use broom instead of generics

strengejacke · Dec 15, 2018 · 677c7de · 677c7de
1 parent 1c1731d
commit 677c7de
Show file tree

Hide file tree

Showing 5 changed files with 14 additions and 15 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -25,7 +25,7 @@ Depends:
     stats,
     utils
 Imports:
-    generics,
+    broom,
     dplyr (>= 0.7.5),
     forcats,
     ggeffects (>= 0.7.0),

diff --git a/NAMESPACE b/NAMESPACE
@@ -76,6 +76,8 @@ export(view_df)
 import(ggplot2)
 importFrom(MASS,glm.nb)
 importFrom(MASS,lda)
+importFrom(broom,augment)
+importFrom(broom,tidy)
 importFrom(dplyr,"%>%")
 importFrom(dplyr,arrange)
 importFrom(dplyr,between)
@@ -95,8 +97,6 @@ importFrom(dplyr,summarize)
 importFrom(dplyr,ungroup)
 importFrom(forcats,fct_reorder)
 importFrom(forcats,fct_rev)
-importFrom(generics,augment)
-importFrom(generics,tidy)
 importFrom(ggeffects,ggeffect)
 importFrom(ggeffects,ggpredict)
 importFrom(glmmTMB,fixef)

diff --git a/NEWS.md b/NEWS.md
@@ -3,7 +3,6 @@
 ## General
 
 * Revised some help-files and vignettes.
-* Replaced package dependency **broom** with **generics**.
 
 ## Removed / Defunct
 

diff --git a/R/sjPlotKfoldCV.R b/R/sjPlotKfoldCV.R
@@ -48,7 +48,7 @@
 #' @importFrom modelr crossv_kfold
 #' @importFrom dplyr mutate ungroup summarise
 #' @importFrom purrr map map2
-#' @importFrom generics augment
+#' @importFrom broom augment
 #' @importFrom tidyr unnest
 #' @importFrom graphics plot
 #' @importFrom stats as.formula formula family poisson glm lm
@@ -113,7 +113,7 @@ sjp.kfold_cv <- function(data, formula, k = 5, fit) {
     # train data
     res <- modelr::crossv_kfold(data, k = k) %>%
       dplyr::mutate(model = purrr::map(.data$train, ~ stats::lm(formula, data = .))) %>%
-      dplyr::mutate(predicted = purrr::map2(.data$model, .data$test, ~ generics::augment(.x, newdata = .y))) %>%
+      dplyr::mutate(predicted = purrr::map2(.data$model, .data$test, ~ broom::augment(.x, newdata = .y))) %>%
       tidyr::unnest(.data$predicted)
 
     # make sure that response vector has an identifiably name

diff --git a/R/tidiers.R b/R/tidiers.R
@@ -58,7 +58,7 @@ get_tidy_data <- function(model, ci.lvl, tf, type, bpe, facets, show.zeroinf, p.
 }
 
 
-#' @importFrom generics tidy
+#' @importFrom broom tidy
 #' @importFrom sjstats p_value
 #' @importFrom stats coef qnorm
 #' @importFrom dplyr mutate
@@ -77,7 +77,7 @@ tidy_generic <- function(model, ci.lvl, facets, p.val) {
   if (inherits(stats::coef(summary(model)), "listof")) {
 
     # get estimates, as data frame
-    dat <- generics::tidy(model, conf.int = FALSE, exponentiate = FALSE)
+    dat <- broom::tidy(model, conf.int = FALSE, exponentiate = FALSE)
 
     # check whether each category should be printed in facets, or
     # in a single graph (with dodged geoms)
@@ -94,7 +94,7 @@ tidy_generic <- function(model, ci.lvl, facets, p.val) {
     if (inherits(model, "lmerModLmerTest")) {
       dat <- tidy_lmerModLmerTest(model, ci.lvl)
     } else {
-      dat <- generics::tidy(model, conf.int = FALSE, effects = "fixed")
+      dat <- broom::tidy(model, conf.int = FALSE, effects = "fixed")
     }
 
 
@@ -185,11 +185,11 @@ tidy_svynb_model <- function(model, ci.lvl) {
 }
 
 
-#' @importFrom generics tidy
+#' @importFrom broom tidy
 #' @importFrom sjstats p_value
 tidy_cox_model <- function(model, ci.lvl) {
   # tidy the model
-  dat <- generics::tidy(model, conf.int = ci.lvl)
+  dat <- broom::tidy(model, conf.int = ci.lvl)
 
   # see if we have p-values. if not, add them
   if (!obj_has_name(dat, "p.value"))
@@ -527,14 +527,14 @@ tidy_stan_model <- function(model, ci.lvl, tf, type, bpe, show.zeroinf, facets,
 }
 
 
-#' @importFrom generics tidy
+#' @importFrom broom tidy
 #' @importFrom sjstats p_value
 #' @importFrom nlme intervals
 tidy_lme_model <- function(model, ci.lvl) {
   # get tidy summary. for lme, this excludes CI,
   # so we compute them separately
 
-  dat <- generics::tidy(model, conf.int = TRUE, conf.level = ci.lvl, effects = "fixed")
+  dat <- broom::tidy(model, conf.int = TRUE, conf.level = ci.lvl, effects = "fixed")
   ci <- as.data.frame(nlme::intervals(model, level = ci.lvl, which = "fixed")$fixed)
 
   dat$conf.low <- ci$lower
@@ -798,7 +798,7 @@ tidy_polr_model <- function(model, ci.lvl) {
 #' @importFrom stats qnorm pnorm
 #' @importFrom rlang .data
 #' @importFrom dplyr mutate
-#' @importFrom generics tidy
+#' @importFrom broom tidy
 #' @importFrom sjmisc var_rename
 tidy_multinom_model <- function(model, ci.lvl, facets) {
 
@@ -811,7 +811,7 @@ tidy_multinom_model <- function(model, ci.lvl, facets) {
 
 
   # get estimates, as data frame
-  dat <- generics::tidy(model, conf.int = FALSE, exponentiate = FALSE)
+  dat <- broom::tidy(model, conf.int = FALSE, exponentiate = FALSE)
 
 
   # add conf. int.