fixing NOTEs and WARNINGs from CRAN

Author: pbiecek

DALEX.Rproj

@@ -1,4 +1,5 @@
 Version: 1.0
+ProjectId: 4d8b6bf2-cb61-4abb-a60f-33d25cb458e9
 
 RestoreWorkspace: No
 SaveWorkspace: No

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: DALEX
 Title: moDel Agnostic Language for Exploration and eXplanation
-Version: 2.5.1
+Version: 2.5.2
 Authors@R: c(person("Przemyslaw", "Biecek", email = "[email protected]", role = c("aut", "cre"), 
              comment = c(ORCID = "0000-0001-8423-1823")),
     person("Szymon", "Maksymiuk", role = "aut",
@@ -22,7 +22,7 @@ Description: Any unverified black box model is the path to failure. Opaqueness l
 License: GPL
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.2
 Depends: R (>= 3.5)
 Imports: 
   ggplot2,

NAMESPACE

@@ -17,7 +17,17 @@ S3method(plot,list)
 S3method(plot,model_diagnostics)
 S3method(plot,model_parts)
 S3method(plot,model_performance)
+S3method(plot,model_performance_boxplot)
+S3method(plot,model_performance_ecdf)
+S3method(plot,model_performance_gain)
+S3method(plot,model_performance_histogram)
+S3method(plot,model_performance_lift)
+S3method(plot,model_performance_prc)
+S3method(plot,model_performance_roc)
 S3method(plot,model_profile)
+S3method(plot,model_profile_aggregates)
+S3method(plot,model_profile_points)
+S3method(plot,model_profile_profiles)
 S3method(plot,predict_diagnostics)
 S3method(plot,predict_parts)
 S3method(plot,predict_profile)

NEWS.md

@@ -1,3 +1,7 @@
+DALEX 2.5.2
+---------------------------------------------------------------
+* fixed NOTE during `devtools::document()` such as `S3 method plot.... needs @export or @exportS3method tag.`
+
 DALEX 2.5.1
 ---------------------------------------------------------------
 * adding the support for calculating kernel SHAP values via `predict_parts()` function

R/data_happiness.R

@@ -1,7 +1,7 @@
 #' World Happiness Report data
 #'
 #' The \code{happiness_train} and \code{happiness_test} datasets are generated
-#' based on the "World Happiness Report" at Kaggle \url{https://www.kaggle.com/datasets/unsdsn/world-happiness}.
+#' based on the "World Happiness Report".
 #'
 #' It contains data for 781 countries and 7 variables. These are:
 #' \itemize{
@@ -19,7 +19,7 @@
 #' @name happiness
 #' @aliases happiness_test happiness_train
 #'
-#' @source World Happiness Report data \url{https://worldhappiness.report/}
+#' @source World Happiness Report data
 #' @usage
 #' data(happiness_train)
 #' data(happiness_test)

R/data_titanic.R

@@ -42,7 +42,7 @@
 #' @source This dataset was copied from the \code{stablelearner} package and went through few variable
 #' transformations. The complete list of persons on the RMS titanic was downloaded from
 #' \url{https://www.encyclopedia-titanica.org} on April 5, 2016. The  information given
-#' in \code{sibsp} and \code{parch} was adopoted from a data set obtained from \url{https://biostat.app.vumc.org/wiki/Main/DataSets}.
+#' in \code{sibsp} and \code{parch} was adopoted from a data set obtained from \code{Kaggle}.
 #' @usage
 #' data(titanic)
 #' data(titanic_imputed)

