Remove distributional optional dependency workaround

.github/workflows/R-CMD-check.yaml

@@ -42,7 +42,7 @@ jobs:
 
       - uses: r-lib/actions/setup-r-dependencies@v2
         with:
-          extra-packages: any::rcmdcheck, ggdist=?ignore-before-r=4.1.0
+          extra-packages: any::rcmdcheck
           needs: check
 
       - uses: r-lib/actions/check-r-package@v2

NEWS.md

@@ -2,6 +2,7 @@
 
 - Added support for scoring multivariate forecasts (#288, big thank you to Sam Abbott and Sebastian Funk). You can find detailed information in the Vignette "Scoring multivariate forecasts". There is a new forecast type, `forecast_multivariate_sample()` and a corresponding `as_forecast_multivariate_sample()` function. To score a multivariate forecast, users are expected to provide a `joint_across` argument which specifies the variables which are forecast jointly.
 - Fixed a small bug with `bias_sample()` when continuous predictions were equal to observations. These ties could lead to incorrect bias scores. New calculations now use mid-ranks to deal with ties.
+- Removed workaround for `{distributional}` native pipe issue which has now been fixed upstream (#947).
 
 # scoringutils 2.1.2
 

vignettes/Deprecated-visualisations.Rmd

@@ -26,6 +26,7 @@ The example data used is the data shipped with the `scoringutils` package. It co
 library(scoringutils)
 library(data.table)
 library(ggplot2)
+library(ggdist)
 library(magrittr)
 library(magrittr) #pipe operator
 ```
@@ -34,7 +35,7 @@ library(magrittr) #pipe operator
 
 In previous versions of `scoringutils`, forecasts and observed values could be visualised using the function `plot_predictions()` and its `make_na()` helper function. The following shows the function code first and then an example. 
 
-```{r eval=require("ggdist", quietly = TRUE)}
+```{r}
 #" @title Plot Predictions vs True Values
 #"
 #" @description
@@ -98,7 +99,6 @@ In previous versions of `scoringutils`, forecasts and observed values could be v
 #"   aes(fill = model, color = model)
 
 
-library(ggdist)
 plot_predictions <- function(data,
                              by = NULL,
                              x = "date",
@@ -278,7 +278,7 @@ In the following are a few examples of using the two functions to create a plot
 
 Visualising the median forecasts for the example data. The truth data is restricted to a period between 2021-05-01 and 2021-07-22. The forecast data is a forecast from the model "EuroCOVIDhub-ensemble" made on the "2021-06-07". All other data is set to `NA`, effectively removing it from the plot. 
 
-```{r eval=require("ggdist", quietly = TRUE)}
+```{r}
 median_forecasts <- example_quantile[quantile_level == 0.5]
 median_forecasts %>%
   make_NA(what = "truth",
@@ -296,7 +296,7 @@ median_forecasts %>%
 
 This is the same plot, but with a variety of prediction intervals shown, instead of just the median. 
 
-```{r eval=require("ggdist", quiet = TRUE)}
+```{r}
 example_quantile %>%
   make_NA(what = "truth",
           target_end_date <= "2021-05-01",
@@ -314,7 +314,7 @@ example_quantile %>%
 
 And a similar plot, this time based on continuous forecasts. The predictions are automatically converted to a quantile-based forecasts for plotting.
 
-```{r eval=require("ggdist", quietly = TRUE)}
+```{r}
 example_sample_continuous %>%
   make_NA(what = "truth",
           target_end_date <= "2021-05-01",
@@ -332,7 +332,7 @@ example_sample_continuous %>%
 
 Displaying two forecasts at a time with additional colours: 
 
-```{r eval=require("ggdist", quietly = TRUE)}
+```{r}
 example_quantile %>%
   make_NA(what = "truth",
           target_end_date > "2021-07-15",