Merge pull request #1 from generable/dev

Dev

R/fits-StanModelFit.R

@@ -54,6 +54,16 @@ StanModelFit <- R6::R6Class("StanModelFit",
       private$stan_data <- stan_data
     },
 
+    save_object = function(file, ...) {
+      # replicate internals of `cmdstanr::CmdStanFit$save_object`
+      private$stan_fit$draws()
+      try(private$stan_fit$sampler_diagnostics(), silent = TRUE)
+      try(private$stan_fit$init(), silent = TRUE)
+      try(private$stan_fit$profiles(), silent = TRUE)
+      saveRDS(self, file = file, ...)
+      invisible(self)
+    },
+
     #' @description Get the underlying 'Stan' fit object.
     get_stan_fit = function() {
       private$stan_fit

R/main-treatment_effect.R

@@ -81,6 +81,8 @@ treatment_effect <- function(fit_post, fit_prior, time_var = "t",
 trajectory_metrics <- function(trajectories, id_var, time_var) {
   checkmate::assert_class(trajectories, "FunctionDraws")
   checkmate::assert_character(id_var, len = 1)
+  # use first provided time_var
+  time_var <- time_var[[1]]
   checkmate::assert_character(time_var, len = 1)
   df <- trajectories$as_data_frame_long()
   df %>%
@@ -120,12 +122,12 @@ predict_new_subjects <- function(fit_post, fit_prior,
                                  prior_param_names = NULL) {
   checkmate::assert_class(fit_post, "TSModelFit")
   checkmate::assert_class(fit_prior, "TSModelFit")
-  t_data <- fit_post$get_data("LON")[[time_var]]
+  t_data <- fit_post$get_data("LON")[[time_var[[1]]]]
   t_pred <- t_pred_auto(t_pred, t_data)
 
   # Pick one subject from each group
   newdat <- group_pred_input(fit_post, group_var, t_pred, time_var)
-
+  newdat <- add_sff_input(newdat, fit_post$get_model())
   # Take draws
   if (is.null(prior_param_names)) {
     prior_param_names <- id_specific_param_names(fit_post)

R/utils-data.R

@@ -3,12 +3,14 @@
 #' @export
 #' @param df original data frame
 #' @param t vector of new time values
-#' @param time_var name of time variable
+#' @param time_var name of time variables
 #' @return new data frame
 extend_df <- function(df, t, time_var) {
   u <- df_unique_factor_rows(df)
   df <- df_replicate_rows(u, length(t))
-  df[[time_var]] <- rep(t, nrow(u))
+  for (v in time_var) {
+    df[[v]] <- rep(t, nrow(u))
+  }
   df
 }
 

R/utils-stancode.R

@@ -67,10 +67,13 @@ stancode_ts_gq <- function(mod, stanname_y, sc_terms_gq) {
     "  vector[n_LON] f_sum_capped = fmin(f_sum, ", cap_val, ");"
   )
   def_ycap <- paste0("  vector[n_LON] ", y_var, "_log_pred_capped;")
+  line_ll <- paste0(
+    "    log_lik[i] = normal_lpdf(", stanname_y, "_log[i] | ", "f_sum[i], sigma);"
+  )
   line_yp <- paste0(
     "    ", sylp, "[i] = normal_rng(", "f_sum[i], sigma);"
   )
-  loop <- paste0("  for(i in 1:n_LON) {\n", line_yp, "\n  }")
+  loop <- paste0("  for(i in 1:n_LON) {\n", line_yp, "\n", line_ll, "\n }")
   line_ycap <- paste0(
     "  ", y_var, "_log_pred_capped = fmin(", sylp, ", ",
     cap_val, ");"