMut, Zip};

/// Into- traits for creating parallelized iterators and/or using [`par_azip!`]
pub mod prelude
{
    #[doc(no_inline)]
    pub use rayon::prelude::{
        IndexedParallelIterator,
        IntoParallelIterator,
        IntoParallelRefIterator,
        IntoParallelRefMutIterator,
        ParallelIterator,
    };

    pub use super::par_azip;
}

pub use self::par::Parallel;
pub use crate::par_azip;

mod impl_par_methods;
mod into_impls;
mod par;
mod send_producer;
mod zipmacro;