Skip to contents

Compile the BHF Stan Model

Source: R/compile_model.R
compile_bhf_model.Rd

Compiles the Stan model for the Bayesian Hybrid Framework. This function should be called once at the beginning of each R session before fitting any models.

Usage

compile_bhf_model(verbose = TRUE, auto_write = TRUE)

Arguments

verbose

Logical. If TRUE, prints compilation progress. Default is TRUE.

auto_write

Logical. If TRUE, saves compiled model to disk for faster loading in future sessions. Default is TRUE.

Value

An object of class stanmodel that can be used with bhf_fit().

Details

This function compiles the Stan model from source code included in the package. Compilation typically takes 30-60 seconds depending on your system.

The compiled model is cached by rstan (if auto_write = TRUE), so subsequent calls to this function in the same or future sessions will be much faster.

Why Manual Compilation?

Unlike packages that pre-compile Stan models during installation, this package requires explicit compilation. This "defensive" approach:

  • Avoids cryptic errors from cached/stale compiled objects

  • Provides clearer error messages when compilation fails

  • Ensures compatibility across different R/Stan/C++ configurations

  • Gives you visibility into what's happening

Troubleshooting

If compilation fails:

  • Ensure you have a C++ toolchain installed (Rtools on Windows, Xcode CLI on macOS, build-essential on Linux)

  • Check that rstan is properly installed: example(stan_model, package = "rstan")

  • See the package vignette for detailed troubleshooting

Examples

if (FALSE) { # \dontrun{
# Compile the model (once per session)
model <- compile_bhf_model()

# Now you can use the model for fitting
fit <- bhf_fit(prepared_data, model = model)
} # }