bilby.core.sampler.pymc.Pymc

class bilby.core.sampler.pymc.Pymc(likelihood, priors, outdir='outdir', label='label', use_ratio=False, plot=False, skip_import_verification=False, **kwargs)[source]

Bases: MCMCSampler

bilby wrapper of the PyMC sampler (https://www.pymc.io/)

All keyword arguments (i.e., the kwargs) passed to run_sampler will be propapated to pymc.sample where appropriate, see documentation for that class for further help. Under Other Parameters, we list commonly used kwargs and the bilby, or where appropriate, PyMC defaults.

Parameters:

draws: int, (1000): The number of sample draws from the posterior per chain.
chains: int, (2): The number of independent MCMC chains to run.
cores: int, (1): The number of CPU cores to use.
tune: int, (500): The number of tuning (or burn-in) samples per chain.
discard_tuned_samples: bool, True: Set whether to automatically discard the tuning samples from the final chains.
step: str, dict: Provide a step method name, or dictionary of step method names keyed to particular variable names (these are case insensitive). If passing a dictionary of methods, the values keyed on particular variables can be lists of methods to form compound steps. If no method is provided for any particular variable then PyMC will automatically decide upon a default, with the first option being the NUTS sampler. The currently allowed methods are ‘NUTS’, ‘HamiltonianMC’, ‘Metropolis’, ‘BinaryMetropolis’, ‘BinaryGibbsMetropolis’, ‘Slice’, and ‘CategoricalGibbsMetropolis’. Note: you cannot provide a PyMC step method function itself here as it is outside of the model context manager.
step_kwargs: dict: Options for steps methods other than NUTS. The dictionary is keyed on lowercase step method names with values being dictionaries of keywords for the given step method.

__init__(likelihood, priors, outdir='outdir', label='label', use_ratio=False, plot=False, skip_import_verification=False, **kwargs)[source]

__call__(*args, **kwargs): Call self as a function.

Methods

`__init__`(likelihood, priors[, outdir, ...])
`calc_likelihood_count`()
`calculate_autocorrelation`(samples[, c])	Uses the emcee.autocorr module to estimate the autocorrelation
`check_draw`(theta[, warning])	Checks if the draw will generate an infinite prior or likelihood
`get_initial_points_from_prior`([npoints])	Method to draw a set of live points from the prior
`get_random_draw_from_prior`()	Get a random draw from the prior distribution
`log_likelihood`(theta)	Parameters:
`log_prior`(theta)	Parameters:
`print_nburn_logging_info`()	Prints logging info as to how nburn was calculated
`prior_transform`(theta)	Prior transform method that is passed into the external sampler.
`run_sampler`()	A template method to run in subclasses
`set_likelihood`()	Convert any bilby likelihoods to PyMC distributions.
`set_prior`()	Set the PyMC prior distributions.
`setup_prior_mapping`()	Set the mapping between predefined bilby priors and the equivalent PyMC distributions.
`write_current_state`()
`write_current_state_and_exit`([signum, frame])	Make sure that if a pool of jobs is running only the parent tries to checkpoint and exit.

Attributes

`check_point_equiv_kwargs`
`constraint_parameter_keys`	list: List of parameters providing prior constraints
`default_kwargs`
`default_nuts_kwargs`
`external_sampler_name`
`fixed_parameter_keys`	list: List of parameter keys that are not being sampled
`hard_exit`
`kwargs`	dict: Container for the kwargs.
`nburn_equiv_kwargs`
`ndim`	int: Number of dimensions of the search parameter space
`npool`
`npool_equiv_kwargs`
`nwalkers_equiv_kwargs`
`sampling_seed_equiv_kwargs`
`sampling_seed_key`	Name of keyword argument for setting the sampling for the specific sampler.
`search_parameter_keys`	list: List of parameter keys that are being sampled

calculate_autocorrelation(samples, c=3)[source]

Uses the emcee.autocorr module to estimate the autocorrelation

Parameters:

samples: array_like: A chain of samples.
c: float: The minimum number of autocorrelation times needed to trust the estimate (default: 3). See emcee.autocorr.integrated_time.

check_draw(theta, warning=True)[source]

Checks if the draw will generate an infinite prior or likelihood

Also catches the output of numpy.nan_to_num.

Parameters:

theta: array_like: Parameter values at which to evaluate likelihood
warning: bool: Whether or not to print a warning

Returns:

bool, cube (nlive,: True if the likelihood and prior are finite, false otherwise

property constraint_parameter_keys: list: List of parameters providing prior constraints

property fixed_parameter_keys: list: List of parameter keys that are not being sampled

get_initial_points_from_prior(npoints=1)[source]

Method to draw a set of live points from the prior

This iterates over draws from the prior until all the samples have a finite prior and likelihood (relevant for constrained priors).

Parameters:

npoints: int: The number of values to return

Returns:

unit_cube, parameters, likelihood: tuple of array_like: unit_cube (nlive, ndim) is an array of the prior samples from the unit cube, parameters (nlive, ndim) is the unit_cube array transformed to the target space, while likelihood (nlive) are the likelihood evaluations.

get_random_draw_from_prior()[source]

Get a random draw from the prior distribution

Returns:

draw: array_like: An ndim-length array of values drawn from the prior. Parameters with delta-function (or fixed) priors are not returned

property kwargs: dict: Container for the kwargs. Has more sophisticated logic in subclasses

log_likelihood(theta)[source]

Parameters:

theta: list: List of values for the likelihood parameters

Returns:

float: Log-likelihood or log-likelihood-ratio given the current: likelihood.parameter values

log_prior(theta)[source]

Parameters:

theta: list: List of sampled values on a unit interval

Returns:

float: Joint ln prior probability of theta

property ndim: int: Number of dimensions of the search parameter space

print_nburn_logging_info()[source]: Prints logging info as to how nburn was calculated

prior_transform(theta)[source]

Prior transform method that is passed into the external sampler.

Parameters:

theta: list: List of sampled values on a unit interval

Returns:

list: Properly rescaled sampled values

run_sampler()[source]: A template method to run in subclasses

sampling_seed_key = 'random_seed': Name of keyword argument for setting the sampling for the specific sampler. If a specific sampler does not have a sampling seed option, then it should be left as None.

property search_parameter_keys: list: List of parameter keys that are being sampled

set_likelihood()[source]: Convert any bilby likelihoods to PyMC distributions.

set_prior()[source]: Set the PyMC prior distributions.

setup_prior_mapping()[source]: Set the mapping between predefined bilby priors and the equivalent PyMC distributions.

write_current_state_and_exit(signum=None, frame=None)[source]

Make sure that if a pool of jobs is running only the parent tries to checkpoint and exit. Only the parent has a ‘pool’ attribute.

For samplers that must hard exit (typically due to non-Python process) use os._exit that cannot be excepted. Other samplers exiting can be caught as a SystemExit.