Use holoviews.TriMesh when it makes it to a release

[basnijholt]

(original issue on GitLab)

opened by Bas Nijholt (@basnijholt) at 2017-12-02T16:19:48.343Z

See [holoviews gitlab:#2143](holoviz/holoviews#2143), this will make plotting with the triangular mesh much faster.

[basnijholt]

originally posted by Bas Nijholt (@basnijholt) at 2017-12-03T16:20:41.405Z on GitLab

This looks pretty cool:

%%output size=200
%%opts TriMesh [filled=True edge_color_index='z' fig_size=200] (cmap='viridis' node_alpha=0)
simplices = ip.tri.simplices
nodes = ip.tri.points
z = ip.values[simplices].mean(axis=1)
hv.TriMesh((np.column_stack([simplices, z]), nodes), vdims='z')

[basnijholt]

originally posted by Bas Nijholt (@basnijholt) at 2017-12-08T22:24:03.858Z on GitLab

lol:

def plot(learner):
    ip = learner.ip()
    simplices = ip.tri.simplices
    nodes = ip.tri.points
    z = ip.values[simplices].mean(axis=1)
    return hv.TriMesh((np.column_stack([simplices, z]), nodes), vdims='z')

%timeit plot(learner)
2.71 ms ± 309 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

%timeit learner.plot(triangles_alpha=True)
216 ms ± 13.9 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)