change order of arguments for metric in gridsearch (#989)

* change order of arguments for metric in gridsearch

* add doctring in gridsearch

* add line break to docstring of gridsearch

Co-authored-by: Grotehans <Clara.Grotehans@EXXETA.com>
Co-authored-by: Julien Herzen <julien@unit8.co>

Author: 3 people

darts/models/forecasting/forecasting_model.py

@@ -680,7 +680,8 @@ def gridsearch(
             If `True`, uses the comparison with the fitted values.
             Raises an error if ``fitted_values`` is not an attribute of `model_class`.
         metric
-            A function that takes two TimeSeries instances as inputs and returns a float error value.
+            A function that takes two TimeSeries instances as inputs (actual and prediction, in this order),
+            and returns a float error value.
         reduction
             A reduction function (mapping array to float) describing how to aggregate the errors obtained
             on the different validation series when backtesting. By default it'll compute the mean of errors.
@@ -764,7 +765,7 @@ def _evaluate_combination(param_combination) -> float:
                 fitted_values = TimeSeries.from_times_and_values(
                     series.time_index, model.fitted_values
                 )
-                error = metric(fitted_values, series)
+                error = metric(series, fitted_values)
             elif val_series is None:  # expanding window mode
                 error = model.backtest(
                     series=series,
@@ -787,7 +788,7 @@ def _evaluate_combination(param_combination) -> float:
                     future_covariates,
                     num_samples=1,
                 )
-                error = metric(pred, val_series)
+                error = metric(val_series, pred)
 
             return float(error)