switch to dontrun for most examples; skip some more tests on cran

R/RW.R

@@ -17,7 +17,7 @@
 #' arguments to be interpreted by the parsing functions in \code{mvgam}
 #' @rdname RW
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# A short example to illustrate CAR(1) models
 #'# Function to simulate CAR1 data with seasonality
 #'sim_corcar1 = function(n = 120,

R/as.data.frame.mvgam.R

@@ -32,7 +32,7 @@
 #'@param ... Ignored
 #'@return A `data.frame`, `matrix`, or `array` containing the posterior draws.
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'sim <- sim_mvgam(family = Gamma())
 #'mod1 <- mvgam(y ~ s(season, bs = 'cc'),
 #'              trend_model = 'AR1',

R/conditional_effects.R

@@ -44,7 +44,7 @@
 #' @author Nicholas J Clark
 #' @seealso \code{\link[marginaleffects]{plot_predictions}}, \code{\link[marginaleffects]{plot_slopes}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data
 #' simdat <- sim_mvgam(family = poisson(),
 #'                     seasonality = 'hierarchical')

R/dynamic.R

@@ -39,7 +39,7 @@
 #' @rdname dynamic
 #' @return a `list` object for internal usage in 'mvgam'
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate a time-varying coefficient
 #'#(as a Gaussian Process with length scale = 10)
 #'set.seed(1111)

R/evaluate_mvgams.R

@@ -61,7 +61,7 @@
 #'stable forecasts
 #'@seealso \code{\link{forecast}}, \code{\link{score}}, \code{\link{lfo_cv}}
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate from a Poisson-AR2 model with a seasonal smooth
 #'set.seed(100)
 #'dat <- sim_mvgam(T = 75,

R/families.R

@@ -66,7 +66,7 @@ student_t = function(link = 'identity'){
 
 #' @rdname mvgam_families
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Example showing how to set up N-mixture models
 #' set.seed(999)
 #'# Simulate observations for species 1, which shows a declining trend and 0.7 detection probability

R/forecast.mvgam.R

@@ -28,7 +28,7 @@ forecast <- function(object, ...){
 #'@rdname forecast.mvgam
 #'@method forecast mvgam
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 3, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc', k = 6),
 #'             trend_model = AR(),

R/hindcast.mvgam.R

@@ -16,7 +16,7 @@ hindcast <- function(object, ...){
 #'@rdname hindcast.mvgam
 #'@method hindcast mvgam
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 3, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),
 #'             trend_model = AR(),

R/index-mvgam.R

@@ -17,7 +17,7 @@ NULL
 #' @return a `list` object of the variables that can be extracted, along
 #' with their aliases
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc', k = 6),
 #'              trend_model = AR(),

R/lfo_cv.mvgam.R

@@ -39,7 +39,7 @@
 #'@references Paul-Christian Bürkner, Jonah Gabry & Aki Vehtari (2020). Approximate leave-future-out cross-validation for Bayesian time series models
 #'Journal of Statistical Computation and Simulation. 90:14, 2499-2523.
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate from a Poisson-AR2 model with a seasonal smooth
 #'set.seed(100)
 #'dat <- sim_mvgam(T = 75,
@@ -127,9 +127,11 @@ lfo_cv.mvgam = function(object,
   if(missing(data)){
     all_data <- object$obs_data
   } else {
-    all_data <- data
+    all_data <- validate_series_time(data,
+                                     name = 'data',
+                                     trend_model = attr(object$model_data, 'trend_model'))
   }
-  N <- max(all_data$time)
+  N <- max(all_data$index..time..index)
 
   # Default minimum training time is 30, or
   # whatever training time allows for at least 10 lfo_cv calculations
@@ -440,7 +442,7 @@ cv_split = function(data, last_train, fc_horizon = 1){
   if(inherits(data, 'list')){
 
     # Find indices of training and testing splits
-    temp_dat = data.frame(time = data$time,
+    temp_dat = data.frame(time = data$index..time..index,
                           series = data$series) %>%
       dplyr::mutate(index = dplyr::row_number()) %>%
       dplyr::arrange(time, series)
@@ -472,12 +474,12 @@ cv_split = function(data, last_train, fc_horizon = 1){
 
