Resolve some warnings and bug fixes on iter

R/sparse.R

@@ -12,6 +12,7 @@
 sample_sparse_tmb <- function(obj, iter, warmup, cores, chains,
                               control=NULL, seed=NULL){
   iter <- iter-warmup
+  obj$env$beSilent()
   sdr <- sdreport(obj, getJointPrecision=TRUE)
   Q <- sdr$jointPrecision
   Qinv <- solve(Q)
@@ -95,7 +96,7 @@ as.tmbfit <- function(x, mle, invf){
   mon$Rhat <- mon$rhat
   ## prepare objects to use the pairs_admb function
   post <- get_post(x, invf, parnames=parnames, TRUE)
-  sp <- x@diagnostics
+  sp <- as.data.frame(x@diagnostics)
   spl <- list()
   for(chain in 1:max(sp$.chain)){
     spl[[chain]] <- as.matrix(sp[sp$.chain==chain,1:6])
@@ -104,7 +105,7 @@ as.tmbfit <- function(x, mle, invf){
             monitor=mon, model='test',
             max_treedepth=x@metadata$max_depth,
             warmup=as.numeric(x@metadata$num_warmup),
-            iter=as.numeric(x@metadata$num_samples),
+            ## iter=as.numeric(x@metadata$num_samples)+as.numeric(x@metadata$num_warmup),
             algorithm='NUTS')
   adfit(x)
 }

R/utils.R

@@ -513,7 +513,7 @@ check_identifiable <- function(model, path=getwd()){
     ## matrix operations.
     x.cur <- (1/chd) * y.cur
   } else if(is(M,"Matrix")){
-    warning( "Use of Q is highly experimental still" )
+    ##  warning( "Use of Q is highly experimental still" )
     stopifnot(require(Matrix))
     # M is actually Q, i.e., the inverse-mass
     # Antidiagonal matrix JJ = I
@@ -756,8 +756,12 @@ extract_sampler_params <- function(fit, inc_warmup=FALSE){
     ind <- -(1:fit$warmup)
     its <- (1:nrow(x[[1]]))[ind]
   }
-  y <- do.call(rbind, lapply(1:length(x), function(i)
-    cbind(chain=i, iteration=its, x[[i]][ind,])))
+  y <- do.call(rbind, lapply(1:length(x), function(i){
+    if(length(its) != NROW(x[[i]][ind,]))
+      warning("Length mismatch in extract_sampler_params: iterations=", length(its),
+              " and draws=", NROW(x[[i]][ind,]))
+    cbind(chain=i, iteration=its, x[[i]][ind,])
+  }))
   return(invisible(as.data.frame(y)))
 }