bayes_yplus

Submodules

bayes_yplus.utils

utils.py Model Utilities

bayes_yplus.utils.gaussian(x: float, center: float, fwhm: float) → float

Evaluate a normalized Gaussian function

Parameters:

xfloat: Position at which to evaluate
centerfloat: Gaussian centroid
fwhmfloat: Gaussian FWHM line width

Returns:

float: Gaussian evaluated at x

bayes_yplus.yplus_model

yplus_model.py YPlusModel definition

class bayes_yplus.yplus_model.YPlusModel(*args, **kwargs)

Bases: BaseModel

Definition of the YPlusModel

Attributes:

baseline_deterministics: Get the deterministic baseline parameter names.
baseline_freeRVs: Get the free baseline parameter names.
cloud_deterministics: Get the deterministic cloud parameter names.
cloud_freeRVs: Get the free cloud parameter names.
hyper_deterministics: Get the deterministic hyper parameter names.
hyper_freeRVs: Get the free hyper parameter names.
labeller: Get the arviz labeller.
unique_solution: Check if posterior samples suggest a unique solution.

Methods

`add_baseline_priors`([prior_baseline_coeffs])	Add baseline priors to the model.
`add_likelihood`()	Add likelihood to the model.
`add_priors`([prior_H_area, prior_H_center, ...])	Add priors and deterministics to the model
`bic`([chain, solution])	Calculate the Bayesian information criterion at the mean point estimate.
`fit`([n, draws, rel_tolerance, ...])	Approximate posterior distribution using Variational Inference (VI).
`graph`()	Generate visualization of the model graph.
`mean_lnlike`([chain, solution])	Evaluate mean log-likelihood over posterior samples.
`null_bic`()	Evaluate the Bayesian Information Criterion for the null hypothesis (baseline only, no clouds)
`predict_baseline`([baseline_params])	Predict the un-normalized baseline model.
`reset_results`()	Reset results and convergence checks.
`sample`([init, n_init, chains, init_kwargs, ...])	Sample posterior distribution using MCMC.
`sample_posterior_predictive`([solution, thin])	Generate posterior predictive samples
`sample_prior_predictive`([samples])	Generate prior predictive samples
`sample_smc`(**kwargs)	Sample posterior distribution using Sequential Monte Carlo.
`solve`([num_gmm_samples, kl_div_threshold])	Identify unique solutions and break the labeling degeneracy.

add_likelihood(): Add likelihood to the model.

add_priors(prior_H_area: float = 1000.0, prior_H_center: Iterable[float] = [0.0, 25.0], prior_H_fwhm: Iterable[float] = [25.0, 10.0], prior_He_H_fwhm_ratio: Iterable[float] = [1.0, 0.1], prior_yplus: float = 0.05, prior_rms: dict[str, float] | None = None, prior_baseline_coeffs: dict[str, Iterable[float]] | None = None, ordered: bool = False)

Add priors and deterministics to the model

Parameters:

prior_H_areafloat, optional: Prior distribution on H RRL line area (mK km s-1), by default 1000.0, where H_area ~ HalfNormal(sigma=prior)
prior_H_centerIterable[float], optional: Prior distribution on H RRL center velocity (km s-1), by default [0, 25.0], where H_center ~ Normal(mu=prior[0], sigma=prior[1])
prior_H_fwhmIterable[float], optional: Prior distribution on H RRL FWHM line width (km s-1), by default [25.0, 10.0], where H_fwhm ~ Gamma(mu=prior[0], sigma=prior[1])
prior_He_H_fwhm_ratioIterable[float], optional: Prior distribution on He/H RRL FWHM line width ratio, by default [1.0, 0.1], where He_H_fwhm_ratio ~ Gamma(mu=prior[0], sigma=prior[1])
prior_yplusfloat, optional: Prior distribution on y+, by default 0.05, where yplus ~ HalfNormal(sigma=prior_yplus)
prior_rmsOptional[dict[str, float]], optional: Prior distribution on spectral rms (K), by default None, where rms ~ HalfNormal(sigma=prior) Keys are dataset names and values are priors. If None, then the spectral rms is taken from dataset.noise and not inferred.
prior_baseline_coeffsOptional[dict[str, Iterable[float]]], optional: Prior distribution on the normalized baseline polynomial coefficients, by default None. Keys are dataset names and values are lists of length baseline_degree+1. If None, use [1.0]*(baseline_degree+1) for each dataset.
orderedbool, optional: If True, assume ordered velocities (optically thin assumption), by default False. If True, the prior distribution on the velocity becomes H_center(cloud = n) ~

prior[0] + sum_i(H_center[i < n]) + Gamma(alpha=2.0, beta=1.0/prior[1])

