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
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
//! Shape manipulation: reshape, transpose, flatten, squeeze, unsqueeze,
//! concatenate, and stack.
use crate::Scalar;
use crate::error::{CoreError, Result};
use super::{Tensor, compute_strides};
impl<T: Scalar> Tensor<T> {
/// Reshape the tensor to a new shape without copying data.
///
/// The total number of elements must remain the same.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3, 4, 5, 6], vec![6]).unwrap();
/// let t = t.reshape(vec![2, 3]).unwrap();
/// assert_eq!(t.shape(), &[2, 3]);
/// ```
pub fn reshape(mut self, new_shape: Vec<usize>) -> Result<Self> {
let new_numel: usize = new_shape.iter().product();
if new_numel != self.numel() {
return Err(CoreError::InvalidShape {
shape: new_shape,
reason: "new shape has different number of elements",
});
}
self.strides = compute_strides(&new_shape);
self.shape = new_shape;
Ok(self)
}
/// Return a reshaped view without consuming the tensor (copies data).
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3, 4], vec![4]).unwrap();
/// let r = t.reshaped(vec![2, 2]).unwrap();
/// assert_eq!(r.shape(), &[2, 2]);
/// assert_eq!(t.shape(), &[4]); // original unchanged
/// ```
pub fn reshaped(&self, new_shape: Vec<usize>) -> Result<Self> {
self.clone().reshape(new_shape)
}
/// Flatten the tensor into a 1-D tensor (consumes self, no copy).
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3, 4], vec![2, 2]).unwrap();
/// let flat = t.flatten();
/// assert_eq!(flat.shape(), &[4]);
/// ```
pub fn flatten(self) -> Self {
let n = self.numel();
Tensor {
data: self.data,
shape: vec![n],
strides: vec![1],
}
}
/// Return a flattened copy of the tensor.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3, 4], vec![2, 2]).unwrap();
/// let flat = t.flattened();
/// assert_eq!(flat.shape(), &[4]);
/// assert_eq!(t.shape(), &[2, 2]); // original unchanged
/// ```
pub fn flattened(&self) -> Self {
let n = self.numel();
Tensor {
data: self.data.clone(),
shape: vec![n],
strides: vec![1],
}
}
/// Transpose a 2-D tensor (matrix). Returns a new tensor with copied data.
///
/// For higher-rank tensors, use [`permute`](Self::permute).
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3, 4, 5, 6], vec![2, 3]).unwrap();
/// let tt = t.transpose().unwrap();
/// assert_eq!(tt.shape(), &[3, 2]);
/// ```
pub fn transpose(&self) -> Result<Self> {
if self.ndim() != 2 {
return Err(CoreError::InvalidArgument {
reason: "transpose() requires a 2-D tensor; use permute() for higher ranks",
});
}
let (rows, cols) = (self.shape[0], self.shape[1]);
let mut data = vec![T::zero(); self.numel()];
for r in 0..rows {
for c in 0..cols {
data[c * rows + r] = self.data[r * cols + c];
}
}
Tensor::from_vec(data, vec![cols, rows])
}
/// Permute the dimensions of the tensor according to the given axes.
///
/// `axes` must be a permutation of `0..ndim`.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::<i32>::arange(24).reshape(vec![2, 3, 4]).unwrap();
/// let p = t.permute(&[2, 0, 1]).unwrap();
/// assert_eq!(p.shape(), &[4, 2, 3]);
/// ```
pub fn permute(&self, axes: &[usize]) -> Result<Self> {
if axes.len() != self.ndim() {
return Err(CoreError::InvalidArgument {
reason: "axes length must match tensor rank",
});
}
// Validate it's a valid permutation
let mut seen = vec![false; self.ndim()];
for &a in axes {
if a >= self.ndim() {
return Err(CoreError::AxisOutOfBounds {
axis: a,
ndim: self.ndim(),
});
}
if seen[a] {
return Err(CoreError::InvalidArgument {
reason: "duplicate axis in permutation",
});
}
seen[a] = true;
}
let new_shape: Vec<usize> = axes.iter().map(|&a| self.shape[a]).collect();
let new_strides = compute_strides(&new_shape);
let new_numel: usize = new_shape.iter().product();
let mut data = vec![T::zero(); new_numel];
// Iterate over every element in the output
let mut out_index = vec![0usize; self.ndim()];
for item in &mut data {
// Map output index back to input index
let mut flat_in = 0;
for (out_ax, &in_ax) in axes.iter().enumerate() {
flat_in += out_index[out_ax] * self.strides[in_ax];
}
*item = self.data[flat_in];
// Increment the output index (odometer style)
for d in (0..self.ndim()).rev() {
out_index[d] += 1;
if out_index[d] < new_shape[d] {
break;
}
out_index[d] = 0;
}
}
Ok(Tensor {
data,
shape: new_shape,
strides: new_strides,
})
}
/// Insert a dimension of size 1 at the given axis.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
/// let t = t.unsqueeze(0).unwrap();
/// assert_eq!(t.shape(), &[1, 3]);
/// ```
pub fn unsqueeze(mut self, axis: usize) -> Result<Self> {
if axis > self.ndim() {
return Err(CoreError::AxisOutOfBounds {
axis,
ndim: self.ndim(),
});
}
self.shape.insert(axis, 1);
self.strides = compute_strides(&self.shape);
Ok(self)
}
/// Remove all dimensions of size 1.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let t = Tensor::from_vec(vec![1, 2, 3], vec![1, 3, 1]).unwrap();
/// let t = t.squeeze();
/// assert_eq!(t.shape(), &[3]);
/// ```
pub fn squeeze(mut self) -> Self {
self.shape.retain(|&d| d != 1);
if self.shape.is_empty() && self.numel() == 1 {
self.shape = vec![];
}
self.strides = compute_strides(&self.shape);
self
}
/// Concatenate a list of tensors along the given axis.