   } else {
     data_train <- data %>%
-      dplyr::filter(time <= last_train) %>%
-      dplyr::arrange(time, series)
+      dplyr::filter(index..time..index <= last_train) %>%
+      dplyr::arrange(index..time..index, series)
 
     data_test <- data %>%
-      dplyr::filter(time > last_train) %>%
-      dplyr::arrange(time, series)
+      dplyr::filter(index..time..index > last_train) %>%
+      dplyr::arrange(index..time..index, series)
 
   }
 

R/logLik.mvgam.R

@@ -20,7 +20,7 @@
 #'original model via the `newdata` argument in \code{\link{mvgam}},
 #'testing observations)
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc', k = 6),

R/loo.mvgam.R

@@ -10,7 +10,7 @@
 #' for details). For `loo_compare.mvgam`, an object of class `compare.loo` (
 #' [loo::loo_compare()] for details)
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate 4 time series with hierarchical seasonality
 #'# and independent AR1 dynamic processes
 #'set.seed(111)

R/lv_correlations.R

@@ -8,7 +8,7 @@
 #'@return A \code{list} object containing the mean posterior correlations
 #'and the full array of posterior correlations
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'simdat <- sim_mvgam()
 #'mod <- mvgam(y ~ s(season, bs = 'cc',
 #'                   k = 6),

R/mcmc_plot.mvgam.R

@@ -22,7 +22,7 @@
 #' @seealso \code{\link{mvgam_draws}} for an overview of some of the shortcut strings
 #' that can be used for argument `variable`
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc', k = 6),
 #'              trend_model = AR(),

R/monotonic.R

@@ -44,7 +44,7 @@
 #' @author Nicholas J Clark
 #' @name monotonic
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate data from a monotonically increasing function
 #' set.seed(123123)
 #' x <- runif(80) * 4 - 1

R/mvgam_diagnostics.R

@@ -15,7 +15,7 @@
 #'
 #' @return The exact form of the output depends on the method.
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc', k = 6),
 #'             trend_model = AR(),

R/pairs.mvgam.R

@@ -14,7 +14,7 @@
 #'   \code{\link[bayesplot:MCMC-scatterplots]{mcmc_pairs}}.
 #'
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),
 #'              trend_model = AR(),

R/plot.mvgam.R

@@ -44,7 +44,7 @@
 #'@author Nicholas J Clark
 #'@return A base R plot or set of plots
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate some time series
 #'dat <- sim_mvgam(T = 80, n_series = 3)
 #'

R/plot_mvgam_factors.R

@@ -16,7 +16,7 @@
 #'@return A \code{dataframe} of factor contributions and,
 #'optionally, a series of base \code{R} plots
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'simdat <- sim_mvgam()
 #'mod <- mvgam(y ~ s(season, bs = 'cc',
 #'                   k = 6),

R/posterior_epred.mvgam.R

@@ -39,7 +39,7 @@
 #' @seealso \code{\link{hindcast.mvgam}} \code{\link{posterior_linpred.mvgam}} \code{\link{posterior_predict.mvgam}}
 #' @aliases posterior_epred
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),
@@ -110,7 +110,7 @@ posterior_epred.mvgam = function(object,
 #' and \code{n_obs} is the number of observations in \code{newdata}
 #' @seealso \code{\link{hindcast.mvgam}} \code{\link{posterior_epred.mvgam}} \code{\link{posterior_predict.mvgam}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),
@@ -193,7 +193,7 @@ posterior_linpred.mvgam = function(object,
 #' and \code{n_obs} is the number of observations in \code{newdata}
 #' @seealso \code{\link{hindcast.mvgam}} \code{\link{posterior_linpred.mvgam}} \code{\link{posterior_epred.mvgam}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),
@@ -260,7 +260,7 @@ posterior_predict.mvgam = function(object,
 #'   quantile estimates as specified via argument \code{probs}.
 #' @seealso \code{\link{hindcast.mvgam}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),

