/rust/registry/src/index.crates.io-1949cf8c6b5b557f/ndarray-0.16.1/src/stacking.rs
Line | Count | Source |
1 | | // Copyright 2014-2020 bluss and ndarray developers. |
2 | | // |
3 | | // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
4 | | // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
5 | | // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
6 | | // option. This file may not be copied, modified, or distributed |
7 | | // except according to those terms. |
8 | | |
9 | | #[cfg(not(feature = "std"))] |
10 | | use alloc::vec::Vec; |
11 | | |
12 | | use crate::dimension; |
13 | | use crate::error::{from_kind, ErrorKind, ShapeError}; |
14 | | use crate::imp_prelude::*; |
15 | | |
16 | | /// Concatenate arrays along the given axis. |
17 | | /// |
18 | | /// ***Errors*** if the arrays have mismatching shapes, apart from along `axis`. |
19 | | /// (may be made more flexible in the future).<br> |
20 | | /// ***Errors*** if `arrays` is empty, if `axis` is out of bounds, |
21 | | /// if the result is larger than is possible to represent. |
22 | | /// |
23 | | /// ``` |
24 | | /// use ndarray::{arr2, Axis, concatenate}; |
25 | | /// |
26 | | /// let a = arr2(&[[2., 2.], |
27 | | /// [3., 3.]]); |
28 | | /// assert!( |
29 | | /// concatenate(Axis(0), &[a.view(), a.view()]) |
30 | | /// == Ok(arr2(&[[2., 2.], |
31 | | /// [3., 3.], |
32 | | /// [2., 2.], |
33 | | /// [3., 3.]])) |
34 | | /// ); |
35 | | /// ``` |
36 | 0 | pub fn concatenate<A, D>(axis: Axis, arrays: &[ArrayView<A, D>]) -> Result<Array<A, D>, ShapeError> |
37 | 0 | where |
38 | 0 | A: Clone, |
39 | 0 | D: RemoveAxis, |
40 | | { |
41 | 0 | if arrays.is_empty() { |
42 | 0 | return Err(from_kind(ErrorKind::Unsupported)); |
43 | 0 | } |
44 | 0 | let mut res_dim = arrays[0].raw_dim(); |
45 | 0 | if axis.index() >= res_dim.ndim() { |
46 | 0 | return Err(from_kind(ErrorKind::OutOfBounds)); |
47 | 0 | } |
48 | 0 | let common_dim = res_dim.remove_axis(axis); |
49 | 0 | if arrays |
50 | 0 | .iter() |
51 | 0 | .any(|a| a.raw_dim().remove_axis(axis) != common_dim) |
52 | | { |
53 | 0 | return Err(from_kind(ErrorKind::IncompatibleShape)); |
54 | 0 | } |
55 | | |
56 | 0 | let stacked_dim = arrays.iter().fold(0, |acc, a| acc + a.len_of(axis)); |
57 | 0 | res_dim.set_axis(axis, stacked_dim); |
58 | 0 | let new_len = dimension::size_of_shape_checked(&res_dim)?; |
59 | | |
60 | | // start with empty array with precomputed capacity |
61 | | // append's handling of empty arrays makes sure `axis` is ok for appending |
62 | 0 | res_dim.set_axis(axis, 0); |
63 | 0 | let mut res = unsafe { |
64 | | // Safety: dimension is size 0 and vec is empty |
65 | 0 | Array::from_shape_vec_unchecked(res_dim, Vec::with_capacity(new_len)) |
66 | | }; |
67 | | |
68 | 0 | for array in arrays { |
69 | 0 | res.append(axis, array.clone())?; |
70 | | } |
71 | 0 | debug_assert_eq!(res.len_of(axis), stacked_dim); |
72 | 0 | Ok(res) |
73 | 0 | } |
74 | | |
75 | | /// Stack arrays along the new axis. |
76 | | /// |
77 | | /// ***Errors*** if the arrays have mismatching shapes. |
78 | | /// ***Errors*** if `arrays` is empty, if `axis` is out of bounds, |
79 | | /// if the result is larger than is possible to represent. |
80 | | /// |
81 | | /// ``` |
82 | | /// extern crate ndarray; |
83 | | /// |
84 | | /// use ndarray::{arr2, arr3, stack, Axis}; |
85 | | /// |
86 | | /// # fn main() { |
87 | | /// |
88 | | /// let a = arr2(&[[2., 2.], |
89 | | /// [3., 3.]]); |
90 | | /// assert!( |
91 | | /// stack(Axis(0), &[a.view(), a.view()]) |
92 | | /// == Ok(arr3(&[[[2., 2.], |
93 | | /// [3., 3.]], |
94 | | /// [[2., 2.], |
95 | | /// [3., 3.]]])) |
96 | | /// ); |
97 | | /// # } |
98 | | /// ``` |
99 | 0 | pub fn stack<A, D>(axis: Axis, arrays: &[ArrayView<A, D>]) -> Result<Array<A, D::Larger>, ShapeError> |
100 | 0 | where |
101 | 0 | A: Clone, |
102 | 0 | D: Dimension, |
103 | 0 | D::Larger: RemoveAxis, |
104 | | { |
105 | 0 | if arrays.is_empty() { |
106 | 0 | return Err(from_kind(ErrorKind::Unsupported)); |
