Skip to contents

Build Stan Data List for a Model Family

Source: R/stan_data.R
build_stan_data.Rd

Converts a preprocessed stacked data frame into the named list format expected by the Stan model's data {} block.

Usage

build_stan_data(
  standards,
  response_variable,
  priors,
  model_family = "logistic4",
  curve_id_map,
  blanks = NULL,
  use_heteroscedastic_noise = FALSE,
  grainsize = 1L
)

Arguments

standards

Data frame. Preprocessed stacked standards with curve_id, concentration, and response column.

response_variable

Character. Response column name.

priors

Named list from compute_dynamic_priors().

model_family

Character. One of the curveRcore model names.

curve_id_map

Named integer vector mapping curve_id values to 1-based Stan indices.

blanks

Data frame or NULL. Blank well data. Must contain curve_id and the response column (on the fitting scale). Passed directly to Stan to anchor the lower asymptote.

use_heteroscedastic_noise

Logical. If TRUE, passes use_heteroscedastic_noise = 1L to Stan, enabling the power-of-mean noise model sigma_i = exp(log_sigma0 + log_sigma_slope * log(|mu_i|)). If FALSE (default), the homoscedastic sigma_obs path is used.

grainsize

Integer. reduce_sum grain size. Default 1. NOTE: grainsize is declared in the Stan data block but reduce_sum is not yet active in the model bodies; this field is reserved for a future threading implementation and is currently ignored by Stan.

Value

Named list suitable for cmdstanr::sample(data = ...).