rustyml 0.11.0

A high-performance machine learning & deep learning library in pure Rust, offering ML algorithms and neural network support
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
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

#![cfg(feature = "utility")]

use approx::assert_abs_diff_eq;
use ndarray::prelude::*;
use rustyml::error::ModelError;
use rustyml::utility::standardize::*;

const EPSILON: f64 = 1e-8;

/// Helper function to calculate mean and standard deviation for validation
fn calculate_mean_and_std(values: impl Iterator<Item = f64>) -> (f64, f64) {
    let values_vec: Vec<f64> = values.collect();
    let mean = values_vec.iter().sum::<f64>() / values_vec.len() as f64;
    let variance =
        values_vec.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / values_vec.len() as f64;
    let std_dev = variance.sqrt();
    (mean, std_dev)
}

#[test]
fn test_standardize_global() {
    let data = array![[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    // After global standardization, the entire array should have mean ≈ 0 and std ≈ 1
    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
}

#[test]
fn test_standardize_by_columns() {
    let data = array![[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]];
    let data2 = data.clone();
    let result = standardize(&data, StandardizationAxis::Column, EPSILON).unwrap();

    // Check mean and standard deviation for each column
    for col_idx in 0..data2.ncols() {
        let col = result.column(col_idx);
        let (mean, std_dev) = calculate_mean_and_std(col.iter().cloned());

        assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
        assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
    }
}

#[test]
fn test_standardize_by_rows() {
    let data = array![[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]];
    let data2 = data.clone();
    let result = standardize(&data, StandardizationAxis::Row, EPSILON).unwrap();

    // Check mean and standard deviation for each row
    for row_idx in 0..data2.nrows() {
        let row = result.row(row_idx);
        let (mean, std_dev) = calculate_mean_and_std(row.iter().cloned());

        assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
        assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
    }
}

#[test]
fn test_standardize_single_value() {
    let data = array![[5.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    // Single value should be standardized to 0 (since variance is 0, so (x-mean)/epsilon = 0)
    assert_abs_diff_eq!(result[[0, 0]], 0.0, epsilon = 1e-10);
}

#[test]
fn test_standardize_identical_values() {
    let data = array![[3.0, 3.0, 3.0], [3.0, 3.0, 3.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    // When all values are identical, standardization should result in all zeros
    for &val in result.iter() {
        assert_abs_diff_eq!(val, 0.0, epsilon = 1e-10);
    }
}

#[test]
fn test_standardize_1d_array() {
    let data = Array1::from(vec![1.0, 2.0, 3.0, 4.0, 5.0]);
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
}

#[test]
fn test_standardize_3d_array() {
    let data =
        Array::from_shape_vec((2, 2, 2), vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0]).unwrap();
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
}

#[test]
fn test_standardize_preserve_shape() {
    let data = array![[1.0, 2.0, 3.0], [4.0, 5.0, 6.0], [7.0, 8.0, 9.0]];
    let original_shape = data.shape().to_vec();
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    assert_eq!(result.shape(), &original_shape[..]);
}

#[test]
fn test_standardize_with_negative_values() {
    let data = array![[-1.0, -2.0], [3.0, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
}

// Input validation tests
#[test]
fn test_standardize_empty_array() {
    let data: Array2<f64> = Array2::zeros((0, 0));
    let result = standardize(&data, StandardizationAxis::Global, EPSILON);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Cannot standardize empty array"));
    }
}

#[test]
fn test_standardize_nan_values() {
    let data = array![[1.0, f64::NAN], [3.0, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Input contains NaN or infinite values"));
    }
}

#[test]
fn test_standardize_infinite_values() {
    let data = array![[1.0, 2.0], [f64::INFINITY, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Input contains NaN or infinite values"));
    }
}

#[test]
fn test_standardize_negative_epsilon() {
    let data = array![[1.0, 2.0], [3.0, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, -1e-8);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Epsilon must be positive and finite"));
    }
}

#[test]
fn test_standardize_zero_epsilon() {
    let data = array![[1.0, 2.0], [3.0, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, 0.0);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Epsilon must be positive and finite"));
    }
}

