reggae¶
Submodules¶
Classes¶
Generalisation of pbjam.star to assist in determination of dipole mixed-mode |
Package Contents¶
- class reggae.DipoleStar(s=None, f=None)¶
Generalisation of pbjam.star to assist in determination of dipole mixed-mode parameters.
- labels = ['$\\Delta\\Pi_0$ (relative)', '$L_0$', '$D_0$', '$\\epsilon_g$', '$\\log...¶
- _extra_ndims = 1¶
- _psd_model¶
- s = None¶
- f = None¶
- classmethod from_pbjam(star)¶
Import and prepare results from PBjam
Takes instances of the either the pbjam.star or pbjam.modeID.modeIDsampler classes and constructs a residual spectrum, by dividing out the l=2,0 (including the background) or just the background respectively.
Parameters¶
- klass: self
reggae.DipoleStar class self.
- star: object
Instance of pbjam.star or pbjam.modeID.modeIDsampler classes.
- __call__(dynamic=False, **kwargs)¶
- summarise(sampler=None)¶
- ptform(u)¶
Turns coordinates on the unit cube into θ_reg object
Parameters¶
- u: list
Point location in the unit hypercube.
Returns¶
- x: jnp.array
Point location in model parameter space.
- inv_ptform(θ_reg, norm)¶
Turns θ_reg object into coordinates on the unit cube
Parameters¶
- θ_reg: ThetaReg object or array-like
Point location in model parameters space.
- norm: jnp.array
Normalisation
- Returns: jnp.array
Point location in the unit hypercube.
- model(theta)¶
Construct the model for the likelihood evaluation
Parameters¶
- theta: jnp.array
Array of model parameters.
Returns¶
- mod: jnp.array
Spectrum model.
- ln_like(theta)¶
Evaluate the model log-likelihood at theta
The log-likelihood is the joint probability of observing the spectrum given the model.
Parameters¶
- theta: jnp.array
Array of model parameters
Returns¶
- lnlike: float
log-likelihood at theta.
- select_n_g()¶
Select the relevant radial g-mode orders
Based on prior estimates of the relevant range period spacing and eps_g, computes g-modes that significantly overlap with the envelope.
These will be used to compute the model.
Returns¶
- n_g: jnp.array
List of radial orders for the g-modes.
- static _prepare_theta_asy(star)¶
Get parameters for asymptotic l=2,0+background model.
Parameters¶
- star: object
Either a pbjam.star or pbjam.modeID.modeIDsampler class instance.
Returns¶
- nu_0: jnp.array
l=0 mode frequencies.
- nu_2: jnp.array
l=2 mode frequencies.
- ThetaAsy: object
ThetaAsy class instance, containing l=2,0 and background model parameters.
- static make_pbjam_model(star, n_samples=50)¶
Construct a spectrum model from PBjam output
Takes either a pbjam.star or pbjam.modeID.modeIDsampler object from a previous PBjam run. This is used to construct a residual spectrum.
In the case of a pbjam.star object being passed, a set of models are drawn and averaged.
Parameters¶
- star: object
Either a pbjam.star or pbjam.modeID.modeIDsampler object.
- n_samples: int, optional
Number of samples to use to construct the mean model. Default is 50.
Returns¶
- mod: jnp.array
Spectrum model.
- simplex(theta_reg, norm, **kwargs)¶
Find maximum log-likelihood using Nelder-Mead downhill simplex minimization.
Parameters¶
- theta_reg: dataclass
A ThetaReg dataclass instance
- norm: jnp.array
An initial point in model parameter space.
- genetic_algorithm(solve_kwargs=None, **kwargs)¶
Find the maximum likelihood using a genetic algorithm.
Any keyword arguments are passed to the yabox.DE initializatoin.
- solve_kwarg: dict, optional
Dictionary of argumentts to pass to the yabox.DE.solve method.