sklears_kernel_approximation/
simple_test.rs

1//! Simple manual test for kernel approximation implementations
2
3use crate::nystroem::*;
4use crate::rbf_sampler::*;
5use scirs2_core::ndarray::array;
6use sklears_core::traits::{Fit, Transform};
7
8pub fn test_implementations() {
9    println!("Testing kernel approximation implementations...");
10
11    // Test data
12    let x = array![[1.0, 2.0], [3.0, 4.0], [5.0, 6.0], [7.0, 8.0]];
13
14    // Test RBF Sampler
15    println!("Testing RBF Sampler...");
16    match test_rbf_sampler(&x) {
17        Ok(_) => println!("✓ RBF Sampler works"),
18        Err(e) => println!("✗ RBF Sampler failed: {}", e),
19    }
20
21    // Test Laplacian Sampler
22    println!("Testing Laplacian Sampler...");
23    match test_laplacian_sampler(&x) {
24        Ok(_) => println!("✓ Laplacian Sampler works"),
25        Err(e) => println!("✗ Laplacian Sampler failed: {}", e),
26    }
27
28    // Test Polynomial Sampler
29    println!("Testing Polynomial Sampler...");
30    match test_polynomial_sampler(&x) {
31        Ok(_) => println!("✓ Polynomial Sampler works"),
32        Err(e) => println!("✗ Polynomial Sampler failed: {}", e),
33    }
34
35    // Test Arc-cosine Sampler
36    println!("Testing Arc-cosine Sampler...");
37    match test_arc_cosine_sampler(&x) {
38        Ok(_) => println!("✓ Arc-cosine Sampler works"),
39        Err(e) => println!("✗ Arc-cosine Sampler failed: {}", e),
40    }
41
42    // Test Nyström Method
43    println!("Testing Nyström Method...");
44    match test_nystroem(&x) {
45        Ok(_) => println!("✓ Nyström Method works"),
46        Err(e) => println!("✗ Nyström Method failed: {}", e),
47    }
48
49    println!("Testing complete!");
50}
51
52fn test_rbf_sampler(
53    x: &scirs2_core::ndarray::Array2<f64>,
54) -> Result<(), Box<dyn std::error::Error>> {
55    let rbf = RBFSampler::new(10).gamma(0.1);
56    let fitted = rbf.fit(x, &())?;
57    let result = fitted.transform(x)?;
58    assert_eq!(result.shape(), &[4, 10]);
59    Ok(())
60}
61
62fn test_laplacian_sampler(
63    x: &scirs2_core::ndarray::Array2<f64>,
64) -> Result<(), Box<dyn std::error::Error>> {
65    let laplacian = LaplacianSampler::new(10).gamma(0.1);
66    let fitted = laplacian.fit(x, &())?;
67    let result = fitted.transform(x)?;
68    assert_eq!(result.shape(), &[4, 10]);
69    Ok(())
70}
71
72fn test_polynomial_sampler(
73    x: &scirs2_core::ndarray::Array2<f64>,
74) -> Result<(), Box<dyn std::error::Error>> {
75    let poly = PolynomialSampler::new(10).degree(2).gamma(1.0).coef0(1.0);
76    let fitted = poly.fit(x, &())?;
77    let result = fitted.transform(x)?;
78    assert_eq!(result.shape(), &[4, 10]);
79    Ok(())
80}
81
82fn test_arc_cosine_sampler(
83    x: &scirs2_core::ndarray::Array2<f64>,
84) -> Result<(), Box<dyn std::error::Error>> {
85    let arc_cosine = ArcCosineSampler::new(10).degree(1);
86    let fitted = arc_cosine.fit(x, &())?;
87    let result = fitted.transform(x)?;
88    assert_eq!(result.shape(), &[4, 10]);
89    Ok(())
90}
91
92fn test_nystroem(x: &scirs2_core::ndarray::Array2<f64>) -> Result<(), Box<dyn std::error::Error>> {
93    // Test with Linear kernel and Random sampling
94    let nystroem = Nystroem::new(Kernel::Linear, 3).sampling_strategy(SamplingStrategy::Random);
95    let fitted = nystroem.fit(x, &())?;
96    let result = fitted.transform(x)?;
97    assert_eq!(result.nrows(), 4);
98
99    // Test with RBF kernel and Leverage Score sampling
100    let nystroem_rbf = Nystroem::new(Kernel::Rbf { gamma: 0.1 }, 3)
101        .sampling_strategy(SamplingStrategy::LeverageScore);
102    let fitted_rbf = nystroem_rbf.fit(x, &())?;
103    let result_rbf = fitted_rbf.transform(x)?;
104    assert_eq!(result_rbf.nrows(), 4);
105
106    Ok(())
107}
sklears_kernel_approximation/simple_test.rs

sklears_kernel_approximation/
simple_test.rs
