Trailing line of scattered animation

[jerpint]

Is it possible to have a trailing scatter animation? I have multiple points on a surface being animated (i.e. random walks) and they are following a trajectory, however I would like to visualize the entire trail of these points through time as it happens (where did they go, did they go where I would want, etc.).

[maartenbreddels]

I think you could 'abuse' the Mesh to do that. For each keyframe, you would repeat the coordinates of the last frame, and then the line indices would point from the current to the previous. In the next keyframe again, etc. Do you understand the idea, if not I could give a more concrete example tomorrow.

[jerpint]

I think I get the idea, but not sure what the most 'clean' and 'pythonic' way to do this would be, do you have an example on hand?

[jerpint]

I should probably mention that I am computing values beforehand; it looks something like this:

xrange = np.arange(-xlim, xlim, 1*step)
yrange = np.arange(-ylim, ylim, 1*step)
x = np.meshgrid(xrange, yrange)

# Compute function over the entire meshgrid and plot
Z = f(x)
X, Y = x
fig = ipv.figure()
ipv.plot_mesh(X, Y, Z, color="cyan")

# Precompute results to some optimization
xx, yy, zz = some_optimization(*args)

# Scatter results
s = ipv.scatter(xx, yy, zz, size=1.25, color=colors, marker="sphere")
ipv.animation_control(s)
ipv.show()

[maartenbreddels]

Sth like this:

import ipyvolume as ipv
import numpy as np

N = 4
x0 = np.linspace(0, 1, N)
x1 = x0 + 0.1
x2 = x0 + 0.2
x3 = x0 + 0.3
y0 = x0*2.0
y1 = x1**2.2
y2 = x2**2.5
y3 = x3**2.7

fig = ipv.figure()
x = np.zeros((3, N*2))
y = np.zeros((3, N*2))
z = x * 0 + 0.5
x[0] = np.concatenate([x0, x1])
x[1] = np.concatenate([x1, x2])
x[2] = np.concatenate([x2, x3])
y[0] = np.concatenate([y0, y1])
y[1] = np.concatenate([y1, y2])
y[2] = np.concatenate([y2, y3])
lines = [[0,N], [1, 1+N], [2, 2+N], [3, 3+N]]
# lines = [0,3, 1, 3, 2, 4]
p = ipv.plot_trisurf(x, y, z, lines=lines);
ipv.animation_control(p)
ipv.show()