Rust Quickstart: Add rscopulas and Fit a Copula

Rscopulas ships a native Rust library crate (rscopulas-core) that provides the same fitting, scoring, and sampling capabilities as the Python API. This guide shows you how to add the crate as a dependency, fit a model, read diagnostics, sample from it, and run the runnable examples included in the repository.

Add the dependency

While developing inside the rscopulas workspace, add the crate as a path dependency alongside ndarray and rand:

[dependencies]
ndarray = "0.17"
rand = "0.9"
rscopulas = { path = "crates/rscopulas-core" }

ndarray provides the array types used for pseudo-observations and samples. rand supplies the RNG required by sample.

Fit a Gaussian copula

Wrap your data in PseudoObs to validate that every entry is strictly inside (0, 1), then call GaussianCopula::fit.

use ndarray::array;
use rand::{SeedableRng, rngs::StdRng};
use rscopulas::{CopulaModel, FitOptions, GaussianCopula, PseudoObs};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let data = PseudoObs::new(array![
        [0.12_f64, 0.18],
        [0.21, 0.25],
        [0.82, 0.79],
    ])?;
    let fit = GaussianCopula::fit(&data, &FitOptions::default())?;
    println!("AIC: {}", fit.diagnostics.aic);
    let mut rng = StdRng::seed_from_u64(7);
    println!("sample:\n{:?}", fit.model.sample(4, &mut rng, &Default::default())?);
    Ok(())
}

fit.diagnostics exposes loglik, aic, bic, converged, and n_iter — the same fields available in the Python API.

Runnable examples

The repository ships several ready-to-run examples under crates/rscopulas-core/examples/. Run them from the workspace root with cargo run:

Each example is self-contained and uses fixtures or hardcoded pseudo-observations, so you can run it without any additional setup.

Next steps