Resolve "(LearnerND) allow any convex hull as domain"

Closes #114 (closed)

Works especially well when combined with !124 (merged)

usage:

import scipy
import adaptive
adaptive.notebook_extension()

def f(xyz):
    x, y, z = xyz
    return x**4 + y**4 + z**4 - (x**2+y**2+z**2)**2

b = [(-1, -1, -1),
     (-1,  1, -1),
     (-1, -1,  1),
     (-1,  1,  1),
     ( 1,  1, -1),
     ( 1, -1, -1)]

bounds = scipy.spatial.ConvexHull(b)
learner = adaptive.LearnerND(f, bounds)
runner = adaptive.Runner(learner, goal=lambda l:l.npoints>1000)
runner.live_info()

Now, if you indeed also have !124 (merged) at your disposal, you are able to run something like this:

learner.plot_isosurface(-0.5)

in order to get this:

otherwise you could run:

learner.plot_slice({1:0})

to get this:

changed the description

added 4 commits

• 18106f5f...3a69cd99 - 2 commits from branch master
• 435405d1 - allow passing a convex hull to LearnerND
• b513dc3a - correctly set bounds

Compare with previous version

added 23 commits

• b513dc3a...ad7a3d23 - 22 commits from branch master
• 9d368d64 - allow passing a convex hull to LearnerND

Compare with previous version

mentioned in issue #117

resolved all discussions

added 1 commit

• f1ef8309 - only do expensive inside_bounds check when the hull is not rectangular

Compare with previous version

added 8 commits

• f1ef8309...6cdd24e9 - 6 commits from branch master
• 50890a99 - allow passing a convex hull to LearnerND
• c16c89ef - only do expensive inside_bounds check when the hull is not rectangular

Compare with previous version

added 1 commit

• c0472fab - only do expensive inside_bounds check when the hull is not rectangular

Compare with previous version

added 4 commits

• c0472fab...7f936b7c - 2 commits from branch master
• cb5facdd - allow passing a convex hull to LearnerND
• 02e66a0b - only do expensive inside_bounds check when the hull is not rectangular

Compare with previous version

We can add a test for adding rectangular bounds as a hull and checking that we get the same result as passing the bounds "as usual".

We should also add an example to the LearnerND documentation!

mentioned in merge request !124 (merged)

added 2 commits

• c59041a1 - added test
• 8a48725c - make test pass by sorting the bound points and adding an eps to the

Compare with previous version

added 19 commits

• 8a48725c...1ffe2209 - 15 commits from branch master
• 939074cc - allow passing a convex hull to LearnerND
• 70ef26e0 - only do expensive inside_bounds check when the hull is not rectangular
• 790a2737 - added test
• e1c43304 - make test pass by sorting the bound points and adding an eps to the

Compare with previous version

added 2 commits

• b90b9d0d - add convexhull example to the docs
• 73326415 - use correct format

Compare with previous version

added 1 commit

• 3c777158 - Update tutorial.LearnerND.rst

Compare with previous version

resolved all discussions

resolved all discussions