ready for CRAN submission

.Rbuildignore

@@ -9,13 +9,17 @@ speedtest_sessioninfo.txt
 .undodir
 pre-commit.sh
 
+.gdbinit
+
 ^.*\.Rproj$
 ^\.Rproj\.user$
 
 Session.vim
 dic-example.R
 speedtest_output.txt
 
+debug.Rprofile
+
 .lintr
 compile_flags.txt
 

.Rinstignore

@@ -1,4 +1,4 @@
 bib$
 RData$
 predlik
-Figures2
+jss3813

DESCRIPTION

@@ -2,7 +2,7 @@ Encoding: UTF-8
 Package: stochvol
 Type: Package
 Title: Efficient Bayesian Inference for Stochastic Volatility (SV) Models
-Version: 3.0.10
+Version: 3.1.0
 Authors@R: c(
 	person("Darjus", "Hosszejni", role = c("aut", "cre"), email = "darjus.hosszejni@wu.ac.at", comment = c(ORCID = "0000-0002-3803-691X")),
 	person("Gregor", "Kastner", role = c("aut"), email = "gregor.kastner@aau.at", comment = c(ORCID = "0000-0002-8237-8271")))

NEWS.md

@@ -1,4 +1,4 @@
-# stochvol 3.0.10
+# stochvol 3.1.0
 
 - NEW FUNCTION 'svlm', which has a formula interface;
   it is a wrapper around 'svsample'; many thanks

tests/testthat/test-cpp.R

@@ -1,4 +1,3 @@
-context("C++")
 test_that("Catch unit tests pass", {
     skip_on_cran()
     expect_cpp_tests_pass("stochvol")

tests/testthat/test-samplers.R

@@ -47,6 +47,9 @@ test_that("vanilla SV passes Geweke test", {
 })
 
 test_that("general SV passes Geweke test", {
+  skip_on_cran()
+  skip_on_ci()
+
   set.seed(60)
   result <- .Call(`_stochvol_geweke_general_cpp`, PACKAGE = "stochvol")
   draws <- result$draws / result$thin
@@ -106,6 +109,9 @@ test_that("general SV passes Geweke test", {
 })
 
 test_that("default fast SV is efficient", {
+  skip_on_cran()
+  skip_on_ci()
+
   set.seed(61)
   n <- 150L
   # Pro-centered
@@ -147,6 +153,9 @@ test_that("default fast SV is efficient", {
 })
 
 test_that("default general SV is efficient", {
+  skip_on_cran()
+  skip_on_ci()
+
   set.seed(61)
   n <- 150L
   # Pro-centered

vignettes/.install_extras

@@ -1,4 +1,4 @@
 predlik
 bib$
 RData$
-Figures2/jss3813-.*\.pdf$
+jss3813

vignettes/article2.Rnw

@@ -98,8 +98,6 @@
 \def\svpars{\bm{\vartheta}}
 \def\svlpars{\bm{\zeta}}
 
-\graphicspath{{Figures2/}}
-
 \author{Darjus Hosszejni\\[.3em]WU Vienna University\\[-.1em]of Economics and Business
 \And Gregor Kastner\\[.3em]University of Klagenfurt}
 \Plainauthor{Darjus Hosszejni, Gregor Kastner}
@@ -180,7 +178,7 @@ knitr::opts_chunk$set(prompt = TRUE,
                       message = FALSE,
                       echo = FALSE,
                       #cache = TRUE,
-                      fig.path = "Figures2/jss3813-",
+                      fig.path = "jss3813-",
                       tidy = FALSE)
 base::options(continue = "+  ", prompt = "R> ", width = 70,
   useFancyQuotes = FALSE)
@@ -926,7 +924,7 @@ Using \fct{covmat}, one can extract the MCMC draws of the implied covariance mat
 <<covn, echo = TRUE, eval=FALSE>>=
 dim(cov_n <- covmat(res))
 @
-shows that we have stored $\Sexpr{res$config$draws/res$config$thin}$ posterior draws of a $\Sexpr{m}\times\Sexpr{m}$ covariance matrix at one point in time, $t = n = \Sexpr{n}$. To check convergence, one can take a look at the trace plot and the autocorrelation function of the log determinant, i.e.,
+shows that we have stored $300$ posterior draws of a $6\times6$ covariance matrix at one point in time, $t = n = 200$. To check convergence, one can take a look at the trace plot and the autocorrelation function of the log determinant, i.e.,
 <<logdetcovn, echo = 2:5, fig.width = 10, fig.height=3.5, results='hide', eval=FALSE>>=
 par(mfrow = c(1, 2), mgp = c(1.7, 0.5, 0), mar = c(3, 3, 1, 1))
 logdet <- function (x) log(det(x))
@@ -943,21 +941,21 @@ The results are visualized in Figure~\ref{fig:logdetcovn}; decent mixing for thi
   \label{fig:logdetcovn}
 \end{figure}
 
-To assess the mixing speed of each individual covariance matrix element, one can check, e.g., the estimated effective sample size (out of $\Sexpr{res$config$draws/res$config$thin}$ draws kept) which is implemented in \pkg{coda}.
+To assess the mixing speed of each individual covariance matrix element, one can check, e.g., the estimated effective sample size (out of $300$ draws kept) which is implemented in \pkg{coda}.
 Again, no major convergence problems are apparent.
 <<covess, echo = TRUE>>=
 round(apply(cov_n, 1:2, coda::effectiveSize))
 @
 
 Assuming that \code{runningstore} was set sufficiently high when sampling, several convenience functions can be used for quick visualizations without having to post-process the MCMC draws. For example, to visualize the time-varying correlation matrices, consider
-<<corimageplot, echo=2, eval=FALSE, results='hide'>>=
+<<corimageplot, echo=2:3, eval=FALSE, results='hide'>>=
 par(mfrow = c(1, 3), xpd = TRUE)
 corimageplot(res, these = seq(1, n, length.out = 3), plotCI = "circle",
   plotdatedist = 2, date.cex = 1.1)
 @
 \begin{figure}[tp]
   \centering
-  \includegraphics[width=\textwidth, clip, trim=22 185 0 174]{jss3813-corimageplot-1}
+  \includegraphics[width=\textwidth, clip, trim=0 167 0 159]{jss3813-corimageplot-1}
   \caption{Three estimated correlation matrices and their posterior uncertainty depicted using circles.
     Inner (outer) radii of the circles illustrate to the posterior mean minus (plus) two standard deviations.
     Colors blue and red represent negative and positive values, respectively.