Isotonic regression doesn't not respect y_min and y_max arguments #6921
Description
Reported by Jonathan Taylor.
While trying to fit an isotonic regression with not trivial y_min and y_max, the fit returned a solution that does not fit the bounds:
import numpy as np
from sklearn import isotonic
X = np.array([1.26336413, 1.31853693, -0.57200917, 0.3072928, -0.70686507,
-0.17614937, -1.59943059, 1.05908504, 1.3958263, 1.90580318,
0.20992272, 0.02836316, -0.08092235, 0.44438247, 0.01791253,
-0.3771914, -0.89577538, -0.37726249, -1.32687569, 0.18013201])
Y = isotonic.isotonic_regression(X, y_min=0, y_max=0.1)
assert(np.all(Y > 0))
assert(np.all(Y < 0.1))