nauticuvs 0.1.2

Fast Discrete Curvelet Transform (FDCT) via wrapping for 2D images
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

//! Unit and integration tests for the curvelet transform crate.

use ndarray::Array2;

use crate::{curvelet_forward, curvelet_inverse, CurveletConfig, CurveletError};

/// Helper: compute relative L2 error between two f32 arrays.
fn relative_l2_error(original: &Array2<f32>, reconstructed: &Array2<f32>) -> f64 {
    let diff_norm_sq: f64 = original
        .iter()
        .zip(reconstructed.iter())
        .map(|(a, b)| (*a as f64 - *b as f64).powi(2))
        .sum();
    let orig_norm_sq: f64 = original.iter().map(|a| (*a as f64).powi(2)).sum();
    if orig_norm_sq < 1e-24 {
        return 0.0;
    }
    (diff_norm_sq / orig_norm_sq).sqrt()
}

// ===== Configuration tests =====

#[test]
fn test_config_valid() {
    let cfg = CurveletConfig::new(5).unwrap();
    assert_eq!(cfg.num_scales, 5);
    assert_eq!(cfg.num_detail_scales(), 3);
}

#[test]
fn test_config_invalid_scale_count() {
    assert!(matches!(
        CurveletConfig::new(1),
        Err(CurveletError::InvalidScaleCount(1))
    ));
    assert!(matches!(
        CurveletConfig::new(11),
        Err(CurveletError::InvalidScaleCount(11))
    ));
}

#[test]
fn test_config_direction_doubling() {
    let cfg = CurveletConfig::new(5).unwrap();
    let d0 = cfg.directions_at_detail_scale(0);
    let d1 = cfg.directions_at_detail_scale(1);
    let d2 = cfg.directions_at_detail_scale(2);
    assert!(d2 >= d1);
    assert!(d1 >= d0);
    assert_eq!(d2, 32);
}

#[test]
fn test_config_custom_directions() {
    let cfg = CurveletConfig::new(5)
        .unwrap()
        .with_directions_per_scale(vec![8, 16, 32])
        .unwrap();
    assert_eq!(cfg.directions_at_detail_scale(0), 8);
    assert_eq!(cfg.directions_at_detail_scale(1), 16);
    assert_eq!(cfg.directions_at_detail_scale(2), 32);
}

#[test]
fn test_config_custom_directions_wrong_length() {
    let result = CurveletConfig::new(5)
        .unwrap()
        .with_directions_per_scale(vec![8, 16]);
    assert!(matches!(
        result,
        Err(CurveletError::DirectionCountMismatch {
            expected: 3,
            got: 2
        })
    ));
}

#[test]
fn test_config_invalid_direction_count() {
    let result = CurveletConfig::new(5).unwrap().with_finest_directions(3);
    assert!(matches!(
        result,
        Err(CurveletError::InvalidDirectionCount(3))
    ));
}

// ===== Input validation tests =====

#[test]
fn test_zero_dimension_error() {
    let img = Array2::<f32>::zeros((0, 10));
    assert!(matches!(
        curvelet_forward(&img, 3),
        Err(CurveletError::ZeroDimension { .. })
    ));
}

#[test]
fn test_non_finite_error() {
    let mut img = Array2::<f32>::zeros((8, 8));
    img[[0, 0]] = f32::NAN;
    assert!(matches!(
        curvelet_forward(&img, 3),
        Err(CurveletError::NonFiniteInput)
    ));
}

// ===== Reconstruction tests (critical correctness) =====

#[test]
fn test_reconstruction_zeros() {
    let img = Array2::<f32>::zeros((16, 16));
    let coeffs = curvelet_forward(&img, 3).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    assert_eq!(recon.dim(), img.dim());
    let max_err: f32 = recon.iter().map(|v| v.abs()).fold(0.0f32, f32::max);
    assert!(
        max_err < 1e-10,
        "zero image reconstruction error: {max_err}"
    );
}

