Say whether calls were right — rate • functai

"Right" means correct for this input, not "nice". A correction becomes a row of data: rated() gives it back with the inputs. Vectorised over call, verdict and answer, so a column of call ids from predict() or augment() can be rated at once.

Usage

rate(
  call,
  verdict = NULL,
  answer = NULL,
  note = NULL,
  reasons = NULL,
  by = NULL,
  origin = NULL,
  sample = NULL,
  folder = NULL
)

Arguments

call: Call ids (the .call column of predict()/augment()).
verdict: "right", "wrong", TRUE, FALSE, or NA to withdraw your earlier rating. Default: "wrong" when answer is given.
answer: The right answers, for "wrong".
note, reasons: Why, in a sentence, and short tags.
by: Who is judging: a person (default: the caller's user). With no person named, the rating is made under this computer's account, which may be shared: it is kept on its own, and never replaces nor is replaced by another rating.
origin: "review" (someone judged it) or "edit" (someone changed the output while using it).
sample: The id of a random draw of calls these ratings belong to.
folder: The log folder (default: where calls are logged here).

Value

The ratings as written, invisibly (a tibble).