NEWS.md
An extra example has been added to ru(), involving different Box-Cox transformation parameters for different component variables.
The list object returned from the function ru() has an extra component named trans_fn. This is an R function that performs the inverse transformation from the transformed variable ρ, on which the generalised ratio-of-uniforms method is performed, back to the original variable θ.
The documentation of the functions ru() and ru_rcpp() now note that if the argument n = 0 is passed then the ratio-of-uniforms bounding box is returned but no simulation is performed. This may be useful if the bounding box itself is of interest.
In the description of the argument logf to ru() and ru_rcpp() it is advised that logf = should be provided explicitly to avoid the potential for argument matching problems.
In plot.ru() the arguments xlabs and ylabs have been super-seeded by an argument var_names, which can be used to provide variable names for use as axis labels.
Improvements to the package manual.
Activated 3rd edition of the testthat package
WARNINGs in the CRAN package check results, like “init.c:60:47: warning: a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes] extern SEXP _revdbayes_RcppExport_registerCCallable();” have been avoided.
Dependence on the spatstat package has been removed.
In the S3 plot method plot.ru() warnings, stemming from stats::regularize.values(), which is called by stats::approx(), have been suppressed.
A very basic S3 print method print.ru() has been added, which simply prints the call to ru() or ru_rcpp().
An extra argument shoof has been added to ru() and ru_rcpp() to enable the user to have control over the initial value used to check that a non-zero convergence indicator is indeed spurious.
Functions that are not intended to be called by the user have been moved to rust-internal.R
Use inherits() to check the class of objects returned from try(), rather than class().
Hyperlinks corrected in vignettes.
pkgdown documentation at https://paulnorthrop.github.io/rust/
Further steps are taken in the opimisations used to set up the ratio-of-uniforms bounding box (specifically in find_a() and in cpp_find_a()) to avoid erroneous (lack of) convergence warnings. These can occur if we start the optimisation algorithm too close to the solution.
LF line endings used in inst/include/rust.h and inst/include/rust_RcppExports.h to avoid CRAN NOTE
The summary method for class “ru” is now set up according to Section 8.1 of the R FAQ at (https://cran.r-project.org/doc/FAQ/R-FAQ.html).
In the extra checks used to try to avoid erroneous convergence warnings lower and upper bounds on the parameters are now used.
Corrected minor typo in the penultimate sentence of the “Cauchy density” subsection in the “When can rust be used vignette”: b1* should be b1+.
In ru() and ru_rcpp() the constant hscale is used to shift logf (and hence scale the target density f) in function logf in the returned object. This helps to avoid over/under-flow when contouring f in plot.ru when d = 2.
The var_names argument to ru_rcpp didn’t work. This has been corrected.
The arguments lower, upper to ru and ru_rcpp are now used, at least partly even when rotate = TRUE. See the updated description of lower, upper in the documentation of ru and ru_rcpp.
That the function logf supplied to ru or ru_rcpp should return -Inf when the density f is zero is stated explicitly in the help files.
ru() did not work when trans = "user" and d > 1. This has been corrected.
Extra checks are used to try to avoid erroneous convergence warnings.
Incorrectly formatted link to the Rcpp Gallery webpage corrected in the vignette “Rusting Faster: Simulation using Rcpp”.
Extra examples provided for ru and for ru_rcpp: (a) Cauchy, (b) half-Cauchy and (c) bivariate normal x bivariate student-t.
Corrected C++ function vecpow to avoid compilation errors on some platforms.
Unnecessary dependence on packages devtools and roxygen2 via Suggests is removed.
Minor edit to vignette: provide link directly to example C++ file user_fns.cpp in src directory, rather than the the (identical) example_user_fns.cpp file in the vignettes directory.
Packages Rcpp (https://CRAN.R-project.org/package=Rcpp) and RcppArmadillo (https://CRAN.R-project.org/package=RcppArmadillo) are used to speed up the computations if the user provides a C++ function to evaluate their target log-density.
New functions: ru_rcpp, find_lambda_rcpp and find_lambda_one_d_rcpp.
New vignette. “Rusting faster: Simulation using Rcpp”.
Bug fixed in plot.ru(): previously plot.ru() failed when d > 2 and no axis label names were provided.
Bug fixed in plot.ru : previously, in the d = 2 case, providing the graphical parameter col produced an error because col = 8 was hard-coded in a call to points. Now the extra argument points_par enables the user to provide a list of arguments to points.
“using pairs()” removed from the last sentence Description of plot.ru() because pairs() is not used when d > 2, rather a single plot is produced for each pair of variables.
Obsolete function rho_to_theta() removed from function ru in ru_sampling.R.
If the user calls ru (or ru_rcpp) with trans = "user" but doesn’t supply phi_to_theta then an error is returned.
plot.ru edited to avoid warning message that occurred in the d=1 case when breaks was supplied as an argument.
The functions rgpd, gpd_sum_stats, gpd_init and gpd_logpost are now exported.
plot.ru() can now be used when d > 2: it produces pairwise plots of the simulated values.
find_lamba(): argument init_lambda added to enable the user to supply an initial value for the Box-Cox transformation vector lambda.
Unnecessary print statement print(dim(phi)) removed from function find_lambda().
Unnecessary print statement print(a_algor) removed from function ru().
Correct lambda$init to lambda$init_psi in ru() when extracting Box-Cox information.
Documentation of ru() updated to include a description of the returned function logf_rho() and simulated values sim_vals_rho and to clarify the meaning of the returned value of f_mode.
ru(): the expression for the inverse Box-Cox transformation in the case where lambda is exactly 0 has been corrected.
find_lambda(): carry out calculation of the target on a shifted log scale to avoid underflow.
Set up plot.ru() so that if the user supplies axis labels then they are used and otherwise the column name(s) of ru_object$sim_vals are used. Also enable plotmath symbols to be rendered in the axis labels.