updated pkgdown site and fixed some small bugs with this

NAMESPACE

@@ -21,3 +21,8 @@ export(strategy_stc)
 export(trial_treatment_effect)
 export(trial_variance)
 importFrom(Rdpack,reprompt)
+importFrom(copula,mvdc)
+importFrom(copula,normalCopula)
+importFrom(copula,rMvdc)
+importFrom(rstanarm,posterior_predict)
+importFrom(rstanarm,stan_glm)

R/IPD_stats.R

@@ -34,9 +34,7 @@
 #' @export
 #'
 strategy_maic <- function(formula = as.formula("y ~ X3 + X4 + trt*X1 + trt*X2"),
-                          R = 1000,
-                          ald) {
-
+                          R = 1000) {
   if (class(formula) != "formula")
     stop("formula argument must be of formula class.")
 
@@ -247,7 +245,7 @@ new_strategy <- function(strategy, ...) {
 hat_Delta_stats <- function(AC.IPD, BC.ALD, strategy, CI = 0.95, ...) {
 
   if (CI <= 0 || CI >= 1) stop("CI argument must be between 0 and 1.")
-  
+
   ##TODO: as method instead?
   if (!inherits(strategy, "strategy"))
     stop("strategy argument must be of a class strategy.")
@@ -329,7 +327,7 @@ IPD_stats.maic <- function(strategy,
                           statistic = maic.boot,
                           R = strategy$R,
                           formula = strategy$formula,
-                          ald = strategy$dat_ALD)
+                          ald = ald)
 
   list(mean = mean(maic_boot$t),
        var = var(maic_boot$t))

R/gcomp_ml.R

@@ -30,6 +30,7 @@ gcomp_ml.boot <- function(data, indices,
 #'
 #' @return Difference of log-odds 
 #' @seealso [strategy_gcomp_ml()], [gcomp_ml.boot()]
+#' @importFrom copula normalCopula mvdc rMvdc
 #' @export
 #'
 gcomp_ml_log_odds_ratio <- function(formula, dat) {
@@ -56,9 +57,9 @@ gcomp_ml_log_odds_ratio <- function(formula, dat) {
   # AC IPD pairwise correlations
   t_rho <- t(rho)  # extract along rows
   cop <-
-    normalCopula(param = t_rho[lower.tri(t_rho)],
-                 dim = n_covariates,
-                 dispstr = "un")
+    copula::normalCopula(param = t_rho[lower.tri(t_rho)],
+                         dim = n_covariates,
+                         dispstr = "un")
 
   # BC covariate means & standard deviations
   mean_sd_margins <- list()
@@ -69,12 +70,12 @@ gcomp_ml_log_odds_ratio <- function(formula, dat) {
   }
 
   # sample covariates from approximate joint distribution using copula
-  mvd <- mvdc(copula = cop,
-              margins = rep("norm", n_covariates),  # Gaussian marginals
-              paramMargins = mean_sd_margins)
+  mvd <- copula::mvdc(copula = cop,
+                      margins = rep("norm", n_covariates),  # Gaussian marginals
+                      paramMargins = mean_sd_margins)
 
   # simulated BC pseudo-population
-  x_star <- as.data.frame(rMvdc(n = 1000, mvd))
+  x_star <- as.data.frame(copula::rMvdc(n = 1000, mvd))
   colnames(x_star) <- covariate_names
 
   # outcome logistic regression fitted to IPD using maximum likelihood

R/gcomp_stan.R

@@ -9,6 +9,8 @@
 #' @template args-ald
 #'
 #' @return A list of \eqn{y^*_A} and \eqn{y^*_C} posterior predictions
+#' @importFrom copula normalCopula mvdc rMvdc
+#' @importFrom rstanarm stan_glm posterior_predict
 #' @export
 #'
 gcomp_stan <- function(formula = as.formula("y ~ X3 + X4 + trt*X1 + trt*X2"),
@@ -27,11 +29,11 @@ gcomp_stan <- function(formula = as.formula("y ~ X3 + X4 + trt*X1 + trt*X2"),
 
   # covariate simulation for BC trial using copula package
   cop <-
-    normalCopula(param = c(rho[1,2],rho[1,3],rho[1,4],
-                                    rho[2,3],rho[2,4],
-                                             rho[3,4]),
-                 dim = n_covariates,
-                 dispstr = "un")  # AC IPD pairwise correlations
+    copula::normalCopula(param = c(rho[1,2],rho[1,3],rho[1,4],
+                                   rho[2,3],rho[2,4],
+                                   rho[3,4]),
+                         dim = n_covariates,
+                         dispstr = "un")  # AC IPD pairwise correlations
 
   # sample covariates from approximate joint distribution using copula
   mvd <- mvdc(copula = cop,
@@ -48,11 +50,11 @@ gcomp_stan <- function(formula = as.formula("y ~ X3 + X4 + trt*X1 + trt*X2"),
   colnames(x_star) <- cov_names
 
   # outcome logistic regression fitted to IPD using MCMC (Stan)
-  outcome.model <- stan_glm(formula,
-                            data = ipd,
-                            family = binomial,
-                            algorithm = "sampling",
-                            iter = 4000, warmup = 2000, chains = 2)
+  outcome.model <- rstanarm::stan_glm(formula,
+                                      data = ipd,
+                                      family = binomial,
+                                      algorithm = "sampling",
+                                      iter = 4000, warmup = 2000, chains = 2)
   # counterfactual datasets
   data.trtA <- data.trtC <- x_star
 
@@ -62,7 +64,7 @@ gcomp_stan <- function(formula = as.formula("y ~ X3 + X4 + trt*X1 + trt*X2"),
 
   # draw binary responses from posterior predictive distribution
   list(
-    y.star.A = posterior_predict(outcome.model, newdata=data.trtA),
-    y.star.C = posterior_predict(outcome.model, newdata=data.trtC))
+    y.star.A = rstanarm::posterior_predict(outcome.model, newdata=data.trtA),
+    y.star.C = rstanarm::posterior_predict(outcome.model, newdata=data.trtC))
 }