def test_cutoff_window_splitter(y, cutoffs, fh, window_length):
"""Test CutoffSplitter."""
if _inputs_are_supported([cutoffs, fh, window_length]):
cv = CutoffSplitter(cutoffs, fh=fh, window_length=window_length)
train_windows, test_windows, cutoffs, n_splits = _check_cv(cv, y)
np.testing.assert_array_equal(cutoffs, cv.get_cutoffs(y))
else:
match = "Unsupported combination of types"
with pytest.raises(TypeError, match=match):
CutoffSplitter(cutoffs, fh=fh, window_length=window_length)
def test_cutoff_window_splitter(y, cutoffs, fh, window_length):
"""Test CutoffSplitter."""
cv = CutoffSplitter(cutoffs, fh=fh, window_length=window_length)
if _cutoffs_fh_window_length_types_are_supported(
cutoffs=cutoffs,
fh=ForecastingHorizon(fh),
window_length=window_length):
train_windows, test_windows, cutoffs, n_splits = _check_cv(cv, y)
np.testing.assert_array_equal(cutoffs, cv.get_cutoffs(y))
else:
match = "Unsupported combination of types"
with pytest.raises(TypeError, match=match):
_check_cv(cv, y)
def _predict_in_sample(self,
fh,
X=None,
return_pred_int=False,
alpha=DEFAULT_ALPHA):
"""Make in-sample prediction using single-step moving-cutoff
predictions
Parameters
----------
fh : np.array
all non-positive (<= 0)
X : pd.DataFrame
return_pred_int : bool
alpha : float or array-like
Returns
-------
y_pred : pd.DataFrame or pd.Series
"""
y_train = self._y
# generate cutoffs from forecasting horizon, note that cutoffs are
# still based on integer indexes,
# so that they can be used with .iloc
cutoffs = fh + len(y_train) - 2
cv = CutoffSplitter(cutoffs, fh=1, window_length=self.window_length_)
return self._predict_moving_cutoff(y_train,
cv,
X=X,
update_params=False,
return_pred_int=return_pred_int,
alpha=alpha)
def test_manual_window_split(y, cutoffs, fh, window_length):
# initiate rolling window cv iterator
cv = CutoffSplitter(cutoffs, fh=fh, window_length=window_length)
# generate and keep splits
training_windows, test_windows, n_splits, _ = generate_and_check_windows(
y, cv)
# check cutoffs
np.testing.assert_array_equal(cutoffs, cv.get_cutoffs(y))
# check training windows
n_incomplete_windows = get_n_incomplete_windows(training_windows,
window_length)
check_windows_dimensions(training_windows, n_incomplete_windows,
window_length)
# check test windows
check_test_windows(test_windows, fh, cutoffs)
def test_cutoff_window_splitter(y, cutoffs, fh, window_length):
"""Test CutoffSplitter."""
cv = CutoffSplitter(cutoffs, fh=fh, window_length=window_length)
train_windows, test_windows, cutoffs, n_splits = _check_cv(cv, y)
np.testing.assert_array_equal(cutoffs, cv.get_cutoffs(y))