pub fn empirical_wasserstein_1d(a: &[f64], b: &[f64]) -> Result<f64, String> {
if a.is_empty() || b.is_empty() {
return Err("samples must be non-empty".to_string());
}
let mut sa = a.to_vec();
let mut sb = b.to_vec();
sa.sort_unstable_by(|x, y| x.total_cmp(y));
sb.sort_unstable_by(|x, y| x.total_cmp(y));
if sa.len() == sb.len() {
let n = sa.len() as f64;
let sum: f64 = sa.iter().zip(sb.iter()).map(|(x, y)| (x - y).abs()).sum();
return Ok(sum / n);
}
let na = sa.len() as f64;
let nb = sb.len() as f64;
let mut ia = 0usize;
let mut ib = 0usize;
let mut cdf_a: f64 = 0.0;
let mut cdf_b: f64 = 0.0;
let mut prev = sa[0].min(sb[0]);
let mut distance = 0.0;
while ia < sa.len() || ib < sb.len() {
let next_a = if ia < sa.len() { sa[ia] } else { f64::INFINITY };
let next_b = if ib < sb.len() { sb[ib] } else { f64::INFINITY };
let x = next_a.min(next_b);
distance += (cdf_a - cdf_b).abs() * (x - prev);
while ia < sa.len() && sa[ia] == x {
cdf_a += 1.0 / na;
ia += 1;
}
while ib < sb.len() && sb[ib] == x {
cdf_b += 1.0 / nb;
ib += 1;
}
prev = x;
}
Ok(distance)
}