validate_mode: Validate that a fit used the expected mode

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

Validate that a fit used the expected mode

Description

Ensures the fit object was created with the expected inference mode. Useful for enforcing mode requirements in downstream analysis.

A fit matches a mode when the mode names either the backend that ran (fit$backend, e.g. "laplace", "mala", "re_cov_gibbs") or the tier it ran in (fit$inference_mode: "exact", "structured", "optimized"). So a fit made with mode = "laplace" validates against both "laplace" and "structured", and a fit made with mode = "auto" validates against the backend and the tier auto chose. A request that tulpa() overrode (see fit$mode_overridden) validates against what ran, not what was asked for.

Usage

validate_mode(fit, expected_mode, error = TRUE)

Arguments

fit	A tulpa_fit object
expected_mode	Mode(s) that should have been used: a character vector of backend and / or tier names, any one of which is accepted.
error	If TRUE (default), error on mismatch. If FALSE, return logical.

Value

If error = FALSE, returns TRUE/FALSE. Otherwise errors on mismatch.

Examples

set.seed(1)
d <- data.frame(x = rnorm(60))
d$y <- rpois(60, exp(0.5 + 0.3 * d$x))
fit <- tulpa(y ~ x, data = d, family = "poisson", mode = "hmc",
             control = list(n_iter = 400L, warmup = 200L, n_chains = 2L,
                            seed = 1L))
validate_mode(fit, "exact")                      # the tier hmc runs in
validate_mode(fit, "structured", error = FALSE)  # FALSE

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