///
/// All tensors must have the same shape except along the concatenation axis.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let a = Tensor::from_vec(vec![1, 2, 3], vec![1, 3]).unwrap();
/// let b = Tensor::from_vec(vec![4, 5, 6], vec![1, 3]).unwrap();
/// let c = Tensor::concat(&[&a, &b], 0).unwrap();
/// assert_eq!(c.shape(), &[2, 3]);
/// ```
pub fn concat(tensors: &[&Tensor<T>], axis: usize) -> Result<Self> {
if tensors.is_empty() {
return Err(CoreError::InvalidArgument {
reason: "cannot concatenate zero tensors",
});
}
let ndim = tensors[0].ndim();
if axis >= ndim {
return Err(CoreError::AxisOutOfBounds { axis, ndim });
}
// Validate shapes match on all axes except `axis`
for t in &tensors[1..] {
if t.ndim() != ndim {
return Err(CoreError::DimensionMismatch {
expected: tensors[0].shape.clone(),
got: t.shape.clone(),
});
}
for (d, (&a, &b)) in tensors[0].shape.iter().zip(t.shape.iter()).enumerate() {
if d != axis && a != b {
return Err(CoreError::DimensionMismatch {
expected: tensors[0].shape.clone(),
got: t.shape.clone(),
});
}
}
}
let mut new_shape = tensors[0].shape.clone();
new_shape[axis] = tensors.iter().map(|t| t.shape[axis]).sum();
let outer: usize = new_shape[..axis].iter().product();
let inner: usize = new_shape[axis + 1..].iter().product();
let total: usize = new_shape.iter().product();
let mut data = Vec::with_capacity(total);
for o in 0..outer {
for t in tensors {
let axis_len = t.shape[axis];
let src_start = o * axis_len * inner;
let src_end = src_start + axis_len * inner;
data.extend_from_slice(&t.data[src_start..src_end]);
}
}
Tensor::from_vec(data, new_shape)
}
/// Stack tensors along a new axis inserted at position `axis`.
///
/// All tensors must have identical shapes.
///
/// # Examples
///
/// ```
/// # use scivex_core::Tensor;
/// let a = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
/// let b = Tensor::from_vec(vec![4, 5, 6], vec![3]).unwrap();
/// let c = Tensor::stack(&[&a, &b], 0).unwrap();
/// assert_eq!(c.shape(), &[2, 3]);
/// ```
pub fn stack(tensors: &[&Tensor<T>], axis: usize) -> Result<Self> {
if tensors.is_empty() {
return Err(CoreError::InvalidArgument {
reason: "cannot stack zero tensors",
});
}
let base_shape = &tensors[0].shape;
if axis > base_shape.len() {
return Err(CoreError::AxisOutOfBounds {
axis,
ndim: base_shape.len() + 1,
});
}
for t in &tensors[1..] {
if t.shape != *base_shape {
return Err(CoreError::DimensionMismatch {
expected: base_shape.clone(),
got: t.shape.clone(),
});
}
}
// Unsqueeze each tensor along the new axis, then concat
let expanded: Vec<Tensor<T>> = tensors
.iter()
// SAFETY: axis is validated above and is within bounds for all tensors.