#[test]
fn test_reconstruction_constant() {
    let img = Array2::from_elem((16, 16), 42.0f32);
    let coeffs = curvelet_forward(&img, 3).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "constant image reconstruction error: {err}");
}

#[test]
fn test_reconstruction_gradient() {
    let n = 32;
    let img = Array2::from_shape_fn((n, n), |(r, c)| (r as f32 + c as f32) / (2.0 * n as f32));
    let coeffs = curvelet_forward(&img, 4).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "gradient reconstruction error: {err}");
}

#[test]
fn test_reconstruction_sinusoid() {
    let n = 64;
    let img = Array2::from_shape_fn((n, n), |(r, c)| {
        let x = r as f32 / n as f32;
        let y = c as f32 / n as f32;
        (2.0 * std::f32::consts::PI * 3.0 * x).sin() + (2.0 * std::f32::consts::PI * 5.0 * y).cos()
    });
    let coeffs = curvelet_forward(&img, 5).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "sinusoid reconstruction error: {err}");
}

#[test]
fn test_reconstruction_random() {
    use rand::Rng;
    let n = 32;
    let mut rng = rand::thread_rng();
    let img = Array2::from_shape_fn((n, n), |_| rng.gen::<f32>());
    let coeffs = curvelet_forward(&img, 4).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "random reconstruction error: {err}");
}

#[test]
fn test_reconstruction_non_square() {
    let img = Array2::from_shape_fn((24, 32), |(r, c)| (r + c) as f32);
    let coeffs = curvelet_forward(&img, 3).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    assert_eq!(recon.dim(), (24, 32));
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "non-square reconstruction error: {err}");
}

#[test]
fn test_reconstruction_5_scales_64() {
    // Previously the worst case: 64×64 with 5 scales
    let n = 64;
    let img = Array2::from_shape_fn((n, n), |(r, c)| {
        ((r as f32 * 0.1).sin() + (c as f32 * 0.07).cos()) * 100.0
    });
    let coeffs = curvelet_forward(&img, 5).unwrap();
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "5-scale 64×64 reconstruction error: {err}");
}

// ===== Coefficient structure =====

#[test]
fn test_coefficient_structure() {
    let img = Array2::from_shape_fn((32, 32), |(r, c)| (r * c) as f32 / 1024.0);
    let coeffs = curvelet_forward(&img, 4).unwrap();
    assert_eq!(coeffs.detail.len(), 2);
    let cfg = CurveletConfig::new(4).unwrap();
    for d in 0..2 {
        assert_eq!(coeffs.detail[d].len(), cfg.directions_at_detail_scale(d));
    }
}

// ===== Thresholding =====

#[test]
fn test_hard_threshold() {
    let img = Array2::from_shape_fn((16, 16), |(r, c)| (r + c) as f32);
    let mut coeffs = curvelet_forward(&img, 3).unwrap();
    let before = coeffs.num_coeffs();
    coeffs.hard_threshold(100.0);
    let zero_count: usize = coeffs
        .detail
        .iter()
        .flat_map(|s| s.iter())
        .flat_map(|sb| sb.iter())
        .filter(|c| c.norm() < 1e-10)
        .count();
    assert!(zero_count > 0);
    assert_eq!(coeffs.num_coeffs(), before);
}

#[test]
fn test_soft_threshold() {
    let img = Array2::from_shape_fn((16, 16), |(r, c)| (r + c) as f32);
    let mut coeffs = curvelet_forward(&img, 3).unwrap();
    coeffs.soft_threshold(50.0);
    for scale in &coeffs.detail {
        for sb in scale {
            for c in sb.iter() {
                assert!(c.norm() >= 0.0);
            }
        }
    }
}

// ===== Edge cases =====

#[test]
fn test_two_scales() {
    let img = Array2::from_elem((8, 8), 1.0f32);
    let coeffs = curvelet_forward(&img, 2).unwrap();
    assert!(coeffs.detail.is_empty());
    let recon = curvelet_inverse(&coeffs).unwrap();
    let err = relative_l2_error(&img, &recon);
    assert!(err < 1e-6, "2-scale reconstruction error: {err}");
}