Module contents

class bayes_yplus.YPlusModel(*args, **kwargs)

Bases: BaseModel

Definition of the YPlusModel

Attributes:

baseline_deterministics: Get the deterministic baseline parameter names.
baseline_freeRVs: Get the free baseline parameter names.
cloud_deterministics: Get the deterministic cloud parameter names.
cloud_freeRVs: Get the free cloud parameter names.
hyper_deterministics: Get the deterministic hyper parameter names.
hyper_freeRVs: Get the free hyper parameter names.
labeller: Get the arviz labeller.
unique_solution: Check if posterior samples suggest a unique solution.

Methods

`add_baseline_priors`([prior_baseline_coeffs])	Add baseline priors to the model.
`add_likelihood`()	Add likelihood to the model.
`add_priors`([prior_H_area, prior_H_center, ...])	Add priors and deterministics to the model
`bic`([chain, solution])	Calculate the Bayesian information criterion at the mean point estimate.
`fit`([n, draws, rel_tolerance, ...])	Approximate posterior distribution using Variational Inference (VI).
`graph`()	Generate visualization of the model graph.
`mean_lnlike`([chain, solution])	Evaluate mean log-likelihood over posterior samples.
`null_bic`()	Evaluate the Bayesian Information Criterion for the null hypothesis (baseline only, no clouds)
`predict_baseline`([baseline_params])	Predict the un-normalized baseline model.
`reset_results`()	Reset results and convergence checks.
`sample`([init, n_init, chains, init_kwargs, ...])	Sample posterior distribution using MCMC.
`sample_posterior_predictive`([solution, thin])	Generate posterior predictive samples
`sample_prior_predictive`([samples])	Generate prior predictive samples
`sample_smc`(**kwargs)	Sample posterior distribution using Sequential Monte Carlo.
`solve`([num_gmm_samples, kl_div_threshold])	Identify unique solutions and break the labeling degeneracy.

add_likelihood(): Add likelihood to the model.

add_priors(prior_H_area: float = 1000.0, prior_H_center: Iterable[float] = [0.0, 25.0], prior_H_fwhm: Iterable[float] = [25.0, 10.0], prior_He_H_fwhm_ratio: Iterable[float] = [1.0, 0.1], prior_yplus: float = 0.05, prior_rms: dict[str, float] | None = None, prior_baseline_coeffs: dict[str, Iterable[float]] | None = None, ordered: bool = False)

Add priors and deterministics to the model

Parameters:

prior_H_areafloat, optional: Prior distribution on H RRL line area (mK km s-1), by default 1000.0, where H_area ~ HalfNormal(sigma=prior)
prior_H_centerIterable[float], optional: Prior distribution on H RRL center velocity (km s-1), by default [0, 25.0], where H_center ~ Normal(mu=prior[0], sigma=prior[1])
prior_H_fwhmIterable[float], optional: Prior distribution on H RRL FWHM line width (km s-1), by default [25.0, 10.0], where H_fwhm ~ Gamma(mu=prior[0], sigma=prior[1])
prior_He_H_fwhm_ratioIterable[float], optional: Prior distribution on He/H RRL FWHM line width ratio, by default [1.0, 0.1], where He_H_fwhm_ratio ~ Gamma(mu=prior[0], sigma=prior[1])
prior_yplusfloat, optional: Prior distribution on y+, by default 0.05, where yplus ~ HalfNormal(sigma=prior_yplus)
prior_rmsOptional[dict[str, float]], optional: Prior distribution on spectral rms (K), by default None, where rms ~ HalfNormal(sigma=prior) Keys are dataset names and values are priors. If None, then the spectral rms is taken from dataset.noise and not inferred.
prior_baseline_coeffsOptional[dict[str, Iterable[float]]], optional: Prior distribution on the normalized baseline polynomial coefficients, by default None. Keys are dataset names and values are lists of length baseline_degree+1. If None, use [1.0]*(baseline_degree+1) for each dataset.
orderedbool, optional: If True, assume ordered velocities (optically thin assumption), by default False. If True, the prior distribution on the velocity becomes H_center(cloud = n) ~

prior[0] + sum_i(H_center[i < n]) + Gamma(alpha=2.0, beta=1.0/prior[1])