#[test]
fn test_standardize_infinite_epsilon() {
    let data = array![[1.0, 2.0], [3.0, 4.0]];
    let result = standardize(&data, StandardizationAxis::Global, f64::INFINITY);

    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
    if let Err(ModelError::InputValidationError(msg)) = result {
        assert!(msg.contains("Epsilon must be positive and finite"));
    }
}

#[test]
fn test_standardize_row_1d_array() {
    let data = Array1::from(vec![1.0, 2.0, 3.0]);
    let result = standardize(&data, StandardizationAxis::Row, EPSILON);

    // 1D array cannot be standardized by rows
    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
}

#[test]
fn test_standardize_column_1d_array() {
    let data = Array1::from(vec![1.0, 2.0, 3.0]);
    let result = standardize(&data, StandardizationAxis::Column, EPSILON);

    // 1D array cannot be standardized by columns
    assert!(matches!(result, Err(ModelError::InputValidationError(_))));
}

#[test]
fn test_standardize_large_values() {
    let data = array![[1e6, 2e6], [3e6, 4e6]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-10);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-10);
}

#[test]
fn test_standardize_small_values() {
    let data = array![[1e-6, 2e-6], [3e-6, 4e-6]];
    let result = standardize(&data, StandardizationAxis::Global, EPSILON).unwrap();

    let mean: f64 = result.iter().sum::<f64>() / result.len() as f64;
    let variance: f64 =
        result.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / result.len() as f64;
    let std_dev = variance.sqrt();

    assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-2);
    assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-2);
}

#[test]
fn test_standardization_axis_equality() {
    // Test StandardizationAxis enum equality
    assert_eq!(StandardizationAxis::Row, StandardizationAxis::Row);
    assert_eq!(StandardizationAxis::Column, StandardizationAxis::Column);
    assert_eq!(StandardizationAxis::Global, StandardizationAxis::Global);

    assert_ne!(StandardizationAxis::Row, StandardizationAxis::Column);
    assert_ne!(StandardizationAxis::Row, StandardizationAxis::Global);
    assert_ne!(StandardizationAxis::Column, StandardizationAxis::Global);
}

#[test]
fn test_standardization_axis_debug() {
    // Test StandardizationAxis Debug implementation
    let row_axis = StandardizationAxis::Row;
    let debug_str = format!("{:?}", row_axis);
    assert_eq!(debug_str, "Row");
}

#[test]
fn test_standardization_axis_clone() {
    // Test StandardizationAxis Clone implementation
    let original = StandardizationAxis::Column;
    let cloned = original.clone();
    assert_eq!(original, cloned);
}

#[test]
fn test_standardization_axis_copy() {
    // Test StandardizationAxis Copy implementation
    let original = StandardizationAxis::Global;
    let copied = original;
    assert_eq!(original, copied);
}

#[test]
fn test_standardize_column_axis_large_array() {
    // Test column standardization on large array
    let data = Array2::from_shape_fn((100, 50), |(i, j)| (i as f64 + j as f64) * 0.1);
    let data2 = data.clone();
    let result = standardize(&data, StandardizationAxis::Column, EPSILON).unwrap();

    // Check mean and standard deviation for each column
    for col_idx in 0..data2.ncols() {
        let col = result.column(col_idx);
        let (mean, std_dev) = calculate_mean_and_std(col.iter().cloned());

        assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
        assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
    }
}

#[test]
fn test_standardize_row_axis_large_array() {
    // Test row standardization on large array
    let data = Array2::from_shape_fn((50, 100), |(i, j)| (i as f64 + j as f64) * 0.1);
    let data2 = data.clone();
    let result = standardize(&data, StandardizationAxis::Row, EPSILON).unwrap();

    // Check mean and standard deviation for each row
    for row_idx in 0..data2.nrows() {
        let row = result.row(row_idx);
        let mean: f64 = row.iter().sum::<f64>() / row.len() as f64;
        let variance: f64 = row.iter().map(|&x| (x - mean).powi(2)).sum::<f64>() / row.len() as f64;
        let std_dev = variance.sqrt();

        assert_abs_diff_eq!(mean, 0.0, epsilon = 1e-4);
        assert_abs_diff_eq!(std_dev, 1.0, epsilon = 1e-4);
    }
}