R/plot_model_parts.R

@@ -9,12 +9,12 @@
 #'
 #' \subsection{variable_importance}{
 #' \itemize{
-#'  \item{\code{max_vars}}{ maximal number of features to be included in the plot. default value is \code{10}}
-#'  \item{\code{show_boxplots}}{ logical if \code{TRUE} (default) boxplot will be plotted to show permutation data.}
-#'  \item{\code{bar_width}}{ width of bars. By default \code{10}}
-#'  \item{\code{desc_sorting}}{ logical. Should the bars be sorted descending? By default \code{TRUE}}
-#'  \item{\code{title}}{ the plot's title, by default \code{'Feature Importance'}}
-#'  \item{\code{subtitle}}{ a character. Plot subtitle. By default \code{NULL} - then subtitle is set to "created for the XXX, YYY model",
+#'  \item{\code{max_vars} -- maximal number of features to be included in the plot. default value is \code{10}}
+#'  \item{\code{show_boxplots} -- logical if \code{TRUE} (default) boxplot will be plotted to show permutation data.}
+#'  \item{\code{bar_width} --  width of bars. By default \code{10}}
+#'  \item{\code{desc_sorting} --  logical. Should the bars be sorted descending? By default \code{TRUE}}
+#'  \item{\code{title} --  the plot's title, by default \code{'Feature Importance'}}
+#'  \item{\code{subtitle} --  a character. Plot subtitle. By default \code{NULL} - then subtitle is set to "created for the XXX, YYY model",
 #'        where XXX, YYY are labels of given explainers.}
 #' }
 #' }

R/plot_model_performance.R

@@ -54,6 +54,7 @@
 #' plot(mp_ranger, mp_glm, mp_lm, geom = "boxplot", show_outliers = 1)
 #' }
 #'
+#' @export
 #
 plot.model_performance <- function(x, ..., geom = "ecdf", show_outliers = 0, ptlabel = "name", lossFunction = loss_function, loss_function = function(x) sqrt(mean(x^2))) {
   if (!(ptlabel %in% c("name", "index"))){
@@ -96,10 +97,10 @@ plot.model_performance <- function(x, ..., geom = "ecdf", show_outliers = 0, ptl
   )
 }
 
-
-plot.model_performance_ecdf <- function(df, nlabels) {
+#' @export
+plot.model_performance_ecdf <- function(x, nlabels, ...) {
   label <- name <- NULL
-  ggplot(df, aes(abs(diff), color = label)) +
+  ggplot(x, aes(abs(diff), color = label)) +
     stat_ecdf(geom = "step") +
     theme_default_dalex() +
     scale_color_manual(name = "Model", values = colors_discrete_drwhy(nlabels)) +
@@ -111,23 +112,24 @@ plot.model_performance_ecdf <- function(df, nlabels) {
     ggtitle(expression(paste("Reverse cumulative distribution of ", group("|", residual, "|"))))
 }
 
-plot.model_performance_boxplot <- function(df, show_outliers, loss_function, nlabels) {
+#' @export
+plot.model_performance_boxplot <- function(x, show_outliers, loss_function, nlabels, ...) {
   label <- name <- NULL
-  pl <- ggplot(df, aes(x = label, y = abs(diff), fill = label)) +
+  pl <- ggplot(x, aes(x = label, y = abs(diff), fill = label)) +
     stat_boxplot(alpha = 0.4, coef = 1000) +
     stat_summary(fun = loss_function, geom = "point", shape = 20, size=10, color="red", fill="red") +
     theme_vertical_default_dalex() +
     scale_fill_manual(name = "Model", values = colors_discrete_drwhy(nlabels)) +
     ylab("") +
-    scale_x_discrete("", limits = rev(levels(df$label))) + # added to fix https://github.com/ModelOriented/DALEX/issues/400
+    scale_x_discrete("", limits = rev(levels(x$label))) + # added to fix https://github.com/ModelOriented/DALEX/issues/400
     ggtitle(
       expression(paste("Boxplots of ", group("|", residual, "|"))),
       "Red dot stands for root mean square of residuals"
     ) +
     coord_flip()
   if (show_outliers > 0) {
-    df$rank <- unlist(tapply(-abs(df$diff), df$label, rank, ties.method = "min"))
-    df_small <- df[df$rank <= show_outliers,]
+    x$rank <- unlist(tapply(-abs(x$diff), x$label, rank, ties.method = "min"))
+    df_small <- x[x$rank <= show_outliers,]
     pl <- pl +
       geom_point(data = df_small) +
       geom_text(data = df_small,
@@ -137,13 +139,14 @@ plot.model_performance_boxplot <- function(df, show_outliers, loss_function, nla
   pl
 }
 
