graph_objects.Mesh3d broken when `x` and `y` is identical, or something else?

[kaimo455]

Broken codes, where are several scenarios:

Scenario 1: x and y are identical

x = np.random.randn(100)
y = x.copy()
z = np.random.randn(100)
fig = go.Figure(data=[go.Mesh3d(x=x, y=y, z=z, opacity=0.50)])
fig.show()

Scenario 2: x and y are created by np.linspace

import plotly.graph_objects as go
import numpy as np
x = np.linspace(-10, 10, 100)
y = np.linspace(-6, 12, 100) 
z = np.random.randn(100)
fig = go.Figure(data=[go.Mesh3d(x=x, y=y, z=z, opacity=0.50)])
fig.show()

The worked codes

import plotly.graph_objects as go
import numpy as np
x = np.random.randn(100)
y = np.random.randn(100)
z = np.random.randn(100)
fig = go.Figure(data=[go.Mesh3d(x=x, y=y, z=z, opacity=0.50)])
fig.show()

[alexcjohnson]

When you don't provide index arrays i, j, k we need to infer how to connect these (x,y,z) points with triangles. The default is Delaunay triangulation looking along the z axis (ie ignoring the z data and just using the data in the xy plane). And when x==y all of the resulting triangles are degenerate (they're just lines) - so it's not surprising the algorithm doesn't know what to do with this. You can get a visible result if you add delaunayaxis='x' (or 'y'), though still the first render shows nothing because the default projection is looking edge-on at the infinitely-thin slice, but if you rotate you'll see it

[kaimo455]

wow, that's cool!
Thank you for your detailed response!