107 | 0 | } |
108 | 0 | let common_dim = arrays[0].raw_dim(); |
109 | | // Avoid panic on `insert_axis` call, return an Err instead of it. |
110 | 0 | if axis.index() > common_dim.ndim() { |
111 | 0 | return Err(from_kind(ErrorKind::OutOfBounds)); |
112 | 0 | } |
113 | 0 | let mut res_dim = common_dim.insert_axis(axis); |
114 | | |
115 | 0 | if arrays.iter().any(|a| a.raw_dim() != common_dim) { |
116 | 0 | return Err(from_kind(ErrorKind::IncompatibleShape)); |
117 | 0 | } |
118 | | |
119 | 0 | res_dim.set_axis(axis, arrays.len()); |
120 | | |
121 | 0 | let new_len = dimension::size_of_shape_checked(&res_dim)?; |
122 | | |
123 | | // start with empty array with precomputed capacity |
124 | | // append's handling of empty arrays makes sure `axis` is ok for appending |
125 | 0 | res_dim.set_axis(axis, 0); |
126 | 0 | let mut res = unsafe { |
127 | | // Safety: dimension is size 0 and vec is empty |
128 | 0 | Array::from_shape_vec_unchecked(res_dim, Vec::with_capacity(new_len)) |
129 | | }; |
130 | | |
131 | 0 | for array in arrays { |
132 | 0 | res.append(axis, array.clone().insert_axis(axis))?; |
133 | | } |
134 | | |
135 | 0 | debug_assert_eq!(res.len_of(axis), arrays.len()); |
136 | 0 | Ok(res) |
137 | 0 | } |
138 | | |
139 | | /// Stack arrays along the new axis. |
140 | | /// |
141 | | /// Uses the [`stack()`] function, calling `ArrayView::from(&a)` on each |
142 | | /// argument `a`. |
143 | | /// |
144 | | /// ***Panics*** if the `stack` function would return an error. |
145 | | /// |
146 | | /// ``` |
147 | | /// extern crate ndarray; |
148 | | /// |
149 | | /// use ndarray::{arr2, arr3, stack, Axis}; |
150 | | /// |
151 | | /// # fn main() { |
152 | | /// |
153 | | /// let a = arr2(&[[1., 2.], |
154 | | /// [3., 4.]]); |
155 | | /// assert_eq!( |
156 | | /// stack![Axis(0), a, a], |
157 | | /// arr3(&[[[1., 2.], |
158 | | /// [3., 4.]], |
159 | | /// [[1., 2.], |
160 | | /// [3., 4.]]]), |
161 | | /// ); |
162 | | /// assert_eq!( |
163 | | /// stack![Axis(1), a, a,], |
164 | | /// arr3(&[[[1., 2.], |
165 | | /// [1., 2.]], |
166 | | /// [[3., 4.], |
167 | | /// [3., 4.]]]), |
168 | | /// ); |
169 | | /// assert_eq!( |
170 | | /// stack![Axis(2), a, a], |
171 | | /// arr3(&[[[1., 1.], |
172 | | /// [2., 2.]], |
173 | | /// [[3., 3.], |
174 | | /// [4., 4.]]]), |
175 | | /// ); |
176 | | /// # } |
177 | | /// ``` |
178 | | #[macro_export] |
179 | | macro_rules! stack { |
180 | | ($axis:expr, $( $array:expr ),+ ,) => { |
181 | | $crate::stack!($axis, $($array),+) |
182 | | }; |
183 | | ($axis:expr, $( $array:expr ),+ ) => { |
184 | | $crate::stack($axis, &[ $($crate::ArrayView::from(&$array) ),* ]).unwrap() |
185 | | }; |
186 | | } |
187 | | |
188 | | /// Concatenate arrays along the given axis. |
189 | | /// |
190 | | /// Uses the [`concatenate()`] function, calling `ArrayView::from(&a)` on each |
191 | | /// argument `a`. |
192 | | /// |
193 | | /// ***Panics*** if the `concatenate` function would return an error. |
194 | | /// |
195 | | /// ``` |
196 | | /// extern crate ndarray; |
197 | | /// |
198 | | /// use ndarray::{arr2, concatenate, Axis}; |
199 | | /// |
200 | | /// # fn main() { |
201 | | /// |
202 | | /// let a = arr2(&[[1., 2.], |
203 | | /// [3., 4.]]); |
204 | | /// assert_eq!( |
205 | | /// concatenate![Axis(0), a, a], |
206 | | /// arr2(&[[1., 2.], |
207 | | /// [3., 4.], |
208 | | /// [1., 2.], |
209 | | /// [3., 4.]]), |
210 | | /// ); |
211 | | /// assert_eq!( |
212 | | /// concatenate![Axis(1), a, a,], |
213 | | /// arr2(&[[1., 2., 1., 2.], |
214 | | /// [3., 4., 3., 4.]]), |
215 | | /// ); |
216 | | /// # } |
217 | | /// ``` |
218 | | #[macro_export] |
219 | | macro_rules! concatenate { |
220 | | ($axis:expr, $( $array:expr ),+ ,) => { |
221 | | $crate::concatenate!($axis, $($array),+) |
222 | | }; |
223 | | ($axis:expr, $( $array:expr ),+ ) => { |
224 | | $crate::concatenate($axis, &[ $($crate::ArrayView::from(&$array) ),* ]).unwrap() |
225 | | }; |
226 | | } |