Struct SavedTransformationNormalGeometry 

pub struct SavedTransformationNormalGeometry {
    pub parameterization: TransformationNormalParameterization,
    pub response_degree: usize,
    pub response_knot_count: usize,
    pub shape_coordinate_count: usize,
    pub cone_carrier_covariate_width: usize,
    pub cone_carrier_row_count: usize,
    pub certified_response_support: (f64, f64),
    pub response_median: f64,
}

Direct-α SCOP-CTN transformation geometry required to replay a saved transformation-normal fit (gam#2306).

Beyond the response knots/degree/median/transform snapshot, direct-α replay — and in particular the certified-domain refusal that crate::transformation_normal::transformation_normal_pit_score raises for out-of-support prediction — needs the coefficient parameterization marker, the response value-basis spec, the shape-coordinate count, the cone-carrier (Khatri-Rao factor) shape, and the response support the positivity cone was certified over. This is a REQUIRED v13 field for a transformation-normal model: loading a CTN payload that lacks it (a v12-or-older squared-γ model) is a typed rejection in [FittedModelPayload::validate_for_persistence], never a heuristic conversion.

Fields

parameterization: TransformationNormalParameterization

Coefficient chart. Only DirectAlpha is written by a v13+ CTN writer.

response_degree: usize

I-spline response value-basis degree (response_degree).

response_knot_count: usize

Number of response knots persisted in transformation_response_knots (a provenance cross-check: replay rebuilds the value basis from those knots and this count must agree).

shape_coordinate_count: usize

Number of shape coordinates p_resp − 1: the non-negativity-constrained I-spline columns. Response column 0 is the unconstrained location field.

cone_carrier_covariate_width: usize

Covariate-side design width p_cov. The Khatri-Rao cone carrier factor is n × p_cov, so the coefficient block is p_resp × p_cov.

cone_carrier_row_count: usize

Number of covariate rows the positivity cone was certified over (the fitted observation count n).

certified_response_support: (f64, f64)

Response support [y_lo, y_hi] (the clamped-knot span) the transform and positivity cone were certified over — the certified domain the out-of-support prediction refusal is defined against.

response_median: f64

Response median anchoring the per-row monotonicity floor ε·(y − median).

Implementations

impl SavedTransformationNormalGeometry

pub fn validate(&self, context: &str) -> Result<(), FittedModelError>

Self-contained structural validation. Cross-checks against the sibling response-basis payload fields (knot count, degree) are done by the caller in validate_for_persistence.

Trait Implementations

impl Clone for SavedTransformationNormalGeometry

fn clone(&self) -> SavedTransformationNormalGeometry

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for SavedTransformationNormalGeometry

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for SavedTransformationNormalGeometry

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for SavedTransformationNormalGeometry

fn eq(&self, other: &SavedTransformationNormalGeometry) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for SavedTransformationNormalGeometry

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for SavedTransformationNormalGeometry