sample_to_quantile.numeric method

[seabbs]

In this code:

scoringutils/R/metrics-sample.R

Lines 305 to 306 in e43e008

	n <- length(observed)
	N <- length(predicted) / n

we convert from a sample format to a quantile format so that we can use the data.frame sample_to_quantile. It would be better if we instead had a helper function to do this.

Reproduced here:

  assert_input_sample(observed, predicted)
  assert_number(range)
  necessary_quantiles <- c((100 - range) / 2, 100 - (100 - range) / 2) / 100

  # this could be its own function, sample_to_quantile.numeric
  # ==========================================================
  n <- length(observed)
  N <- length(predicted) / n
  dt <- data.table(
    observed = rep(observed, each = N),
    predicted = as.vector(t(predicted))
  )
  quantile_dt <- sample_to_quantile(dt, necessary_quantiles)