sampling_vertex_vw doesn't work as expected

[IKupriyanov-HORIS]

The quality of the resulting geometries is unacceptably bad:

from lets_plot import *
LetsPlot.setup_html()

def f(x, y, n):
    z = np.sin(x * 3 * np.pi / n)
    z += np.sin(y * 3 * np.pi / n)
    z += x * 3 / n
    z += y * 5 / n
    return z

n = 50
x = np.arange(n)
y = np.arange(n)
X, Y = np.meshgrid(x, y)
Z = f(X, Y, n)
data = dict(
    x=X.reshape(-1),
    y=Y.reshape(-1),
    z=Z.reshape(-1))

p = ggplot(data, aes('x','y', z='z')) \
    + theme_void() + ggsize(200, 200)
p + geom_contour(sampling=sampling_vertex_vw(100))

[IKupriyanov-HORIS]

Fixed via 80fcbf2