fix(regressor): correctly cap the labels in predict (#1662)

updates pre-commit

Author: aron-bram

.pre-commit-config.yaml

@@ -4,7 +4,7 @@
 repos:
 
   - repo: https://github.com/pycqa/isort
-    rev: 5.10.1
+    rev: 5.11.5
     hooks:
       - id: isort
         name: isort imports autosklearn
@@ -15,7 +15,7 @@ repos:
         files: test/.*
 
   - repo: https://github.com/psf/black
-    rev: 22.10.0
+    rev: 23.3.0
     hooks:
       - id: black
         name: black formatter autosklearn
@@ -31,15 +31,15 @@ repos:
 
   # This is disabled as most modules fail this
   - repo: https://github.com/pycqa/pydocstyle
-    rev: 6.1.1
+    rev: 6.3.0
     hooks:
       - id: pydocstyle
         files: DISABLED # autosklearn/.*
         always_run: false
         additional_dependencies: ["toml"] # Needed to parse pyproject.toml
 
   - repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v0.990
+    rev: v1.2.0
     hooks:
       - id: mypy
         name: mypy auto-sklearn

autosklearn/pipeline/regression.py

@@ -106,12 +106,35 @@ def iterative_fit(self, X, y, n_iter=1, **fit_params):
         )
 
     def predict(self, X, batch_size=None):
+        """Predict the classes using the selected model.
+
+        Predicted values are capped to approximately the maximum and minimum labels
+        seen during training.
+
+        Parameters
+        ----------
+        X : array-like, shape = (n_samples, n_features)
+
+        batch_size: int or None, defaults to None
+            batch_size controls whether the pipeline will be
+            called on small chunks of the data. Useful when calling the
+            predict method on the whole array X results in a MemoryError.
+
+        Returns
+        -------
+        array, shape=(n_samples,) if n_classes == 2 else (n_samples, n_classes)
+            Returns the predicted values"""
         y = super().predict(X, batch_size=batch_size)
-        y[y > (2 * self.y_max_)] = 2 * self.y_max_
+
+        if self.y_max_ > 0:
+            y[y > (2 * self.y_max_)] = 2 * self.y_max_
+        elif self.y_max_ < 0:
+            y[y > (0.5 * self.y_max_)] = 0.5 * self.y_max_
         if self.y_min_ < 0:
             y[y < (2 * self.y_min_)] = 2 * self.y_min_
         elif self.y_min_ > 0:
             y[y < (0.5 * self.y_min_)] = 0.5 * self.y_min_
+
         return y
 
     def _get_hyperparameter_search_space(