svod-tensor 0.1.0-alpha.3

High-level lazy tensor API for the Svod ML compiler
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
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167

use ndarray::{Array1, Array3, array};

use crate::Tensor;

// =========================================================================
// Codegen-required tests
// =========================================================================

crate::codegen_tests! {
    fn test_to_vec_computed(config) {
        let a = Tensor::from_slice([1.0f32, 2.0, 3.0]);
        let b = Tensor::from_slice([10.0f32, 20.0, 30.0]);
        let mut c = &a + &b;
        c.realize_with(&config).unwrap();
        let v = c.as_vec::<f32>().unwrap();
        assert_eq!(v, vec![11.0, 22.0, 33.0]);
    }
}

// === from_ndarray ===

#[test]
fn test_from_ndarray_1d() {
    let arr = array![1.0f32, 2.0, 3.0];
    let t = Tensor::from_ndarray(&arr);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view.shape(), &[3]);
    assert_eq!(view, arr.view().into_dyn());
}

#[test]
fn test_from_ndarray_2d() {
    let arr = array![[1i32, 2], [3, 4]];
    let t = Tensor::from_ndarray(&arr);
    let view = t.array_view::<i32>().unwrap();
    assert_eq!(view.shape(), &[2, 2]);
    assert_eq!(view, arr.view().into_dyn());
}

#[test]
fn test_from_ndarray_3d() {
    let arr = Array3::from_shape_vec((2, 1, 3), vec![1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0]).unwrap();
    let t = Tensor::from_ndarray(&arr);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view.shape(), &[2, 1, 3]);
    assert_eq!(view, arr.view().into_dyn());
}

#[test]
fn test_from_ndarray_0d() {
    let arr = ndarray::arr0(42.0f32);
    let t = Tensor::from_ndarray(&arr);
    let v = t.as_vec::<f32>().unwrap();
    assert_eq!(v, vec![42.0]);
}

#[test]
fn test_from_ndarray_empty() {
    let arr = Array1::<f32>::zeros(0);
    let t = Tensor::from_ndarray(&arr);
    let result = t.as_ndarray::<f32>().unwrap();
    assert_eq!(result.shape(), &[0]);
    assert_eq!(result.len(), 0);
}

#[test]
fn test_from_ndarray_view() {
    let arr = array![10.0f32, 20.0, 30.0];
    let view = arr.view();
    let t = Tensor::from_ndarray(&view);
    let tv = t.array_view::<f32>().unwrap();
    assert_eq!(tv, arr.view().into_dyn());
}

#[test]
fn test_from_ndarray_fortran() {
    // Fortran (column-major) — hits slow path, still correct logical order
    let arr = ndarray::Array2::from_shape_vec(ndarray::ShapeBuilder::f((2, 2)), vec![1.0f32, 3.0, 2.0, 4.0]).unwrap();
    let t = Tensor::from_ndarray(&arr);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view.shape(), &[2, 2]);
    assert_eq!(view, arr.view().into_dyn());
}

// === to_vec ===

#[test]
fn test_to_vec_f32() {
    let t = Tensor::from_slice([1.0f32, 2.0, 3.0]);
    let v = t.as_vec::<f32>().unwrap();
    assert_eq!(v, vec![1.0, 2.0, 3.0]);
}

// === array_view ===

#[test]
fn test_array_view_basic() {
    let arr = array![[1.0f32, 2.0], [3.0, 4.0]];
    let t = Tensor::from_ndarray(&arr);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view, arr.view().into_dyn());
}

#[test]
fn test_array_view_on_from_slice() {
    // from_slice retains buffer — no realize needed
    let t = Tensor::from_slice([1.0f32, 2.0, 3.0]);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view.as_slice().unwrap(), &[1.0, 2.0, 3.0]);
}

#[test]
fn test_array_view_unrealized() {
    let a = Tensor::from_slice([1.0f32, 2.0]);
    let b = Tensor::from_slice([3.0f32, 4.0]);
    let c = &a + &b; // lazy, no buffer
    assert!(c.array_view::<f32>().is_err());
}

#[test]
fn test_array_view_dtype_mismatch() {
    let t = Tensor::from_slice([1.0f32, 2.0, 3.0]);
    assert!(t.array_view::<i32>().is_err());
}

// === array_view_mut ===

#[test]
fn test_array_view_mut_write() {
    let t = Tensor::from_ndarray(&ndarray::Array2::<f32>::zeros((2, 3)));
    t.array_view_mut::<f32>().unwrap()[[1, 2]] = 42.0;
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view[[1, 2]], 42.0);
    assert_eq!(view[[0, 0]], 0.0);
}

#[test]
fn test_array_view_mut_fill() {
    let t = Tensor::from_slice([0.0f32; 4]);
    t.array_view_mut::<f32>().unwrap().fill(7.0);
    assert_eq!(t.array_view::<f32>().unwrap().as_slice().unwrap(), &[7.0; 4]);
}

#[test]
fn test_array_view_mut_unrealized() {
    let a = Tensor::from_slice([1.0f32, 2.0]);
    let c = &a + &a;
    assert!(c.array_view_mut::<f32>().is_err());
}

// === roundtrip ===

#[test]
fn test_roundtrip_ndarray() {
    let original = array![[1.0f32, 2.0, 3.0], [4.0, 5.0, 6.0]];
    let t = Tensor::from_ndarray(&original);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view, original.view().into_dyn());
}

#[test]
fn test_roundtrip_vec() {
    let original = vec![1.0f32, 2.0, 3.0, 4.0, 5.0];
    let t = Tensor::from_slice(&original);
    let view = t.array_view::<f32>().unwrap();
    assert_eq!(view.as_slice().unwrap(), &original);
}