MAINT Parameters validation for sklearn.model_selection.cross_val_pre…

…dict (scikit-learn#26252)

Co-authored-by: jeremie du boisberranger <[email protected]>

sklearn/model_selection/_validation.py

@@ -884,6 +884,30 @@ def _score(estimator, X_test, y_test, scorer, error_score="raise"):
  return scores
 
 
+@validate_params(
+ {
+ "estimator": [HasMethods(["fit", "predict"])],
+ "X": ["array-like", "sparse matrix"],
+ "y": ["array-like", None],
+ "groups": ["array-like", None],
+ "cv": ["cv_object"],
+ "n_jobs": [Integral, None],
+ "verbose": ["verbose"],
+ "fit_params": [dict, None],
+ "pre_dispatch": [Integral, str, None],
+ "method": [
+ StrOptions(
+ {
+ "predict",
+ "predict_proba",
+ "predict_log_proba",
+ "decision_function",
+ }
+ )
+ ],
+ },
+ prefer_skip_nested_validation=False, # estimator is not validated yet
+)
 def cross_val_predict(
  estimator,
  X,
@@ -912,10 +936,11 @@ def cross_val_predict(
 
  Parameters
  ----------
- estimator : estimator object implementing 'fit' and 'predict'
- The object to use to fit the data.
+ estimator : estimator
+ The estimator instance to use to fit the data. It must implement a `fit`
+ method and the method given by the `method` parameter.
 
- X : array-like of shape (n_samples, n_features)
+ X : {array-like, sparse matrix} of shape (n_samples, n_features)
  The data to fit. Can be, for example a list, or an array at least 2d.
 
  y : array-like of shape (n_samples,) or (n_samples, n_outputs), \

sklearn/tests/test_public_functions.py

@@ -269,6 +269,7 @@ def _check_function_param_validation(
  "sklearn.metrics.top_k_accuracy_score",
  "sklearn.metrics.v_measure_score",
  "sklearn.metrics.zero_one_loss",
+ "sklearn.model_selection.cross_val_predict",
  "sklearn.model_selection.cross_val_score",
  "sklearn.model_selection.cross_validate",
  "sklearn.model_selection.learning_curve",