Usage

use nuts_rs::{CpuLogpFunc,CpuMath,LogpError,DiagGradNutsSettings,Chain,SampleStats,Settings};use thiserror::Error;use rand::thread_rng;// Define a function that computes the unnormalized posterior density// and its gradient.#[derive(Debug)]structPosteriorDensity{}// The density might fail in a recoverable or non-recoverable manner...#[derive(Debug,Error)]enumPosteriorLogpError{}implLogpErrorforPosteriorLogpError{fnis_recoverable(&self) ->bool{false}}implCpuLogpFuncforPosteriorDensity{typeLogpError =PosteriorLogpError;// Only used for transforming adaptation.typeTransformParams =();// We define a 10 dimensional normal distributionfndim(&self) ->usize{10}// The normal likelihood with mean 3 and its gradient.fnlogp(&mutself,position:&[f64],grad:&mut[f64]) ->Result<f64,Self::LogpError>{let mu =3f64;let logp = position.iter().copied().zip(grad.iter_mut()).map(|(x, grad)|{let diff = x - mu;*grad = -diff;                -diff* diff /2f64}).sum();returnOk(logp)}}fnmain(){// We get the default sampler argumentsletmut settings =DiagGradNutsSettings::default();// and modify as we like    settings.num_tune =1000;    settings.maxdepth =3;// small value just for testing...// We instanciate our posterior density functionlet logp_func =PosteriorDensity{};let math =CpuMath::new(logp_func);let chain =0;letmut rng =thread_rng();letmut sampler = settings.new_chain(0, math,&mut rng);// Set to some initial position and start drawing samples.    sampler.set_position(&vec![0f64;10]).expect("Unrecoverable error during init");letmut trace =vec![];// Collection of all drawsfor _in0..2000{let(draw, info) = sampler.draw().expect("Unrecoverable error during sampling");        trace.push(draw.clone());println!("Draw: {:?}", draw);}}

Implementation details