Struct PiecewiseView 

pub struct PiecewiseView<'a, R: Scalar> {
    data: &'a PiecewiseDiscount,
    params: Cow<'a, [R]>,
    coefficients: Option<Vec<R>>,
    long_coefficients: Option<Vec<R>>,
    long_ys: Option<Vec<R>>,
}

PiecewiseDiscount lifted into one pricing mode.

Borrows the storage; owns only this mode’s parameters (and for the passive mode, borrows those too).

Fields

data: &'a PiecewiseDiscount

params: Cow<'a, [R]>

coefficients: Option<Vec<R>>

A global scheme’s coefficients, solved once here and shared by every query against this view. None under a local scheme, which reads only the pillars bracketing a query and so has nothing to precompute.

On a mixed curve this belongs to the short side and is solved over the anchor-to-changeover knots alone.

long_coefficients: Option<Vec<R>>

The long side’s coefficients on a mixed curve, solved over the changeover-to-last knots alone. Neither side’s solve reads the other’s ordinates: that independence is what keeps the two halves from having to agree on a derivative at the join.

long_ys: Option<Vec<R>>

The long side’s ordinates in its shifted frame, built once per lift for the same reason its taus live on the curve: a query allocates nothing.

Implementations

impl<'a, R: Scalar> PiecewiseView<'a, R>

fn build(data: &'a PiecewiseDiscount, params: Cow<'a, [R]>) -> Self

The one place a view is built: solves a global scheme’s coefficients and leaves a local scheme’s None.

Every constructor routes through here so the solve cannot be forgotten on one path and silently produce a curve that reads as LogLinear.

pub fn parameters(&self) -> &[R]

This mode’s parameters. For the passive mode these are the stored ones, borrowed.

pub fn with_parameters( data: &'a PiecewiseDiscount, params: &[R], ) -> Result<Self, CurveError>

A view over the same storage with params substituted — how the risk lift installs tape-registered handles without touching the storage.

Errors

CurveError::ParameterCountMismatch if params is not one parameter per free pillar. Checked here so a wrong-length substitution fails at the seam instead of indexing out of bounds mid-valuation.

fn log_df(&self, tau: f64) -> R

The log discount factor at tau, under this curve’s scheme. The anchor (0, 0) is the implicit left end.

Trait Implementations

impl<'a, R: Clone + Scalar> Clone for PiecewiseView<'a, R>

fn clone(&self) -> PiecewiseView<'a, R>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<'a, R: Debug + Scalar> Debug for PiecewiseView<'a, R>

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

Formats the value using the given formatter.

impl<R: Scalar> YieldCurve<R> for PiecewiseView<'_, R>

fn reference_date(&self) -> NaiveDate

The date at which the discount factor is 1.

fn discount(&self, date: NaiveDate) -> Result<R, CurveError>

The discount factor to date.

fn log_discount(&self, date: NaiveDate) -> Result<R, CurveError>

The log discount factor to date.

fn discount_at_ordinal(&self, ordinal: i32) -> Result<R, CurveError>

The discount factor to the proleptic-Gregorian day number ordinal.

fn log_discount_at_ordinal(&self, ordinal: i32) -> Result<R, CurveError>

The log discount factor at ordinal; see discount_at_ordinal.

fn rate_day_count(&self) -> DayCountFraction

The day count this curve forms a rate under.

fn zero_rate(&self, date: NaiveDate) -> Result<R, CurveError>

The continuously-compounded zero rate to date.

fn forward_rate( &self, start: NaiveDate, end: NaiveDate, ) -> Result<R, CurveError>

The continuously-compounded forward rate between start and end.