-plot.model_performance_histogram <- function(df, nlabels) {
+#' @export
+plot.model_performance_histogram <- function(x, nlabels, ...) {
   diff <- label <- NULL
   # commented to keep it consistent with other plots
   # see: https://github.com/ModelOriented/DALEX/issues/400
   # if (length(levels(df$label)) > 1) levels(df$label) <- rev(levels(df$label))
 
-  ggplot(df, aes(diff, fill = label)) +
+  ggplot(x, aes(diff, fill = label)) +
     geom_histogram(bins = 100) +
     facet_wrap(~label, ncol = 1) +
     theme_default_dalex() + xlab("residuals") + theme(legend.position = "none") +
@@ -153,8 +156,9 @@ plot.model_performance_histogram <- function(df, nlabels) {
 }
 
 # precision-recall curve
-plot.model_performance_prc <- function(df, nlabels) {
-  dfl <- split(df, factor(df$label))
+#' @export
+plot.model_performance_prc <- function(x, nlabels, ...) {
+  dfl <- split(x, factor(x$label))
   prcdfl <- lapply(dfl, function(df) {
     pred_sorted <- df[order(df$predicted, decreasing = TRUE), ]
 
@@ -178,8 +182,9 @@ plot.model_performance_prc <- function(df, nlabels) {
 }
 
 # receiver operating characteristic
-plot.model_performance_roc <- function(df, nlabels) {
-  dfl <- split(df, factor(df$label))
+#' @export
+plot.model_performance_roc <- function(x, nlabels, ...) {
+  dfl <- split(x, factor(x$label))
   rocdfl <- lapply(dfl, function(df) {
     # assuming that y = 0/1 where 1 is the positive
     tpr_tmp <- tapply(df$observed, df$predicted, sum)
@@ -203,8 +208,9 @@ plot.model_performance_roc <- function(df, nlabels) {
     ggtitle("Receiver Operator Characteristic")
 }
 
-plot.model_performance_gain <- function(df, nlabels) {
-  dfl <- split(df, factor(df$label))
+#' @export
+plot.model_performance_gain <- function(x, nlabels, ...) {
+  dfl <- split(x, factor(x$label))
   rocdfl <- lapply(dfl, function(df) {
     pred_sorted <- df[order(df$predicted, decreasing = TRUE), ]
 
@@ -214,7 +220,7 @@ plot.model_performance_gain <- function(df, nlabels) {
     data.frame(lift = lift, pr = pr, label = df$label[1])
   })
   rocdf <- do.call(rbind, rocdfl)
-  max_lift <- sum(df$observed)/nrow(df)
+  max_lift <- sum(x$observed)/nrow(x)
 
   pr <- lift <- label <- NULL
   ggplot(rocdf, aes(x = pr, y = lift, color = label)) +
@@ -228,9 +234,9 @@ plot.model_performance_gain <- function(df, nlabels) {
 
 }
 
-
-plot.model_performance_lift <- function(df, nlabels) {
-  dfl <- split(df, factor(df$label))
+#' @export
+plot.model_performance_lift <- function(x, nlabels, ...) {
+  dfl <- split(x, factor(x$label))
   rocdfl <- lapply(dfl, function(df) {
     pred_sorted <- df[order(df$predicted, decreasing = TRUE), ]
 
