from sklearn.svm import LinearSVC
from sklearn.base import clone

from sklearn.datasets import make_classification
import numpy as np

rng = np.random.RandomState()

X, y = make_classification(
    n_samples=100,
    n_features=5,
    n_informative=3,
    n_classes=4,
    random_state=0,
)

# Create dataset with repetitions and corresponding sample weights
sample_weight = rng.randint(0, 10, size=X.shape[0])
X_resampled_by_weights = np.repeat(X, sample_weight, axis=0)
y_resampled_by_weights = np.repeat(y, sample_weight)

est_sw = LinearSVC(dual=False,class_weight="balanced").fit(X, y, sample_weight=sample_weight)
est_dup = LinearSVC(dual=False,class_weight="balanced").fit(
    X_resampled_by_weights, y_resampled_by_weights, sample_weight=None
)

np.testing.assert_allclose(est_sw.coef_, est_dup.coef_,rtol=1e-10,atol=1e-10)
np.testing.assert_allclose(
    est_sw.decision_function(X_resampled_by_weights),
    est_dup.decision_function(X_resampled_by_weights),
    rtol=1e-10,
    atol=1e-10
)