Function calibrate 

pub fn calibrate(
    reference: NaiveDate,
    rungs: &[Rung],
    fixings: &FixingStore,
    fx: &FxRates,
    tolerance: Tolerance,
    scheme: Interpolation,
) -> Result<Calibrated, CalibrationError>

Calibrates a discount curve so that every rung prices to zero.

Pillars solve in order, each against the instrument quoted for it. A piecewise log-linear curve extends its last segment’s forward beyond its last pillar, so a rung maturing at pillar j depends only on pillars up to j — which is what makes the sequential solve valid, and what lets each step carry a curve holding only the pillars solved so far.

Each rung’s instrument both discounts and projects on the curve under solve — the self-discounting form. For a projection curve solved against a discount curve already known, see calibrate_with.

Each Newton step runs in forward mode, seeded at the pillar being solved: one input, one output, value and derivative from a single pass. No tape is opened and no reverse sweep runs.

Errors

CalibrationError if the ladder is malformed, a pillar does not converge, a pillar has no effect on its own instrument, or a valuation fails.