Compatibility issue: _IsotonicRegression class not inheriting from scikit-learn's BaseEstimator

[cyrfar]

When calibrating a model using classifier.calibrate(X_cal, Y_cal) and calibration_method='isotonic', an AttributeError is raised during probability prediction. The error message says that _IsotonicRegression object has no attribute __sklearn_tags__. This seems to be due to the _IsotonicRegression class not inheriting from scikit-learn's BaseEstimator, which is required for compatibility with scikit-learn. Similar errors occurs with other calibration methods.

To Reproduce
Steps to reproduce the behavior:

1. Install hiclass and scikit-learn in a Python environment.
2. Run the full example code from the docs:

from sklearn.ensemble import RandomForestClassifier

from hiclass import LocalClassifierPerNode

# Define data
X_train = [[1], [2], [3], [4]]
X_test = [[4], [3], [2], [1]]
X_cal = [[5], [6], [7], [8]]
Y_train = [
    ["Animal", "Mammal", "Sheep"],
    ["Animal", "Mammal", "Cow"],
    ["Animal", "Reptile", "Snake"],
    ["Animal", "Reptile", "Lizard"],
]

Y_cal = [
    ["Animal", "Mammal", "Cow"],
    ["Animal", "Mammal", "Sheep"],
    ["Animal", "Reptile", "Lizard"],
    ["Animal", "Reptile", "Snake"],
]

# Use random forest classifiers for every node
rf = RandomForestClassifier()

# Use local classifier per node with isotonic regression as calibration method
classifier = LocalClassifierPerNode(
    local_classifier=rf, calibration_method="isotonic", probability_combiner="multiply"
)

# Train local classifier per node
classifier.fit(X_train, Y_train)

# Calibrate local classifier per node
classifier.calibrate(X_cal, Y_cal)

# Predict probabilities
probabilities = classifier.predict_proba(X_test)

# Print probabilities and labels for the last level
print(classifier.classes_[2])
print(probabilities)

Expected behavior
The code should run without errors and return calibrated probability predictions for the test set.

Screenshots
N/A. Here is the error trace

AttributeError: The following error was raised: '_IsotonicRegression' object has no attribute '__sklearn_tags__'. It seems that there are no classes that implement sklearn_tagsin the MRO and/or all classes in the MRO callsuper().sklearn_tags(). Make sure to inherit from BaseEstimatorwhich implementssklearn_tags(or alternatively definesklearn_tagsbut we don't recommend this approach). Note thatBaseEstimator needs to be on the right side of other Mixins in the inheritance order.

Desktop (please complete the following information):

• OS: MacOS Tahoe 26.2
• Version: Python 3.12, hiclass 5.0.4, scikit-learn 1.8.0

Additional context
The issue goes away when downgrading to scikit-learn 1.7.2