Function thin_random 

pub fn thin_random(
    points: &[ThinPoint3],
    target_count: usize,
    seed: u64,
) -> Vec<ThinPoint3>

Random thinning: keep target_count points selected uniformly without replacement via a deterministic Fisher-Yates shuffle.

The shuffle is driven by an internal LCG (lcg_next) rather than the rand crate, in keeping with the project’s no rand dependency policy. Identical seed values produce identical outputs across runs and across platforms (the LCG state is a u64 and only wrapping_* arithmetic is used). Output points are returned in input order — the shuffle is applied to an index vector, the first target_count indices are retained, and the final point vector is materialised in the original order.

Edge cases:

• Empty input returns an empty Vec.
• target_count >= points.len() returns a copy of the input (no thinning).
• target_count == 0 returns an empty Vec.

Examples

use oxigdal_algorithms::raster::{ThinPoint3, thin_random};

let pts: Vec<_> = (0..100)
    .map(|i| ThinPoint3::new(i as f64, 0.0, 0.0))
    .collect();
let a = thin_random(&pts, 10, 0xC0FFEE);
let b = thin_random(&pts, 10, 0xC0FFEE);
assert_eq!(a, b); // deterministic for a given seed
assert_eq!(a.len(), 10);