Merge branch 'cran-pretest'

DESCRIPTION

@@ -1,12 +1,13 @@
 Package: Hmsc
 Title: Hierarchical Model of Species Communities
 Type: Package
-Version: 3.0-12
+Version: 3.0-13
 Authors@R: c(person(given = "Gleb", family="Tikhonov", role="aut"),
    person(given="Otso", family="Ovaskainen",
       email="otso.ovaskainen@helsinki.fi",
-      comment = c(ORCID="0000-0001-9750-4421"), role=c("aut","cre")),
-   person(given="Jari", family="Oksanen", role="aut"),
+      comment = c(ORCID="0000-0001-9750-4421"), role=c("aut", "cre")),
+   person(given="Jari", family="Oksanen", role=c("aut"),
+   email="jhoksane@gmail.com"),
    person(given="Melinda", family=c("de", "Jonge"), role="aut"),
    person(given="Oystein", family="Opedal", role="aut"),
    person(given="Tad", family="Dallas", role="aut"))
@@ -16,11 +17,11 @@ Description: Hierarchical Modelling of Species Communities (HMSC) is
    Markov chain Monte Carlo (MCMC) sampling (Tikhonov et al. (2020)
    <doi:10.1111/2041-210X.13345>).
 License: GPL-3 | file LICENSE
-URL: https://www.helsinki.fi/en/researchgroups/statistical-ecology/hmsc
+URL: https://www.helsinki.fi/en/researchgroups/statistical-ecology/software/hmsc
 BugReports: https://github.com/hmsc-r/HMSC/issues/
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 7.1.2
+RoxygenNote: 7.2.1
 Suggests: R.rsp,
    testthat,
    corrplot

R/c.Hmsc.R

@@ -15,15 +15,14 @@
 #' @return An \code{\link{Hmsc}} model with chains of posterior
 #'     samples.
 #' @examples
-#' ## Fitted model with two chains
-#' TD$m
-#' ## New chains: check carefully that these are sampled exactly like
-#' ## the previous model
-#' m2 <- sampleMcmc(TD$m, nChains=1, samples=100, thin=1, transient=50,
-#'     verbose=0)
-#' ## Now three chains
-#' mnew <- c(TD$m, m2)
-#' mnew
+#' ## Fit a toy model with two chains
+#' m1 <- sampleMcmc(TD$m, samples=10, transient=5, nChains=2, verbose=0)
+#' ## Need more data? Add chains: check carefully that these are
+#' ## sampled exactly like the previous model
+#' m2 <- sampleMcmc(TD$m, nChains=2, samples=10, transient=5, verbose=0)
+#' ## Now four chains
+#' m4 <- c(m1, m2)
+#' m4
 #'
 #' @export
 `c.Hmsc` <-

R/predict.R

@@ -76,7 +76,7 @@ predict.Hmsc = function(object, post=poolMcmcChains(object$postList), XData=NULL
                         Gradient=NULL, Yc=NULL, mcmcStep=1, expected=FALSE,
                         predictEtaMean=FALSE, predictEtaMeanField=FALSE,
                         nParallel = 1,
-                        useSocket = .Platform$OS.type == "windows", ...)
+                        useSocket = TRUE, ...)
 {
    ## check valid nParallel
    nParallel <- min(nParallel, detectCores())

R/sampleMcmc.R

@@ -81,7 +81,7 @@ sampleMcmc =
     function(hM, samples, transient=0, thin=1, initPar=NULL,
              verbose, adaptNf=rep(transient,hM$nr),
              nChains=1, nParallel=1,
-             useSocket = .Platform$OS.type == "windows",
+             useSocket = TRUE,
              dataParList=NULL, updater=list(),
              fromPrior = FALSE, alignPost = TRUE)
 {

inst/NEWS.md

@@ -1,4 +1,4 @@
-Version 3.0-12 (not released, only in github)
+Version 3.0-13
 ==============
 
 ### New Features
@@ -18,7 +18,7 @@ Version 3.0-12 (not released, only in github)
   but other platforms can profit from the use of fork clusters which
   may have lower memory use and are faster to set up, and also may be
   marginally faster. The choice can be made with new argument
-  `useSocket` which `TRUE` in Windows.
+  `useSocket` (defaults `TRUE`).
 
 * Updaters in `sampleMcmc` can occasionally fail in extreme `Hmsc`
   models. This is no longer an error that stops analysis, but sampling
@@ -27,7 +27,7 @@ Version 3.0-12 (not released, only in github)
   of failures, the sampling is safe to use. If there are updater
   errors only in some chains, these chains can be removed, but other
   chains can be used. See
-  [issue #123](https://github.com/hmsc-r/HMCS/issue/123).
+  [issue #123](https://github.com/hmsc-r/HMSC/issues/123).
 
