.mean() and .moment(1) returning incorrect answers for sharp, non-zero values at extrema

[sschmidt23]

It looks like the .mean() and .moment(1) methods produce faulty answers when there is a sudden rise to a non-zero value at either extrema of a qp.interp grid (may happen in others, but I found for sure in qp.interp). For example, if you try:

import numpy as np
import qp
from scipy.stats import norm
zgrid = np.linspace(0,3,301)
func1 = norm(loc=3.0, scale=0.05)
func2 = norm(loc=3.0, scale=0.25)
x1 = func1.pdf(zgrid)
x2 = func2.pdf(zgrid)
ydata = np.vstack((x1, x2)) 
ens = qp.Ensemble(qp.interp, data=dict(xvals=zgrid, yvals = ydata))
ens.mean()

you get:

array([[3.19960439],
           [2.84842853]])

The first value, with the very small half-Gaussian centered right on the end point of 3.0, is incorrectly overshot, while the broader half-Gaussian with scale=0.25 produces the correct answer for 2.8484. So, this problem seems to only occur for a rapid rise in the PDF. I noticed this while computing means for some FZBoost PDFs where the mean was being reported as high as z=3.5 when the grid stopped at z=3.0, which is obviously a problem.