tulpa_joint_axis_specs_from_grid: Rebuild axis specs from an already-assembled outer grid

In tulpa: Templated Unified Library for Posterior Approximation in Bayesian Hierarchical Models

Rebuild axis specs from an already-assembled outer grid

Description

Rebuilds each axis's declared spec (support, log/linear coordinate, prior) from a theta_grid's columns, keyed off column names – the same metadata the multi-block joint driver recovers rather than carrying a second description of the axes it already assembled. A column holding one value across every cell is treated as a fixed setting rather than an axis (e.g. a fixed NB-size node) and dropped. Intended for a caller that needs an axis's declared span, coordinate or prior off a settled grid without restating tulpa's own hyperprior declarations alongside it.

Usage

tulpa_joint_axis_specs_from_grid(
  theta_grid,
  copy_slab = "exponential",
  folded_axes = NULL,
  logchol = .hp_logchol_designs(theta_grid)
)

Arguments

theta_grid	A named ⁠[n_cells x n_axes]⁠ matrix (see tulpa_theta_matrix()).
copy_slab	"exponential" or "flat"; the copy-scale axis's continuum measure (see tulpa_hyper_check_copy_slab()).
folded_axes	Optional names of axes folded onto ⁠[0, Inf)⁠.
logchol	Free-covariance block design(s) as returned by .hp_logchol_designs(); defaults to the design theta_grid itself declares.

Value

A named list of per-axis spec lists, or NULL if theta_grid has no axis names.

See Also

tulpa_theta_matrix(), tulpa_grid_log_quad(), tulpa_hyper_grid_supports()

Examples

set.seed(1)
S <- 30L                                   # spatial units in a chain
nb <- lapply(seq_len(S), function(s) setdiff(c(s - 1L, s + 1L), c(0L, S + 1L)))
nn <- lengths(nb)
field <- as.numeric(scale(cumsum(rnorm(S, 0, 0.4))))
idx <- rep(seq_len(S), each = 6L); n <- length(idx); x <- rnorm(n)
y <- rbinom(n, 1L, plogis(-0.2 + 0.6 * x + field[idx]))
prior <- list(type = "icar", n_spatial_units = S, spatial_idx = idx,
              adj_row_ptr = c(0L, cumsum(nn)), adj_col_idx = unlist(nb) - 1L,
              n_neighbors = nn, tau_grid = c(0.5, 1, 2, 4, 8))
fit <- tulpa_nested_laplace(y, rep(1L, n), cbind(1, x), prior = prior,
                            family = "binomial",
                            control = list(progress = FALSE))
specs <- tulpa_joint_axis_specs_from_grid(tulpa_theta_matrix(fit))
specs[[1]]$log_scale

tulpa documentation built on Oct. 19, 2026, 5:09 p.m.