 * New experimental function `pcomputePredictedValues` with more
   aggressive parallelization than `computePredictedValues`. In old
@@ -46,8 +46,8 @@ Version 3.0-12 (not released, only in github)
   instead of spatial coordinates.
 
 * `constructGradient` provides wider choice of coordinates for
-  centroid, including user-set and infinite (meaning no spatial
-  dependence) coordinates.
+  centroid of `new_unit`, including user-set and infinite (meaning no
+  spatial dependence) coordinates.
 
 * Detect cases when user tries to analyse posterior samples of
   non-sampled `Hmsc` object to avoid confusing error messages such as
@@ -85,7 +85,7 @@ Version 3.0-12 (not released, only in github)
   `X` instead of model frame `XData` and model formula
   `XFormula`. Concerns functions `biPlot`,
   `computeVariancePartitioning` and `constructGradient`. Fixes
-  [issue #126](https://github.com/dshmsc-r/HMSC/issues/126).
+  [issue #126](https://github.com/hmsc-r/HMSC/issues/126).
 
 * `biPlot` has improved handling of colour scaling of continuous
   variables.

man/Hmsc-package.Rd

@@ -94,7 +94,7 @@ Hierarchical Modelling of Species Communities (Hmsc) is a
 \seealso{
 Useful links:
 \itemize{
-  \item \url{https://www.helsinki.fi/en/researchgroups/statistical-ecology/hmsc}
+  \item \url{https://www.helsinki.fi/en/researchgroups/statistical-ecology/software/hmsc}
   \item Report bugs at \url{https://github.com/hmsc-r/HMSC/issues/}
 }
 

tests/Examples/Hmsc-Ex.Rout.save

@@ -1,6 +1,6 @@
 
-R version 4.1.2 (2021-11-01) -- "Bird Hippie"
-Copyright (C) 2021 The R Foundation for Statistical Computing
+R version 4.2.1 (2022-06-23) -- "Funny-Looking Kid"
+Copyright (C) 2022 The R Foundation for Statistical Computing
 Platform: aarch64-apple-darwin20 (64-bit)
 
 R is free software and comes with ABSOLUTELY NO WARRANTY.
@@ -129,20 +129,22 @@ Loading required package: coda
 > 
 > ### ** Examples
 > 
-> ## Fitted model with two chains
-> TD$m
-Hmsc object with 50 sampling units, 4 species, 3 covariates, 3 traits and 2 random levels
-Posterior MCMC sampling with 2 chains each with 100 samples, thin 1 and transient 50 
-> ## New chains: check carefully that these are sampled exactly like
-> ## the previous model
-> m2 <- sampleMcmc(TD$m, nChains=1, samples=100, thin=1, transient=50,
-+     verbose=0)
+> ## Fit a toy model with two chains
+> m1 <- sampleMcmc(TD$m, samples=10, transient=5, nChains=2, verbose=0)
+setting updater$Gamma2=FALSE due to specified phylogeny matrix
+Computing chain 1
+Computing chain 2
+> ## Need more data? Add chains: check carefully that these are
+> ## sampled exactly like the previous model
+> m2 <- sampleMcmc(TD$m, nChains=2, samples=10, transient=5, verbose=0)
 setting updater$Gamma2=FALSE due to specified phylogeny matrix
-> ## Now three chains
-> mnew <- c(TD$m, m2)
-> mnew
+Computing chain 1
+Computing chain 2
+> ## Now four chains
+> m4 <- c(m1, m2)
+> m4
 Hmsc object with 50 sampling units, 4 species, 3 covariates, 3 traits and 2 random levels
-Posterior MCMC sampling with 3 chains each with 100 samples, thin 1 and transient 50 
+Posterior MCMC sampling with 4 chains each with 10 samples, thin 1 and transient 5 
 > 
 > 
 > 
@@ -173,7 +175,7 @@ Posterior MCMC sampling with 3 chains each with 100 samples, thin 1 and transien
 > 
 > ### Name: computePredictedValues
 > ### Title: computePredictedValues
-> ### Aliases: computePredictedValues
+> ### Aliases: computePredictedValues pcomputePredictedValues
 > 
 > ### ** Examples
 > 
@@ -549,7 +551,7 @@ Chain 1, iteration 10 of 10 (sampling)
 > cleanEx()
 > options(digits = 7L)
 > base::cat("Time elapsed: ", proc.time() - base::get("ptime", pos = 'CheckExEnv'),"\n")
-Time elapsed:  4.28 0.087 4.392 0.001 0.001 
+Time elapsed:  2.824 0.061 2.9 0.004 0.007 
 > grDevices::dev.off()
 null device 
           1