bump version, adjusting CRAN checks

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: lognorm
 Title: Functions for the Lognormal Distribution
-Version: 0.1.8
+Version: 0.1.9
 Author: Thomas Wutzler
 Maintainer: Thomas Wutzler <[email protected]>
 Description: The lognormal distribution  
@@ -13,7 +13,7 @@ Description: The lognormal distribution
   difference of two correlated lognormally distributed variables
   (Lo 2012 <doi:10.1155/2012/838397>).
 Imports: Matrix
-Suggests: testthat, knitr,  dplyr, ggplot2, mvtnorm, purrr, tidyr
+Suggests: markdown, rmarkdown, testthat, knitr,  dplyr, ggplot2, mvtnorm, purrr, tidyr
 VignetteBuilder: knitr
 License: GPL-2
 LazyData: true

NEWS.md

@@ -1,5 +1,5 @@
 # lognorm 0.1.9
-- Improved handling of missing values in computeEffectiveNumObs
+- Less bias with missing values in computeEffectiveNumObs
 - New function varCor to compute unbiased variance of uncorrelated time series
 - seCor now based on varCor, which reduces bias for small number of effective
   observations.

R/coefLognorm.R

@@ -64,7 +64,7 @@ getParmsLognormForLowerAndUpperLog <- function(
 #'    i.e. practical maximum
 #' @param sigmaFac sigmaFac=2 is 95\% sigmaFac=2.6 is 99\% interval.
 #'
-#' @return numeric matrix wiht columns `mu` and `sigma`, the parameter of the
+#' @return numeric matrix with columns `mu` and `sigma`, the parameter of the
 #'   lognormal distribution. Rows correspond to rows of inputs.
 #' @export
 #' @examples
@@ -171,7 +171,7 @@ getParmsLognormForExpval <- function(mean, sigmaStar){
 #' NA values should be stripped before the computation proceeds.
 #'
 #' @return numeric vector with components \code{mu} and \code{sigma},
-#' ie.., the center parameter (mean at log scale, log(median)) and 
+#' i.e., the center parameter (mean at log scale, log(median)) and 
 #' the scale parameter (standard deviation at log scale)
 #' 
 #' @examples 

R/lognormalDiff.R

@@ -1,6 +1,6 @@
 #' Inference on the difference of two lognormals
 #' 
-#' The distribtuion of y = a - b + s, where a and b are two lognormal random
+#' The distribution of y = a - b + s, where a and b are two lognormal random
 #' variables and s is a constant to be estimated, can be approximated
 #' by a lognormal distribution.
 #' 

README.Rmd

@@ -72,7 +72,7 @@ abline(v = getLognormMode(coefSum["mu"], coefSum["sigma"]), lty = "dashed")
 abline(v = getLognormMedian(coefSum["mu"], coefSum["sigma"]), lty = "dotdash")
 ```
 
-The sum of the expected values is conserved, while the mutliplicative standard 
+The sum of the expected values is conserved, while the multiplicative standard 
 deviation decreases during aggregation:
 
 ```{r}

README.md

@@ -58,7 +58,7 @@ abline(v = getLognormMedian(coefSum["mu"], coefSum["sigma"]), lty = "dotdash")
 
 ![](tools/README-example-1.png)
 
