[BUG] Forecasting regressor fix and notebook (#2885)

* stop returning an array

* notebook

* Create window.png

* fix test and typo

* remove unfinished sentence

Author: TonyBagnall

aeon/classification/compose/_pipeline.py

@@ -40,7 +40,7 @@ class ClassifierPipeline(BaseCollectionPipeline, BaseClassifier):
         A transform or list of transformers to use prior to classification.
         List of tuples (str, transformer) of transformers can also be passed, where
         the str is used to name the transformer.
-        The objecst are cloned prior, as such the state of the input will not be
+        The objects are cloned prior, as such the state of the input will not be
         modified by fitting the pipeline.
     estimator : aeon or sklearn classifier
         A classifier to use at the end of the pipeline.

aeon/forecasting/_regression.py

@@ -66,6 +66,10 @@ def _fit(self, y, exog=None):
         else:
             self.regressor_ = self.regressor
         y = y.squeeze()
+        if self.window < 1 or self.window > len(y) - 3:
+            raise ValueError(
+                f" window value {self.window} is invalid for series " f"length {len(y)}"
+            )
         X = np.lib.stride_tricks.sliding_window_view(y, window_shape=self.window)
         # Ignore the final horizon values: need to store these for pred with empty y
         X = X[: -self.horizon]
@@ -91,35 +95,13 @@ def _predict(self, y=None, exog=None):
 
         Returns
         -------
-        np.ndarray
+        float
             single prediction self.horizon steps ahead of y.
         """
         if y is None:
-            return self.regressor_.predict(self.last_)
+            return self.regressor_.predict(self.last_)[0]
         last = y[:, -self.window :]
-        return self.regressor_.predict(last)
-
-    def _forecast(self, y, exog=None):
-        """
-        Forecast the next horizon steps ahead.
-
-        Parameters
-        ----------
-        y : np.ndarray
-            A time series to predict the next horizon value for.
-        exog : np.ndarray, default=None
-            Optional exogenous time series data. Included for interface
-            compatibility but ignored in this estimator.
-
-        Returns
-        -------
-        np.ndarray
-            single prediction self.horizon steps ahead of y.
-
-        NOTE: deal with horizons
-        """
-        self.fit(y, exog)
-        return self.predict()
+        return self.regressor_.predict(last)[0]
 
     @classmethod
     def _get_test_params(cls, parameter_set: str = "default"):

aeon/forecasting/base.py

@@ -111,7 +111,8 @@ def predict(self, y=None, exog=None):
             y = self._convert_y(y, self.axis)
         if exog is not None:
             raise NotImplementedError("Exogenous variables not yet supported")
-        return self._predict(y, exog)
+        x = self._predict(y, exog)
+        return x
 
     @abstractmethod
     def _predict(self, y=None, exog=None): ...

aeon/forecasting/tests/test_regressor.py

@@ -1,14 +1,16 @@
 """Test the regression forecaster."""
 
+import numpy as np
+import pytest
 from sklearn.linear_model import LinearRegression
 
-from aeon.datasets import load_airline
 from aeon.forecasting import RegressionForecaster
+from aeon.regression import DummyRegressor
 
 
 def test_regression_forecaster():
     """Test the regression forecaster."""
-    y = load_airline()
+    y = np.random.rand(100)
     f = RegressionForecaster(window=10)
     f.fit(y)
     p = f.predict()
@@ -20,3 +22,13 @@ def test_regression_forecaster():
     f2.fit(y)
     p2 = f2.predict()
     assert p == p2
+    f2 = RegressionForecaster(regressor=DummyRegressor(), window=10)
+    f2.fit(y)
+    f2.predict()
+
+    with pytest.raises(ValueError):
+        f = RegressionForecaster(window=-1)
+        f.fit(y)
+    with pytest.raises(ValueError):
+        f = RegressionForecaster(window=101)
+        f.fit(y)