R/ppc.mvgam.R

@@ -47,7 +47,7 @@
 #'@author Nicholas J Clark
 #'@seealso \code{\link{pp_check.mvgam}}, \code{\link{predict.mvgam}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some smooth effects and fit a model
 #' set.seed(0)
 #' dat <- mgcv::gamSim(1, n = 200, scale = 2)
@@ -785,7 +785,7 @@ ppc.mvgam = function(object, newdata, data_test, series = 1, type = 'hist',
 #' package.
 #' @seealso \code{\link{ppc}} \code{\link{predict.mvgam}}
 #' @examples
-#' \donttest{
+#' \dontrun{
 #'simdat <- sim_mvgam(seasonality = 'hierarchical')
 #'mod <- mvgam(y ~ series +
 #'               s(season, bs = 'cc', k = 6) +

R/predict.mvgam.R

@@ -60,6 +60,32 @@
 #'   `predict.gam(object, type = "terms", se.fit = TRUE)`, where mean contributions
 #'   from each effect are returned in `matrix` form while standard errors (representing
 #'   the interval: `(max(probs) - min(probs)) / 2`) are returned in a separate `matrix`
+#'@examples
+#'\dontrun{
+#'# Simulate 4 time series with hierarchical seasonality
+#'# and independent AR1 dynamic processes
+#'set.seed(111)
+#'simdat <- sim_mvgam(seasonality = 'hierarchical',
+#'                    trend_model = 'AR1',
+#'                    family = gaussian())
+#'
+#'# Fit a model with shared seasonality
+#'mod1 <- mvgam(y ~ s(season, bs = 'cc', k = 6),
+#'              data = simdat$data_train,
+#'              family = gaussian(),
+#'              trend_model = AR(),
+#'              burnin = 300,
+#'              samples = 300,
+#'              chains = 2)
+#'
+#'# Generate predictions against observed data
+#'preds <- predict(mod1, summary = TRUE)
+#'head(preds)
+#'
+#'# Generate predictions against test data
+#'preds <- predict(mod1, newdata = simdat$data_test, summary = TRUE)
+#'head(preds)
+#'}
 #'@export
 predict.mvgam = function(object,
                          newdata,

R/residuals.mvgam.R

@@ -22,7 +22,7 @@
 #'   \code{robust}). The remaining columns starting with \code{Q} contain
 #'   quantile estimates as specified via argument \code{probs}.
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = 'AR1')
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),

R/score.mvgam_forecast.R

@@ -31,7 +31,7 @@
 #'posterior empirical quantiles. Intervals are not calculated when using `elpd` because forecasts
 #'will only contain the linear predictors
 #'@examples
-#'\donttest{
+#'\dontrun{
 #'# Simulate observations for three count-valued time series
 #'data <- sim_mvgam()
 #'# Fit a dynamic model using 'newdata' to automatically produce forecasts

R/update.mvgam.R

@@ -16,7 +16,7 @@
 #' for other functions in the package. See \code{\link{mvgam-class}} for details.
 #' Use `methods(class = "mvgam")` for an overview on available methods.
 #' @examples
-#' \donttest{
+#' \dontrun{
 #' # Simulate some data and fit a Poisson AR1 model
 #' simdat <- sim_mvgam(n_series = 1, trend_model = AR())
 #' mod <- mvgam(y ~ s(season, bs = 'cc'),

cran-comments.md

@@ -4,7 +4,7 @@
 * DESCRIPTION has been shortened appropriately
 * Missing Rd tags have been added for all exported functions
 * Examples that previously used non-exported functions have been fixed
-* `\dontrun` has been replaced with `\donttest` throughout, but note that many of these examples take some time to run due to the need for 'Stan' models to compile. Using `--run-donttest`, the examples take between 38 and 50 minutes, depending on the test environment. A very similar R package that is on CRAN ('brms') sticks to the `\dontrun` convention because of this 
+* `\dontrun` was replaced with `\donttest` throughout, but this is likely to cause problems in any later CRAN checks because many of these examples take some time to run due to the need for 'Stan' models to compile. Using `--run-donttest`, the examples take between 38 and 50 minutes, depending on the test environment. A very similar R package that is on CRAN ('brms') sticks to the `\dontrun` convention because of this, so I have elected to only use `\donttest` in the examples for the package's primary function `mvgam()` 
 * `cat()` has been replaced with `message()` throughout
 * `onexit()` has been used as suggested to ensure the user's `par` is not changed