-The sum of the expected values is conserved, while the mutliplicative standard deviation decreases during aggregation:
+The sum of the expected values is conserved, while the multiplicative standard deviation decreases during aggregation:
 
 ``` r
 c( 

cran-comments.md

@@ -1,31 +1,17 @@
 ## Note
-Added support for computing moments and statistics.
-
-Changed doi format (brackets) in DESCRIPTION according to 
-comment from Uwe Ligges from 
-10.1641/0006-3568(2001)051[0341:lndats]2.0.co;2
-to
-10.1641/0006-3568(2001)051%5B0341:lndats%5D2.0.co;2
-
-And changed back to original doi after second comment 
-by Uwe Ligges
-
-Changed again to escaped form after email from Kurt Hornik.
-
-Removed link to md file in vignette. (The link worked nice on github but not
-for the installed package.)
+* better support for missing values
 
 ## Test environments
-with old doi
-* local R 3.4.4 on Mint17 64bit
-* Ubuntu 14.04.5 LTS (on travis-ci), 14.04
-* win-builder (devel)
-* R-Hub
-
-with updated doi and link:
-* local R 3.4.4 on Mint17 64bit
+* local R 4.0.4 on Mint21 64bit
+* R-hub
+** Ubuntu Linux 20.04.1 LTS, R-release, GCC
+** Fedora Linux, R-devel, clang, gfortran
+** windows-x86_64-devel (with env_vars=c(R_COMPILE_AND_INSTALL_PACKAGES = "always")
 
 ## R CMD check results
-Note on problem with doi
+* 1 Note on (possibly) invalid DOIs: 
+** After changing to escaped form after email from Kurt Hornik in a previous release.
+*** In contrast Uwe Ligges was in favour of the plain version. This went back and forth 
+    already several times
+** The plain correct version is: 10.1641/0006-3568(2001)051[0341:LNDATS]2.0.CO;2
 
-

inst/WORDLIST

@@ -0,0 +1,42 @@
+Limpert
+Zieba
+aggregateCorrelated
+al
+ams
+anc
+autocorrelated
+computeEffectiveNumObs
+doi
+effCor
+estimateDiffLognormal
+estimateSumLognormal
+et
+gapfilled
+getParmsLognormForLowerAndUpper
+getParmsLognormForLowerAndUpperLog
+ij
+inlinedocs
+isGapFilled
+isTransScale
+ln
+lndats
+logmean
+lognormalSum
+lognormally
+md
+nObs
+na
+nfin
+operatorname
+pDiffLognormalSample
+precomputed
+roxygen
+scaleLogToOrig
+sd
+seCor
+sigmaFac
+sigmaStar
+varCor
+μ
+ρ
+σ

tests/testthat/test_autoCorr.R

@@ -70,6 +70,7 @@ test_that("computeEffectiveNumObs",{
 
 test_that("computeEffectiveNumObs with NA",{
   # generate autocorrelated time series
+  set.seed(0815)
   res <- stats::filter(rnorm(1000), filter = rep(1,8), circular = TRUE)
   effAcfFull <- computeEffectiveAutoCorr(res)
   # tail NA
@@ -84,7 +85,7 @@ test_that("computeEffectiveNumObs with NA",{
   expect_true(nEff <=  sum(is.finite(res)))
   # positive bias not accounting for missings
   nEff3 <- computeEffectiveNumObs(res, na.rm = TRUE, exact.na = FALSE)
-  expect_true(nEff3 > nEff) 
+  expect_true(nEff3 > nEff) # may fail due to only 7 random numbers
   nEff <- computeEffectiveNumObs(res, na.rm = TRUE, effAcf = effAcfFull)
   expect_true(nEff <=  sum(is.finite(res)))
   expect_true(nEff >=  1)

vignettes/lognorm.Rmd

@@ -10,7 +10,7 @@ vignette: >
 ---
 ```{r eval=FALSE, include=FALSE}
 # twDev::genVigs()
-rmarkdown::render("lognorm.Rmd","md_document") 
+#rmarkdown::render("lognorm.Rmd","md_document") 
 ```
 
 ```{r setup, include=FALSE}
@@ -32,10 +32,10 @@ knit_hooks$set(spar = function(before, options, envir) {
 })
 library(lognorm) 
 if (!require(ggplot2) || !require(dplyr) || !require(purrr)) {
-	print("To generate this vignette, ggplo2, dplyr, and purrr are required.")
-	exit(0)
+	print("To generate this vignette, ggplot2, dplyr, and purrr are required.")
+	knit_exit()
 }
-themeTw <- theme_bw(base_size = 10) + 
+themeTw <- ggplot2::theme_bw(base_size = 10) + 
   theme(axis.title = element_text(size = 9))
 ```
 

vignettes/lognormalDiff.Rmd

@@ -9,7 +9,7 @@ vignette: >
 ---
 ```{r eval=FALSE, include=FALSE}
 # twDev::genVigs()
-rmarkdown::render("lognormalDiff.Rmd","md_document") 
+#rmarkdown::render("lognormalDiff.Rmd","md_document") 
 ```
 
 ```{r setup, include=FALSE}
@@ -31,10 +31,10 @@ knit_hooks$set(spar = function(before, options, envir) {
 })
 library(lognorm) 
 if (!require(ggplot2) || !require(dplyr) || !require(tidyr) || !require(purrr)) {
-	print("To generate this vignette, ggplo2, dplyr, tidyr, and purrr are required.")
-	exit(0)
+	print("To generate this vignette, ggplot2, dplyr, tidyr, and purrr are required.")
+	knit_exit()
 }
-themeTw <- theme_bw(base_size = 10) + 
+themeTw <- ggplot2::theme_bw(base_size = 10) + 
   theme(axis.title = element_text(size = 9))
 ```
 
@@ -108,12 +108,12 @@ theme(axis.title.x = element_blank())
 
 ## Test if difference is significantly different from zero
 
-The probability of the zero quantile needs to be larger than a significane 
+The probability of the zero quantile needs to be larger than a significance 
 level.
 We can compute it based on the lognormal approximation.
 
 Since Lo12 is only accurate if the expected difference is small compared
-to the expected sum, the probiblity of the difference being larger than zero can be estimated by a sampling both terms.
+to the expected sum, the probability of the difference being larger than zero can be estimated by a sampling both terms.
 