.map(|t| {
(*t).clone()
.unsqueeze(axis)
.expect("axis is valid for all tensors since shapes were validated above")
})
.collect();
let refs: Vec<&Tensor<T>> = expanded.iter().collect();
Tensor::concat(&refs, axis)
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_reshape() {
let t = Tensor::from_vec(vec![1, 2, 3, 4, 5, 6], vec![6]).unwrap();
let t = t.reshape(vec![2, 3]).unwrap();
assert_eq!(t.shape(), &[2, 3]);
assert_eq!(t.strides(), &[3, 1]);
assert_eq!(*t.get(&[1, 0]).unwrap(), 4);
}
#[test]
fn test_reshape_invalid() {
let t = Tensor::from_vec(vec![1, 2, 3, 4], vec![4]).unwrap();
assert!(t.reshape(vec![3, 2]).is_err());
}
#[test]
fn test_flatten() {
let t = Tensor::from_vec(vec![1, 2, 3, 4, 5, 6], vec![2, 3]).unwrap();
let flat = t.flatten();
assert_eq!(flat.shape(), &[6]);
assert_eq!(flat.as_slice(), &[1, 2, 3, 4, 5, 6]);
}
#[test]
fn test_transpose() {
// [[1, 2, 3],
// [4, 5, 6]]
let t = Tensor::from_vec(vec![1, 2, 3, 4, 5, 6], vec![2, 3]).unwrap();
let tt = t.transpose().unwrap();
assert_eq!(tt.shape(), &[3, 2]);
assert_eq!(*tt.get(&[0, 0]).unwrap(), 1);
assert_eq!(*tt.get(&[0, 1]).unwrap(), 4);
assert_eq!(*tt.get(&[2, 0]).unwrap(), 3);
assert_eq!(*tt.get(&[2, 1]).unwrap(), 6);
}
#[test]
fn test_transpose_not_2d() {
let t = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
assert!(t.transpose().is_err());
}
#[test]
fn test_permute() {
// Shape [2, 3, 4] -> permute [2, 0, 1] -> shape [4, 2, 3]
let t = Tensor::<i32>::arange(24).reshape(vec![2, 3, 4]).unwrap();
let p = t.permute(&[2, 0, 1]).unwrap();
assert_eq!(p.shape(), &[4, 2, 3]);
// Element at [0, 0, 0] in original is at [0, 0, 0] in permuted
assert_eq!(*p.get(&[0, 0, 0]).unwrap(), 0);
// Element at [1, 2, 3] in original -> permuted[3, 1, 2]
assert_eq!(*p.get(&[3, 1, 2]).unwrap(), *t.get(&[1, 2, 3]).unwrap());
}
#[test]
fn test_unsqueeze_squeeze() {
let t = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
let t = t.unsqueeze(0).unwrap();
assert_eq!(t.shape(), &[1, 3]);
let t = t.squeeze();
assert_eq!(t.shape(), &[3]);
}
#[test]
fn test_concat() {
let a = Tensor::from_vec(vec![1, 2, 3], vec![1, 3]).unwrap();
let b = Tensor::from_vec(vec![4, 5, 6], vec![1, 3]).unwrap();
let c = Tensor::concat(&[&a, &b], 0).unwrap();
assert_eq!(c.shape(), &[2, 3]);
assert_eq!(c.as_slice(), &[1, 2, 3, 4, 5, 6]);
}
#[test]
fn test_concat_axis1() {
let a = Tensor::from_vec(vec![1, 2, 3, 4], vec![2, 2]).unwrap();
let b = Tensor::from_vec(vec![5, 6, 7, 8], vec![2, 2]).unwrap();
let c = Tensor::concat(&[&a, &b], 1).unwrap();
assert_eq!(c.shape(), &[2, 4]);
assert_eq!(c.as_slice(), &[1, 2, 5, 6, 3, 4, 7, 8]);
}
#[test]
fn test_stack() {
let a = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
let b = Tensor::from_vec(vec![4, 5, 6], vec![3]).unwrap();
let c = Tensor::stack(&[&a, &b], 0).unwrap();
assert_eq!(c.shape(), &[2, 3]);
assert_eq!(c.as_slice(), &[1, 2, 3, 4, 5, 6]);
}
#[test]
fn test_stack_axis1() {
let a = Tensor::from_vec(vec![1, 2, 3], vec![3]).unwrap();
let b = Tensor::from_vec(vec![4, 5, 6], vec![3]).unwrap();
let c = Tensor::stack(&[&a, &b], 1).unwrap();
assert_eq!(c.shape(), &[3, 2]);
assert_eq!(c.as_slice(), &[1, 4, 2, 5, 3, 6]);
}
#[test]
fn test_concat_shape_mismatch() {
let a = Tensor::from_vec(vec![1, 2, 3], vec![1, 3]).unwrap();
let b = Tensor::from_vec(vec![4, 5], vec![1, 2]).unwrap();
assert!(Tensor::concat(&[&a, &b], 0).is_err());
}
}