Kmeans fit sideeffect: sample_weight is parameter is modified #17203
Description
Describe the bug
The array passed in to fit() or fit_predict for the argument sample weight is modified (rescaled) in the method with side effects outside the call.
Steps/Code to Reproduce
Example:
import numpy as np
from sklearn.cluster import KMeans
X = np.array([[1], [2], [4]])
sample_weights = np.array([0.5, 0.2, 0.3])
kmeans = KMeans(n_clusters=2, random_state=4321)
y_kmeans = kmeans.fit_predict(X, sample_weight=sample_weights)
print(sample_weights)