Changes in version 2.1.0                        

cmdstanr backend support

  - bmgarch() now accepts backend = "cmdstanr" in addition to the
    default "rstan". Both MCMC (NUTS) and variational Bayes
    (sampling_algorithm = "VB") are supported with the new backend.

  - summary(), print(), plot(), forecast(), fitted(), loo(), and
    model_weights() all work with cmdstanr-fitted objects.

  - Stan function a_b_scale_jacobian renamed to a_b_scale_log_jac to
    avoid a future name collision with Stan 2.39's reserved _jacobian
    suffix (affects inst/stan/functions/jacobian.stan).

  - The iterations argument now defaults to 2000 for MCMC and 30000 for
    VB when left unspecified. Previously the single default of 2000 was
    too low for ADVI gradient ascent.

  - The backend field is stored in every bmgarch object so that loo()
    refits and other operations that call back into bmgarch()
    automatically use the same backend.

  - Threading defaults: cmdstanr models compiled with stan_threads =
    TRUE now default to threads_per_chain = 1 (MCMC) and threads = 1
    (VB). Users can override via ....

Bug fixes

  - Fixed ensemble covariance underestimation in .weighted_samples().
    The between-model variance term required by the law of total
    variance, Σ_m w_m (μ̂_m − μ̂_mix)(μ̂_m − μ̂_mix)ᵀ, is now added to
    the weighted combination of H and H_forecasted. This corrects the
    ensemble posterior predictive covariance for both backcasts and
    forecasts whenever models with differing mean structures are
    combined. The fix is a no-op for single-model objects.

  - model_weights() and loo() now work with cmdstanr objects: log_lik
    extraction and chain/iteration metadata no longer rely on rstan S4
    slot access (@sim).

  - Weighted forecasts (forecast(..., weights = mw)) now work with
    cmdstanr backends via a new .extract_param_list() helper that
    replicates rstan::extract() output for cmdstanr draw objects.

  - Forecast draws passed to rstan::gqs() are now correctly extracted
    via posterior::as_draws_matrix() for cmdstanr fits, fixing a "no
    dimnames" error.

Tests

  - Added tests/testthat/test-backends.R covering summary, forecast, and
    model_weights for all three combinations: rstan/MCMC, cmdstanr/MCMC,
    and cmdstanr/VB on a CCC model.

                 Changes in version 2.0.0 (2023-09-11)                  

  - All stan models are rewritten to match the new rstan 2.26.0 array
    syntax.

                 Changes in version 1.1.0 (2021-12-14)                  

  - Meanstructure now also takes the VAR argument for a VAR(1) model.
  - Added variational Bayes VB as a sampling algorithm option. bmgarch
    now takes VB and the standard MCMC argument in sampling_algorithm.
    VB is inherited from rstan and is still experimental - use with
    caution.
  - Updated rstantools to facilitate rstan upgrade (according to rstan
    developer guide)
  - standat now checks for constant vectors in data, and returns error
    if there's no variance (cf. discourse mc-stan)

                 Changes in version 1.0.1 (2021-06-14)                  

  - loo function now takes m-ahead argument for the "backward"
    approximation allowing one to tune the forcast to arbitrary steps
    ahead.

  - Fixed package dependency versions.

                 Changes in version 1.0.0 (2020-09-17)                  

Initial CRAN release