@@ -240,7 +246,7 @@ plot.model_performance_lift <- function(df, nlabels) {
     data.frame(lift = lift/pr, pr = pr, label = df$label[1])
   })
   rocdf <- do.call(rbind, rocdfl)
-  max_lift <- sum(df$observed)/nrow(df)
+  max_lift <- sum(x$observed)/nrow(x)
 
   pr <- lift <- label <- NULL
   ggplot(rocdf, aes(x = pr, y = lift/max_lift, color = label)) +

R/plot_model_profile.R

@@ -8,14 +8,14 @@
 #'
 #' \subsection{aggregates}{
 #' \itemize{
-#'  \item{\code{color}}{ a character. Either name of a color, or hex code for a color,
+#'  \item{\code{color} --  a character. Either name of a color, or hex code for a color,
 #'   or \code{_label_} if models shall be colored, or \code{_ids_} if instances shall be colored}
-#'  \item{\code{size}}{ a numeric. Size of lines to be plotted}
-#'  \item{\code{alpha}}{ a numeric between \code{0} and \code{1}. Opacity of lines}
-#'  \item{\code{facet_ncol}}{ number of columns for the \code{\link[ggplot2]{facet_wrap}}}
-#'  \item{\code{variables}}{ if not \code{NULL} then only \code{variables} will be presented}
-#'  \item{\code{title}}{ a character. Partial and accumulated dependence explainers have deafult value.}
-#'  \item{\code{subtitle}}{ a character. If \code{NULL} value will be dependent on model usage.}
+#'  \item{\code{size} --  a numeric. Size of lines to be plotted}
+#'  \item{\code{alpha} --  a numeric between \code{0} and \code{1}. Opacity of lines}
+#'  \item{\code{facet_ncol} --  number of columns for the \code{\link[ggplot2]{facet_wrap}}}
+#'  \item{\code{variables} --  if not \code{NULL} then only \code{variables} will be presented}
+#'  \item{\code{title} --  a character. Partial and accumulated dependence explainers have deafult value.}
+#'  \item{\code{subtitle} --  a character. If \code{NULL} value will be dependent on model usage.}
 #' }
 #' }
 #'
@@ -63,6 +63,7 @@ plot.model_profile <- function(x, ..., geom = "aggregates") {
   )
 }
 
+#' @export
 plot.model_profile_aggregates <- function(x, ...) {
 #  plot(x$agr_profiles, ..., color = x$color)
   # fix for https://github.com/ModelOriented/DALEX/issues/237
@@ -73,11 +74,13 @@ plot.model_profile_aggregates <- function(x, ...) {
   do.call(plot, tmp)
 }
 
+#' @export
 plot.model_profile_profiles <- function(x, ...) {
   plot(x$cp_profiles, ..., size = 0.5, color = "grey") +
     ingredients::show_aggregated_profiles(x$agr_profiles, ..., size = 2, color = x$color)
 }
 
+#' @export
 plot.model_profile_points <- function(x, ...) {
   plot(x$cp_profiles, ..., color = "grey", size = 0.5) +
     ingredients::show_aggregated_profiles(x$agr_profiles, ..., size = 2, color = x$color) +

