Enum Prior 

pub enum Prior<const D: usize> {
    Uniform {
        bounds: [(f64, f64); D],
    },
    Gaussian {
        mean: [f64; D],
        cov_diag: [f64; D],
        truncation_sigmas: f64,
    },
    Custom {
        bounds: [(f64, f64); D],
        density: Arc<dyn Fn([f64; D]) -> f64 + Send + Sync>,
    },
}

Probability density / sampling region over R^D.

All variants are validated by Prior::validate. The sample-space described here is the support of the prior; quadrature points are drawn inside this support and transformed if needed.

Variants

Uniform

Uniform density on the axis-aligned box bounds[i] = (lo, hi).

Fields

bounds: [(f64, f64); D]

Per-axis lower / upper bounds. Lower must be strictly less than upper.

Gaussian

Axis-aligned Gaussian. cov_diag[i] holds σ² along axis i. Truncated to ±k·σ in Quadrature::Sobol / Quadrature::LatinHypercube via inverse-CDF; the truncation_sigmas field controls k (default 4.0).

Fields

mean: [f64; D]

Per-axis means.

cov_diag: [f64; D]

Per-axis variances (must be > 0).

truncation_sigmas: f64

Truncation in standard deviations (samples are clamped after inverse-CDF). Default 4.0 captures > 99.99 % of the mass and avoids runaway tails.

Custom

Arbitrary density specified pointwise. The closure must return a non-negative density. Quadrature for Custom priors is restricted to Sobol / LatinHypercube over an axis-aligned bounding box that the caller supplies (the closure is not used by quadrature; only by importance-weighting inside evaluate_area_loss).

Fields

bounds: [(f64, f64); D]

Bounding box used by sampling-based quadratures.

density: Arc<dyn Fn([f64; D]) -> f64 + Send + Sync>

Density evaluated at each sampled point. Must be non-negative.

Implementations

impl<const D: usize> Prior<D>

pub fn validate(&self) -> Result<(), AreaError>

Reject malformed priors before any quadrature samples are drawn.

pub fn bounding_box(&self) -> [(f64, f64); D]

Axis-aligned bounding box used as the search space for [Quadrature::WorstCaseSearch] and the integration domain for tensor-product / sampled quadratures.

Trait Implementations

impl<const D: usize> Clone for Prior<D>

fn clone(&self) -> Prior<D>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<const D: usize> Debug for Prior<D>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.