deep_causality_tensor 0.2.0

Tensor data structure for for deep_causality crate.
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

/*
 * SPDX-License-Identifier: MIT
 * Copyright (c) "2025" . The DeepCausality Authors and Contributors. All Rights Reserved.
 */

use deep_causality_tensor::{CausalTensor, CausalTensorMathExt, CausalTensorStackExt, Tensor};

pub fn main() {
    println!("\n--- CausalTensor Example ---");

    // 1. Creating a new CausalTensor
    // Tensors can be created from a flat Vec of data and a shape.
    println!("\n1. Creating a 2x3 tensor:");
    let data = vec![1, 2, 3, 4, 5, 6];
    let shape = vec![2, 3];
    let tensor = CausalTensor::new(data, shape).unwrap();
    println!("   Tensor: {}", tensor);
    println!("   Shape: {:?}", tensor.shape());
    println!("   Is empty: {}", tensor.is_empty());
    println!("   Number of dimensions: {}", tensor.num_dim());
    println!("   Total elements: {}", tensor.len());

    // 2. Accessing elements
    // Elements can be accessed using their multi-dimensional index.
    println!("\n2. Accessing element at [1, 2]:");
    let element = tensor.get(&[1, 2]).unwrap();
    println!("   Value: {}", element);
    assert_eq!(*element, 6);

    // 3. Shape Manipulation
    // Tensors can be reshaped or flattened without copying the underlying data.
    println!("\n3. Reshaping the tensor to 3x2:");
    let reshaped_tensor = tensor.reshape(&[3, 2]).unwrap();
    println!("   Reshaped Tensor: {}", reshaped_tensor);
    assert_eq!(reshaped_tensor.shape(), &[3, 2]);

    println!("\n   Flattening the tensor (ravel):");
    let raveled_tensor = tensor.ravel(); // Consumes the original tensor
    println!("   Raveled Tensor: {}", raveled_tensor);
    assert_eq!(raveled_tensor.shape(), &[6]);

    // 4. Tensor-Scalar Arithmetic
    // Tensors support element-wise arithmetic with scalars.
    println!("\n4. Tensor-Scalar Arithmetic (add 10 to each element):");
    let tensor = CausalTensor::new(vec![1.0, 2.0, 3.0, 4.0], vec![2, 2]).unwrap();
    let added_tensor = &tensor + 10.0;
    println!("   Original: {}", tensor);
    println!("   Result:   {}", added_tensor);
    assert_eq!(added_tensor.as_slice(), &[11.0, 12.0, 13.0, 14.0]);

    // 5. Reduction Operations
    // You can reduce tensors along axes using operations like sum and mean.
    println!("\n5. Reduction Operations on a 2x3 tensor:");
    let tensor = CausalTensor::new(vec![1, 2, 3, 4, 5, 6], vec![2, 3]).unwrap();
    println!("   Original Tensor: {}", tensor);

    // Sum along axis 0 (columns)
    let sum_axis0 = tensor.sum_axes(&[0]).unwrap();
    println!("   Sum along axis 0: {}", sum_axis0);
    assert_eq!(sum_axis0.as_slice(), &[5, 7, 9]);

    // Mean of all elements (full reduction)
    let tensor_f64 = CausalTensor::new(vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0], vec![2, 3]).unwrap();
    let mean_all = tensor_f64.mean_axes(&[]).unwrap();
    println!("   Mean of all elements: {}", mean_all);
    assert_eq!(mean_all.as_slice(), &[3.5]);

    // 6. Sorting a 1D Tensor
    // You can get the indices that would sort a 1D tensor.
    println!("\n6. Sorting a 1D tensor:");
    let tensor_1d = CausalTensor::new(vec![3, 1, 4, 1, 5, 9], vec![6]).unwrap();
    println!("   Original 1D Tensor: {}", tensor_1d);
    let sorted_indices = tensor_1d.arg_sort().unwrap();
    println!("   Sorted indices: {:?}", sorted_indices);
    assert_eq!(sorted_indices, vec![1, 3, 0, 2, 4, 5]);

    // 7. Tensor-Tensor Arithmetic
    // Tensors can be added, subtracted, etc., with other tensors.
    // This supports broadcasting for compatible shapes.
    println!("\n7. Tensor-Tensor Addition with Broadcasting:");
    let t1 = CausalTensor::new(vec![1, 2, 3, 4, 5, 6], vec![2, 3]).unwrap();
    // This tensor will be broadcasted to match the shape of t1.
    let t2 = CausalTensor::new(vec![10, 20, 30], vec![1, 3]).unwrap();
    println!("   Tensor 1: {}", t1);
    println!("   Tensor 2 (to be broadcasted): {}", t2);

    // Tensor ops can be by value, by ref, or any combination i.e. &t1 + t2
    let result = &t1 + &t2;
    println!("   Result (t1 + t2): {}", result);
    assert_eq!(result.as_slice(), &[11, 22, 33, 14, 25, 36]);

    // 8. Logarithmic Functions. Import CausalTensorLogMathExt
    // Tensors with floating-point data support element-wise logarithmic operations.
    println!("\n8. Logarithmic Functions on a 2x3 tensor:");
    let tensor_f64_log = CausalTensor::new(
        vec![1.0, std::f64::consts::E, 10.0, 100.0, 4.0, 16.0],
        vec![2, 3],
    )
    .unwrap();
    println!("   Original Tensor: {}", tensor_f64_log);

    let log_nat_tensor = tensor_f64_log.log_nat().unwrap();
    println!("   Natural Log (ln): {}", log_nat_tensor);

    let log2_tensor = tensor_f64_log.log2().unwrap();
    println!("   Base 2 Log (log2): {}", log2_tensor);

    let log10_tensor = tensor_f64_log.log10().unwrap();
    println!("   Base 10 Log (log10): {}", log10_tensor);

    // 9. Stacking Tensors. Import CausalTensorCollectionExt
    // A slice of tensors can be stacked along a new axis.
    println!("\n9. Stacking two 2-element vectors:");
    let tensor_a = CausalTensor::<i32>::new(vec![1, 2], vec![2]).unwrap();
    println!("   Tensor A: {}", tensor_a);
    let tensor_b = CausalTensor::<i32>::new(vec![3, 4], vec![2]).unwrap();
    println!("   Tensor B: {}", tensor_b);

    let tensors_to_stack = [tensor_a, tensor_b];

    // Stack along axis 0
    let stacked_axis_0 = tensors_to_stack.stack(0).unwrap();
    println!(
        "   Stacked along axis 0 (new shape {{:?}}): {{}} {:?} {:?}",
        stacked_axis_0.shape(),
        stacked_axis_0
    );
    assert_eq!(stacked_axis_0.shape(), &[2, 2]);
    assert_eq!(stacked_axis_0.as_slice(), &[1, 2, 3, 4]);

    // Stack along axis 1
    let stacked_axis_1 = tensors_to_stack.stack(1).unwrap();
    println!(
        "   Stacked along axis 1 (new shape {{:?}}): {{}} {:?} {:?}",
        stacked_axis_1.shape(),
        stacked_axis_1
    );
    assert_eq!(stacked_axis_1.shape(), &[2, 2]);
    assert_eq!(stacked_axis_1.as_slice(), &[1, 3, 2, 4]);

    println!("\nAll examples executed successfully!");
}