Template Class Interpolator1D

Class Documentation

template<size_t NF = 1>
class Interpolator1D

A class to handle 1D interpolation.

Template Parameters:

NF – Number of quantities to interpolate

Public Types

using PostProcess = void (*)(std::array<double, NF>&)

Type of a function that post-processes the interpolated values at a point.

Takes the full array of NF interpolated values at a single point and modifies it in place – see setPostProcess().

Public Functions

inline Interpolator1D(const std::vector<double> &x, const std::array<std::vector<double>, NF> &f, const gsl_interp_type *interpType = gsl_interp_steffen)

Construct an Interpolator1D.

Parameters:

  • x – Locations of sample points to be interpolated

  • f – Values of sample points to be interpolated

  • interpType – The type of interpolation to use

inline Interpolator1D(const std::vector<double> &x, const std::vector<double> &f, const gsl_interp_type *interpType = gsl_interp_steffen)
inline virtual ~Interpolator1D()
Interpolator1D(const Interpolator1D&) = delete
auto operator=(const Interpolator1D&) -> Interpolator1D& = delete
Interpolator1D(Interpolator1D&&) = delete
auto operator=(Interpolator1D&&) -> Interpolator1D& = delete
inline auto xMin() const

Get minimum allowed x.

Returns:

Minimum allowed x

inline auto xMax() const

Get maximum allowed x.

Returns:

Maximum allowed x

inline auto xRange() const

Get allowed range in x.

Returns:

Allowed range in x

inline void setPostProcess(const PostProcess postProcess)

Set a function to apply to the interpolated values at every point.

When set, every point evaluation – operator()(x), and operator()(x, idx), which then evaluates every quantity so that postProcess sees all of them together – passes its full array of interpolated values through postProcess before returning. This exists for constraints that tie quantities together, which interpolating each quantity independently cannot respect on its own: e.g. stellar surface mass fractions, whose sum independent (even monotonicity- preserving) interpolation can push above 1 between sample points (see tracks::renormalizeSurfaceAbundances()). The integ() methods are unaffected, since a constraint on point values does not in general translate into one on integrals.

Parameters:

postProcess – The function to apply, or nullptr (the default) for none

inline auto operator()(double x) const

Interpolate to a given point.

Return type is a double if NF = 1, and a std::array otherwise.

Parameters:

x – The point to which to interpolate

Returns:

The interpolated value(s), passed through the function set by setPostProcess(), if any

inline auto operator()(double x, size_t idx) const

Interpolate a single quantity to a given point.

Parameters:

  • x – The point to which to interpolate

  • idx – Quantity to interpolate; must be <= NF

Returns:

The interpolated value, consistent with operator()(x)[idx] – including any post-processing set by setPostProcess()

inline auto integ(double x0, double x1) const

Integrate the interpolant over [x0, x1].

Return type is a double if NF = 1, and a std::array otherwise – mirrors operator()(double) exactly, except that each quantity’s interpolant is integrated over [x0, x1] (via gsl_interp_eval_integ) rather than evaluated at a single point.

Parameters:

  • x0 – Lower limit of integration

  • x1 – Upper limit of integration

Returns:

The integral value(s)

inline auto integ(double x0, double x1, size_t idx) const

Integrate a single quantity’s interpolant over [x0, x1].

Parameters:

  • x0 – Lower limit of integration

  • x1 – Upper limit of integration

  • idx – Quantity to integrate; must be <= NF

Returns:

The integral value

inline auto integ(const Interpolator1D &g, double x0, double x1, size_t idx) const -> double

Integrate the product f(x)*g(x) over [x0, x1] for a single field.

Evaluates \(\int_{x_0}^{x_1} f_\mathrm{idx}(x)\, g_\mathrm{idx}(x)\, dx\) where \(f(x)\) is this Interpolator1D and \(g(x)\) is a second Interpolator1D. Both Interpolator1D objects are required to have the same NF – this is guaranteed at compile time by the template signature. The integration range is the intersection of [x0, x1] with both interpolators’ domains; returns 0 if that intersection is empty.

Parameters:

  • g – A second Interpolator1D; must have the same NF as this one

  • x0 – Lower limit of integration

  • x1 – Upper limit of integration

  • idx – Field index to integrate; must be < NF

Returns:

The integral value as a double

inline auto integ(const Interpolator1D &g, double x0, double x1) const

Integrate the product f(x)*g(x) over [x0, x1] for all fields.

Evaluates \(\int_{x_0}^{x_1} f(x)\, g(x)\, dx\) where \(f(x)\) is this Interpolator1D and \(g(x)\) is a second Interpolator1D. Both Interpolator1D objects are required to have the same NF – this is guaranteed at compile time by the template signature. The integration range is the intersection of [x0, x1] with both interpolators’ domains; returns 0 (or an array of zeros) if that intersection is empty. Return type mirrors integ(double, double): a double if NF = 1, a std::array<double, NF> otherwise.

Parameters:

  • g – A second Interpolator1D; must have the same NF as this one

  • x0 – Lower limit of integration

  • x1 – Upper limit of integration

Returns:

The integral value(s)