Error when using sample_weight paramater in ```fit``` function of LocalClassifierPerParentNode and SGDClassifier

[johanmazelanssi]

Describe the bug
Using a sample_weight parameter in fit function of LocalClassifierPerParentNode and SGDClassifier cause an error inside the scikit-learn code

To Reproduce

import numpy as np

from sklearn.utils.class_weight import compute_sample_weight

from sklearn.linear_model import SGDClassifier

from hiclass import LocalClassifierPerParentNode


random_state = 0

# Example from https://hiclass.readthedocs.io/en/latest/auto_examples/plot_empty_levels.html
# Define data
X_train = [[1, 2], [3, 4], [5, 6], [7, 8], [9, 10]]
X_test = [[9, 10], [7, 8], [5, 6], [3, 4], [1, 2]]
Y_train = np.array(
    [
        ["Bird"],
        ["Reptile", "Snake"],
        ["Reptile", "Lizard"],
        ["Mammal", "Cat"],
        ["Mammal", "Wolf", "Dog"],
    ],
    dtype=object,
)

clf = SGDClassifier(loss="log_loss",
                    n_jobs=1,
                    shuffle=True,
                    random_state=random_state)

classifier = LocalClassifierPerParentNode(local_classifier=clf)

Y_train_mod=["_".join(l) for l in Y_train]
print(f"Y_train_mod: {Y_train_mod}")
sample_weight_l = compute_sample_weight(class_weight="balanced", y=Y_train_mod)

classifier.fit(X_train, Y_train, sample_weight=sample_weight_l)

predictions = classifier.predict(X_test)

print(f"predictions: {predictions}")

The observed output is ValueError: Provided ``coef_`` does not match dataset.

Expected behavior
The script should execute without an error.

Versions:

• Python 3.12.1
• HiClass 5.0.4

Additional information
The problem looks specific to SGDClassifier. If I use clf = RandomForestClassifier(max_depth=20, random_state=random_state), the script is ok.