Describe the workflow you want to enable

Would be good to get sample_weight support for QuantileTransformer for dealing with sparse or imbalanced data, a la #15601.

scaler = QuantileTransformer(output_distribution="normal")

scaler.fit(X, sample_weight=w)

Describe your proposed solution

As far as I know it would just require adding the weight argument to the quantiles_ computation in np.nanpercentile.

KBinsDiscretizer supports sample_weight and with strategy='quantile', encode='ordinal' this behavior can be achieved but it is much, much slower.

Describe alternatives you've considered, if relevant

No response

Additional context

No response