Struct CalibrationConfig 

pub struct CalibrationConfig {

    pub flight_path_m: f64,
    pub fit_background: bool,
    pub restarts: usize,
    pub xatol: f64,
    pub fatol: f64,
    pub max_iter: usize,
    pub ic_n_energies: usize,
    pub ic_n_tau: usize,
    pub psr_fwhm_ns: f64,
    pub fit_t0: bool,
    pub fit_l_scale: bool,
    pub position_t0_center_us: f64,
    pub position_l_scale_center: f64,
    pub position_t0_prior_us: Option<f64>,
    pub position_l_scale_prior: Option<f64>,
}

Configuration for calibrate_resolution.

Fields

flight_path_m: f64

Flight-path length (m).

fit_background: bool

Fit a low-order baseline (anorm + constant + linear) instead of anorm only.

restarts: usize

Number of optimizer restarts (perturbed starts; keep the best).

xatol: f64

Nelder–Mead simplex-spread tolerance.

fatol: f64

Nelder–Mead objective-range tolerance.

max_iter: usize

Nelder–Mead maximum iterations.

ic_n_energies: usize

IC synthesis grid resolution (energies × τ-samples per kernel).

ic_n_tau: usize

psr_fwhm_ns: f64

SNS PSR (accumulator-ring) channel-triangle FWHM in ns, folded into the IC family only (default DEFAULT_PSR_FWHM_NS; 0.0 disables the fold). Tabulated/UDR (FTS) kernels already carry the fold in the file itself (header: “folded triang FWHM 350 ns PSR”) and are structurally never re-folded here — applying it twice would double-count the burst. When the family is IkedaCarpenter { fit_psr: true } this value is the fit’s starting point instead of a pin, clamped into the 0.05–1 µs fit box: a width in (1, 10] µs is a legal pin but an out-of-box start — the fit then starts at the box top (announced by a stderr warning), and if it stays there it reports psr_fwhm_us:upper in CalibrationResult::bounds_hit. Nonzero widths above PSR_FWHM_PIN_CEILING_US (10 µs = 10 000 ns) are rejected as a ns↔µs unit slip — see that constant for the quadratic-cost rationale.

fit_t0: bool

Fit the SAMMY TOF-zero t0 (µs) as a SHARED energy-scale parameter. Default false — position is pinned at position_t0_center_us so calibration is a pure shape/width fit (matching SAMMY, where t0/L are a separate energy-scale calibration). Opt in only with a metrology prior; see with_position_prior.

fit_l_scale: bool

Fit the flight-path scale L_scale as a shared energy-scale parameter. Default false. A free L_scale shares the asymmetric-kernel lag’s 1/√E basis and corrupts the calibrated width — fit it only under a prior.

position_t0_center_us: f64

Prior mean (and pinned value when fit_t0 is false) of the TOF zero t0 (µs). Default 0.0. Lets a caller inject a pre-calibrated t0.

position_l_scale_center: f64

Prior mean (and pinned value when fit_l_scale is false) of L_scale. Default 1.0.

position_t0_prior_us: Option<f64>

Gaussian prior σ on t0 (µs); None = flat (bounded only). When set, adds ((t0 − center)/σ)² to the data χ² (a metrology penalty, not part of the reported chi2_dof).

position_l_scale_prior: Option<f64>

Gaussian prior σ on L_scale; None = flat. See position_t0_prior_us.

Implementations

impl CalibrationConfig

pub fn with_position_prior( self, t0_center_us: f64, l_scale_center: f64, sigma_t0_us: f64, sigma_l_scale: f64, ) -> Self

Enable a SHARED, metrology-priored energy-scale (t0, L_scale) fit: sets fit_t0/fit_l_scale, the prior means (*_center), and the Gaussian prior σ. Use this for joint energy-scale or cross-family identifiability work; the default config pins position (pure shape/width calibration). Pass the prior σ from the instrument’s independent flight-path / timing metrology — a loose σ marginalizes position (weak, honest shape-only discrimination), a tight σ pins it.

Trait Implementations

impl Clone for CalibrationConfig

fn clone(&self) -> CalibrationConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CalibrationConfig

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

Formats the value using the given formatter.

impl Default for CalibrationConfig

fn default() -> Self

Returns the “default value” for a type.