Renders the labels of parameter quantities from their
structured label parts: catalog quantities (the label_parts column of
parameter_catalog()) and the columns of mixed and marginal posterior
draws (the quantities draw metadata, see posterior_metadata()). The same
renderer produces the catalog aliases, the rows of the summary tables, the
column names of mixed posteriors, plot legends, and warning text, so a label
shown by BayesTools is always a label of the same quantity everywhere.
parameter_labels(
x,
style = c("selector", "table", "plot", "warning"),
formula_prefix = TRUE,
simplify = FALSE,
vocabulary = NULL
)a parameter catalog (or its quantities or aliases table), a
parameter selection returned by parameter_catalog_resolve(), a data
frame with a label_parts column (e.g. the quantities attribute of an
estimates table), or a list of label parts.
one of "selector" (the exact selector, i.e., the canonical
name of a catalog quantity; also for the quantities of an extension
provider that have no label parts), "table" (the summary-table row label),
"plot" (the plot-legend label: the level text of a level cell), or
"warning" (the row label used in warnings and footnotes).
whether table and warning labels show the formula
parameter, as in (mu) x. Defaults to TRUE.
whether random-effect labels use their simplified display
arguments (for example, sd for a sole random intercept). Defaults to
FALSE.
optional named character vector renaming the
random-effect quantities in the labels, e.g. c(sd = "tau", cor = "rho")
renders (mu) study: tau(intercept) for (mu) study: sd(intercept).
Quantities without an entry keep their names. Defaults to NULL.
a character vector of labels.
Square brackets after a factor term hold a level label (g[b]);
interaction cells name the level of every factor (g[b]:h[v], g[b]:x);
contrast coefficients that are not level cells are written with curly
braces (g{1}). Selectors percent-escape syntax-sensitive level characters
and quote interaction tokens as the catalog does; table, plot, and warning
labels show the level labels as they are. The quantities of an ordered
term (prior_ordered()) follow the same grammar: its total is a coefficient
of the term that is not a level, f{total} (f:x{total} for a term with a
continuous predictor, written on the last component that is not an ordinary
factor), and the share of the total reached at a level is a per-level
quantity of the ordered factor, f[share: mid] (f[share: mid]:x, with
the tag on the component of its ordered factor). When the ordinary factors
of the term index several slices, the slice follows the last ordinary
factor: f{total}:g{1}, f[share: mid]:g{1}, g{1}:f{total}. The
inclusion indicator of a mixture total reads f{total} (inclusion). The
label parts of catalog
aliases render the aliases with style = "table"; under a vocabulary,
they render the aliases of the renamed random-effect quantities (including
the pairwise aliases of the shared correlation of cs() and hcs()
blocks).