R/plot_predict_parts.R

@@ -9,35 +9,35 @@
 #'
 #' \subsection{break_down}{
 #' \itemize{
-#'  \item{\code{max_features}}{ maximal number of features to be included in the plot. default value is \code{10}}
-#'  \item{\code{min_max}}{ a range of OX axis. By default \code{NA}, therefore it will be extracted from the contributions of \code{x}.
+#'  \item{\code{max_features} --  maximal number of features to be included in the plot. default value is \code{10}}
+#'  \item{\code{min_max} --  a range of OX axis. By default \code{NA}, therefore it will be extracted from the contributions of \code{x}.
 #'    But it can be set to some constants, useful if these plots are to be used for comparisons.}
-#'  \item{\code{add_contributions}}{ if \code{TRUE}, variable contributions will be added to the plot.}
-#'  \item{\code{shift_contributions}}{ number describing how much labels should be shifted to the right, as a fraction of range. By default equal to \code{0.05}.}
-#'  \item{\code{vcolors}}{ If \code{NA} (default), DrWhy colors are used.}
-#'  \item{\code{vnames}}{ a character vector, if specified then will be used as labels on OY axis. By default \code{NULL}.}
-#'  \item{\code{digits}}{ number of decimal places (\code{\link{round}}) or significant digits (\code{\link{signif}}) to be used.}
-#'  \item{\code{rounding_function}}{ a function to be used for rounding numbers.}
-#'  \item{\code{plot_distributions}}{ if \code{TRUE} then distributions of conditional propotions will be plotted. This requires \code{keep_distributions=TRUE} in the
-#'    \code{\link{break_down}}, \code{\link{local_attributions}}, or \code{\link{local_interactions}}.}
-#'  \item{\code{baseline}}{ if numeric then veritical line starts in \code{baseline}.}
-#'  \item{\code{title}}{ a character. Plot title. By default \code{"Break Down profile"}.}
-#'  \item{\code{subtitle}}{ a character. Plot subtitle. By default \code{NULL} - then subtitle is set to "created for the XXX, YYY model",
+#'  \item{\code{add_contributions} --  if \code{TRUE}, variable contributions will be added to the plot.}
+#'  \item{\code{shift_contributions} --  number describing how much labels should be shifted to the right, as a fraction of range. By default equal to \code{0.05}.}
+#'  \item{\code{vcolors} --  If \code{NA} (default), DrWhy colors are used.}
+#'  \item{\code{vnames} --  a character vector, if specified then will be used as labels on OY axis. By default \code{NULL}.}
+#'  \item{\code{digits} --  number of decimal places (\code{\link{round}}) or significant digits (\code{\link{signif}}) to be used.}
+#'  \item{\code{rounding_function} --  a function to be used for rounding numbers.}
+#'  \item{\code{plot_distributions} --  if \code{TRUE} then distributions of conditional propotions will be plotted. This requires \code{keep_distributions=TRUE} in the
+#'    \code{break_down}, \code{local_attributions}, or \code{local_interactions}.}
+#'  \item{\code{baseline} --  if numeric then veritical line starts in \code{baseline}.}
+#'  \item{\code{title} --  a character. Plot title. By default \code{"Break Down profile"}.}
+#'  \item{\code{subtitle} --  a character. Plot subtitle. By default \code{NULL} - then subtitle is set to "created for the XXX, YYY model",
 #'        where XXX, YYY are labels of given explainers.}
-#'  \item{\code{max_vars}}{ alias for the \code{max_features} parameter.}
+#'  \item{\code{max_vars} --  alias for the \code{max_features} parameter.}
 #' }
 #' }
 #' \subsection{shap}{
 #' \itemize{
-#'  \item{\code{show_boxplots}}{ logical if \code{TRUE} (default) boxplot will be plotted to show uncertanity of attributions.}
-#'  \item{\code{vcolors}}{ If \code{NA} (default), DrWhy colors are used.}
-#'  \item{\code{max_features}}{ maximal number of features to be included in the plot. default value is \code{10}}
-#'  \item{\code{max_vars}}{ alias for the \code{max_features} parameter.}
+#'  \item{\code{show_boxplots} -- logical if \code{TRUE} (default) boxplot will be plotted to show uncertanity of attributions.}
+#'  \item{\code{vcolors} --  If \code{NA} (default), DrWhy colors are used.}
+#'  \item{\code{max_features} --  maximal number of features to be included in the plot. default value is \code{10}}
+#'  \item{\code{max_vars} --  alias for the \code{max_features} parameter.}
 #' }
 #' }
 #' \subsection{oscillations}{
 #' \itemize{
-#'  \item{\code{bar_width}}{ width of bars. By default \code{10}}
+#'  \item{\code{bar_width} --  width of bars. By default \code{10}}
 #' }
 #' }
 #'