posterior_extraction_helpers: Helper functions for extracting and formatting posterior distributions

Description

Internal helper functions to extract posterior samples from JAGS fits and reformat them for further processing (summary tables, diagnostics, plots). These functions consolidate common logic that was duplicated across runjags_estimates_table, .diagnostics_plot_data, and plotting functions.

Helper to remove all columns associated with a parameter

Usage

.extract_posterior_samples(fit, as_list = FALSE)
.remove_auxiliary_parameters(
  model_samples,
  prior_list,
  remove_parameters = NULL
)
.remove_parameter_columns(model_samples, prior, par_name)
.filter_parameters(
  prior_list,
  remove_parameters = NULL,
  remove_formulas = NULL,
  keep_parameters = NULL,
  keep_formulas = NULL,
  remove_spike_0 = TRUE
)
.process_spike_and_slab(
  model_samples,
  prior_list,
  par,
  conditional = FALSE,
  remove_inclusion = FALSE,
  warnings = NULL
)
.apply_parameter_transformations(
  model_samples,
  transformations,
  prior_list,
  transform_factors = FALSE
)
.transform_factor_contrasts(
  model_samples,
  prior_list,
  transform_factors = FALSE,
  transformations = NULL
)
.rename_factor_levels(model_samples, prior_list)

Value

matrix or mcmc.list of posterior samples

list with cleaned model_samples and updated prior_list

updated model_samples matrix

list with filtered model_samples and prior_list

list with updated model_samples, prior_list, and warnings

updated model_samples matrix

updated model_samples matrix with renamed columns

Arguments

fit: a JAGS model fit object
as_list: whether to return samples as mcmc.list (TRUE) or merged matrix (FALSE)
model_samples: matrix of posterior samples
prior_list: list of prior objects
remove_parameters: character vector of parameter names to remove, or TRUE to remove all non-formula parameters. If "bias" is specified and the bias prior contains PET, PEESE, or weightfunction priors, the corresponding parameters (PET, PEESE, omega) are also added to the removal list.
prior: prior object for the parameter
par_name: name of the parameter
remove_formulas: character vector of formula names whose parameters should be removed
keep_parameters: character vector of parameter names to keep (all others removed unless in keep_formulas). If "bias" is specified and the bias prior contains PET, PEESE, or weightfunction priors, the corresponding parameters (PET, PEESE, omega) are also added to the keep list.
keep_formulas: character vector of formula names whose parameters should be kept (all others removed unless in keep_parameters)
remove_spike_0: whether to remove spike at 0 priors
par: parameter name
conditional: whether to compute conditional summary
remove_inclusion: whether to remove inclusion indicators
warnings: character vector for collecting warnings
transformations: list of transformations to apply
transform_factors: whether to transform orthonormal/meandif to differences