Skip to content

BOT: Fix #825: Remove Metrics package from Imports #1084

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
nikosbosse wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

BOT: Fix #825: Remove Metrics package from Imports #1084

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion DESCRIPTION
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,14 +59,14 @@ Imports:
data.table (>= 1.16.0),
ggplot2 (>= 3.4.0),
methods,
Metrics,
purrr,
scoringRules (>= 1.1.3),
stats
Suggests:
ggdist,
kableExtra,
knitr,
Metrics,
rmarkdown,
testthat (>= 3.1.9),
vdiffr
Expand Down
3 changes: 0 additions & 3 deletions NAMESPACE
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,9 +109,6 @@ export(transform_forecasts)
export(underprediction_quantile)
export(underprediction_sample)
export(wis)
importFrom(Metrics,ae)
importFrom(Metrics,ape)
importFrom(Metrics,se)
importFrom(checkmate,assert)
importFrom(checkmate,assert_character)
importFrom(checkmate,assert_class)
Expand Down
13 changes: 6 additions & 7 deletions R/class-forecast-point.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,6 @@ is_forecast_point <- function(x) {
}


#' @importFrom Metrics se ae ape
#' @importFrom stats na.omit
#' @importFrom data.table setattr copy
#' @rdname score
Expand All @@ -100,9 +99,9 @@ score.forecast_point <- function(forecast, metrics = get_metrics(forecast), ...)
#'
#' @description
#' For point forecasts, the default scoring rules are:
#' - "ae_point" = [ae()][Metrics::ae()]
#' - "se_point" = [se()][Metrics::se()]
#' - "ape" = [ape()][Metrics::ape()]
#' - "ae_point" = absolute error
#' - "se_point" = squared error
#' - "ape" = absolute percentage error
#'
#' A note of caution: Every scoring rule for a point forecast
#' is implicitly minimised by a specific aspect of the predictive distribution
Expand Down Expand Up @@ -145,9 +144,9 @@ score.forecast_point <- function(forecast, metrics = get_metrics(forecast), ...)
#' Journal of the American Statistical Association.
get_metrics.forecast_point <- function(x, select = NULL, exclude = NULL, ...) {
all <- list(
ae_point = Metrics::ae,
se_point = Metrics::se,
ape = Metrics::ape
ae_point = function(actual, predicted) abs(actual - predicted),
se_point = function(actual, predicted) (actual - predicted)^2,
ape = function(actual, predicted) abs(actual - predicted) / abs(actual)
)
select_metrics(all, select, exclude)
}
Expand Down
6 changes: 3 additions & 3 deletions man/get_metrics.forecast_point.Rd
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

50 changes: 50 additions & 0 deletions tests/testthat/test-class-forecast-point.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -107,3 +107,53 @@ test_that("get_metrics.forecast_point() works as expected", {
c("ae_point", "se_point", "ape")
)
})

test_that("internal ae replacement produces identical results", {
observed <- c(1, -15, 22, 0, 5.5)
predicted <- c(5, 6, 7, 0, 5.5)
ae_fn <- get_metrics(example_point, select = "ae_point")[[1]]
expect_identical(ae_fn(observed, predicted), abs(observed - predicted))
expect_identical(ae_fn(5, 5), 0)
expect_identical(ae_fn(-10, 5), 15)
})

test_that("internal se replacement produces identical results", {
observed <- c(1, -15, 22, 0, 5.5)
predicted <- c(5, 6, 7, 0, 5.5)
se_fn <- get_metrics(example_point, select = "se_point")[[1]]
expect_identical(se_fn(observed, predicted), (observed - predicted)^2)
expect_identical(se_fn(5, 5), 0)
expect_identical(se_fn(-10, 5), 225)
})

test_that("internal ape replacement produces identical results", {
observed <- c(1, -15, 22, 5.5, 100)
predicted <- c(5, 6, 7, 0, 100)
ape_fn <- get_metrics(example_point, select = "ape")[[1]]
expect_equal(ape_fn(observed, predicted), abs(observed - predicted) / abs(observed))
expect_identical(ape_fn(5, 5), 0)
expect_identical(ape_fn(0, 5), Inf)
})

test_that("Metrics package is not in DESCRIPTION Imports", {
desc_text <- readLines(system.file("DESCRIPTION", package = "scoringutils"))
imports_lines <- desc_text[grepl("^Imports:|^\\s+Metrics", desc_text)]
expect_false(any(grepl("\\bMetrics\\b", imports_lines)))
})

test_that("score() with point forecasts produces correct results after Metrics removal", {
scores <- score(example_point)
input <- na.omit(as.data.table(example_point))
expect_equal(scores$ae_point, abs(input$observed - input$predicted))
expect_equal(scores$se_point, (input$observed - input$predicted)^2)
expect_equal(scores$ape, abs(input$observed - input$predicted) / abs(input$observed))
expect_true(all(c("ae_point", "se_point", "ape") %in% colnames(scores)))
})

test_that("get_metrics.forecast_point() returns expected functions", {
metrics <- get_metrics(example_point)
expect_type(metrics, "list")
expect_named(metrics, c("ae_point", "se_point", "ape"))
expect_true(all(vapply(metrics, is.function, logical(1))))
expect_true(all(vapply(metrics, function(f) length(formals(f)) == 2, logical(1))))
})
14 changes: 7 additions & 7 deletions vignettes/scoring-rules.Rmd
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,11 +58,11 @@ observed <- rnorm(n, 5, 4)^2
predicted_mu <- mean(observed)
predicted_not_mu <- predicted_mu - rnorm(n, 10, 2)

mean(Metrics::ae(observed, predicted_mu))
mean(Metrics::ae(observed, predicted_not_mu))
mean(abs(observed - predicted_mu))
mean(abs(observed - predicted_not_mu))

mean(Metrics::se(observed, predicted_mu))
mean(Metrics::se(observed, predicted_not_mu))
mean((observed - predicted_mu)^2)
mean((observed - predicted_not_mu)^2)
```


Expand All @@ -73,7 +73,7 @@ mean(Metrics::se(observed, predicted_not_mu))

**Forecast**: $\hat{y}$, a real number, the median of the forecaster's predictive distribution.

The absolute error is the absolute difference between the predicted and the observed values. See `?Metrics::ae`.
The absolute error is the absolute difference between the predicted and the observed values.

$$\text{ae} = |y - \hat{y}|$$

Expand All @@ -85,7 +85,7 @@ The absolute error is only an appropriate rule if $\hat{y}$ corresponds to the m

**Forecast**: $\hat{y}$, a real number, the mean of the forecaster's predictive distribution.

The squared error is the squared difference between the predicted and the observed values. See `?Metrics::se`.
The squared error is the squared difference between the predicted and the observed values.

$$\text{se} = (y - \hat{y})^2$$
The squared error is only an appropriate rule if $\hat{y}$ corresponds to the mean of the forecaster's predictive distribution. Otherwise, results will be misleading (see @gneitingMakingEvaluatingPoint2011).
Expand All @@ -96,7 +96,7 @@ The squared error is only an appropriate rule if $\hat{y}$ corresponds to the me

**Forecast**: $\hat{y}$, a real number

The absolute percentage error is the absolute percent difference between the predicted and the observed values. See `?Metrics::ape`.
The absolute percentage error is the absolute percent difference between the predicted and the observed values.

$$\text{ape} = \frac{|y - \hat{y}|}{|y|}$$

Expand Down