Need to skip "bad" values during scale transformation

[ASmirnov-HORIS]

The following code produces an exception ends can't be None:

import numpy as np

from lets_plot import *
LetsPlot.setup_html()

np.random.seed(42)

n = 1000
x = np.linspace(-10, 10, n)
y = x ** 2 + np.random.normal(size=n)
z = 2.0 ** x + np.random.normal(size=n)
data = {'x': x, 'y': y, 'z': z}

ggplot(data, aes(x='x', y='y')) + \
    geom_point(aes(color='z')) + \
    scale_y_log10() + scale_color_gradient(low='red', high='green', trans='log10')

Without trans parameter I have this:

It would be better if the color were transformed logarithmically but there are a few bad values in the data (with negative z). The exception message doesn't explain what exactly goes wrong and what I should do with my data. Also I want to keep my data immutable, so I prefer to have some parameter for skipping non-colorable data - then "bad" points would be colored gray or something.