splimes 0.1.0

Spline interpolation over irregularly sampled time series, on SIMD/parallel CPU or GPU (wgpu), with BigDecimal values.
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

use crate::types::Spline;

/// Apply fast path optimization to spline types based on dataset characteristics
///
/// This function automatically downgrades complex spline types to simpler ones
/// for better performance when dealing with large datasets or specific conditions.
///
/// # Arguments
/// * `spline_type` - The requested spline type
/// * `measurement_count` - Number of input measurements
///
/// # Returns
/// Optimized spline type that provides better performance characteristics
#[must_use]
pub const fn apply_fast_path(spline: Spline, measurement_count: usize) -> Spline {
	match spline {
		// Linear always stays linear - it's already optimal
		Spline::Linear => Spline::Linear,

		// Quadratic optimizations
		Spline::Quadratic => {
			if measurement_count > 5000 {
				// Very large datasets: degrade to linear for speed
				Spline::Linear
			} else {
				Spline::Quadratic
			}
		}

		// Cubic optimizations
		Spline::Cubic => {
			if measurement_count > 5000 {
				// Very large datasets: degrade to linear
				Spline::Linear
			} else if measurement_count >= 2500 {
				// Large datasets: degrade to quadratic
				Spline::Quadratic
			} else {
				Spline::Cubic
			}
		}

		// Polynomial optimizations
		Spline::Polynomial(degree, bounds_factor) => {
			match degree {
				1 => Spline::Linear,                            // Degree 1 is linear
				2 => Spline::Quadratic,                         // Degree 2 is quadratic
				3 => Spline::Cubic,                             // Degree 3 is cubic
				_ => Spline::Polynomial(degree, bounds_factor), // Small datasets: keep polynomial
			}
		}
	}
}