Repository navigation
Learner2D.plot() returns NaN #181
Description
Example:
import adaptive
adaptive.notebook_extension()
import numpy as np
bounds = [(1, 1.7), (250, 550)]
learner = adaptive.Learner2D(lambda x: x[0] * x[1], bounds=bounds)
adaptive.runner.simple(learner, goal=lambda l: l.npoints >= 2000)
plot_data = learner.plot().Image.I.data
print(np.sum(np.isnan(plot_data)))
>>> 47The issue is the precision in Learner2D._scale (which is called by Learner2D.ip, which itself is used to make the plot):
points, values = learner._data_in_bounds()
points = learner._scale(points)
print(np.min(points[:, 0]), np.max(points[:, 0]))
>>> -0.5000000000000001 0.49999999999999983Points lying exactly at x = 0.5 will be outside of the convex hull of points and scipy.interpolate.LinearNDInterpolator fills them with nan.
Activity
basnijholt commentedon Apr 9, 2019
Thanks for reporting! I'll fix it tomorrow.
2D: do not return points outside the bounds, closes #181
2D: do not return points outside the bounds, closes #181
basnijholt commentedon Apr 10, 2019
OK this problem only really appears for plotting because
prints
and happens inside the
plotfunction becausex = y = np.linspace(-0.5, 0.5, n).Merge pull request #182 from python-adaptive/fix_nan_2d
2D: do not return points outside the bounds, closes #181
2D: do not return points outside the bounds, closes #181
2D: do not return points outside the bounds, closes #181
2D: do not return points outside the bounds, closes #181