 ```{r}
   mu1 = log(120)

vignettes/lognormalDiff.md

@@ -99,12 +99,12 @@ A check by random numbers (dotted lines) shows close correspondence.
 
 ## Test if difference is significantly different from zero
 
-The probability of the zero quantile needs to be larger than a significane 
+The probability of the zero quantile needs to be larger than a significance 
 level.
 We can compute it based on the lognormal approximation.
 
 Since Lo12 is only accurate if the expected difference is small compared
-to the expected sum, the probiblity of the difference being larger than zero can be estimated by a sampling both terms.
+to the expected sum, the probability of the difference being larger than zero can be estimated by a sampling both terms.
 
 
 ```r

vignettes/lognormalSum.Rmd

@@ -9,7 +9,7 @@ vignette: >
 ---
 ```{r eval=FALSE, include=FALSE}
 # twDev::genVigs()
-rmarkdown::render("lognormalSum.Rmd","md_document") 
+#rmarkdown::render("lognormalSum.Rmd","md_document") 
 ```
 
 ```{r setup, include=FALSE}
@@ -31,10 +31,10 @@ knit_hooks$set(spar = function(before, options, envir) {
 })
 library(lognorm) 
 if (!require(ggplot2) || !require(dplyr) || !require(tidyr) || !require(purrr)) {
-	print("To generate this vignette, ggplo2, dplyr, tidyr, and purrr are required.")
-	exit(0)
+	print("To generate this vignette, ggplot2, dplyr, tidyr, and purrr are required.")
+	knit_exit()
 }
-themeTw <- theme_bw(base_size = 10) + 
+themeTw <- ggplot2::theme_bw(base_size = 10) + 
   theme(axis.title = element_text(size = 9))
 ```
 
@@ -172,7 +172,7 @@ coefSum <- estimateSumLognormal( theta[,1], theta[,2], effAcf = c(1,acf1) )
 setNames(exp(coefSum["sigma"]), "sigmaStar")
 ```
 
-Its expected value corresponds to the summ of expected values (100*10).
+Its expected value corresponds to the sum of expected values (100*10).
 
 ```{r}
 (sumExp <- getLognormMoments( coefSum[1], coefSum[2])[1,"mean"])
@@ -201,9 +201,9 @@ theme(axis.title.x = element_blank()) +
 theme(legend.title = element_blank())
 ```
 
-The mean is the sum devided by the number of observations, $n$.
-While the multiplicatie standard deviation does not change by this operation,
-the location parameter is obtained by deviding by $n$ at original scale, hence,
+The mean is the sum divided by the number of observations, $n$.
+While the multiplicative standard deviation does not change by this operation,
+the location parameter is obtained by dividing by $n$ at original scale, hence,
 subtracting $log(n)$ at log-scale.
 ```{r}
 (coefMean <- setNames(c(coefSum["mu"] - log(nObs), coefSum["sigma"]), c("mu","sigma")))

vignettes/lognormalSum.md

@@ -112,7 +112,7 @@ variables. The multiplicative uncertainty has decreased from 1.7.
     ## sigmaStar 
     ##  1.077687
 
-Its expected value corresponds to the summ of expected values (100\*10).
+Its expected value corresponds to the sum of expected values (100\*10).
 
     (sumExp <- getLognormMoments( coefSum[1], coefSum[2])[1,"mean"])
 
@@ -124,9 +124,9 @@ the distribution of the 10000 repetitions.
 
 <img src="lognormalSum_files/figure-markdown_strict/pdfSum100-1.png" style="display:block; margin: auto" />
 
-The mean is the sum devided by the number of observations, *n*. While
-the multiplicatie standard deviation does not change by this operation,
-the location parameter is obtained by deviding by *n* at original scale,
+The mean is the sum divided by the number of observations, *n*. While
+the multiplicative standard deviation does not change by this operation,
+the location parameter is obtained by dividing by *n* at original scale,
 hence, subtracting *l**o**g*(*n*) at log-scale.
 
     (coefMean <- setNames(c(coefSum["mu"] - log(nObs), coefSum["sigma"]), c("mu","sigma")))