Plotting overlaying fields in different colors.

[nathanLaubeuf]

Cannot plot fields wrapped with vis.overlay in different colors since f03e6bd.
This used to be supported when color was an attribute of the field.
Can it be fixed, maybe with a different indexing of vis.plot's color argument?

[holl-]

The easiest way to do it is to stack your data along a channel dimension that is not called 'vector'.
Then the different components will automatically get different colors.

plot(CenteredGrid(Noise(colors='col1,col2'), x_axis=10))
plot(math.random_uniform(instance(x=10), channel(colors=2, vector='x,y')))

You can additionally specify the colors directly, either as cycle indices or via hex code:

plot(CenteredGrid(Noise(colors='col1,col2'), x_axis=10), color=wrap([2, 3], channel('colors')))
plot(math.random_uniform(instance(x=10), channel(colors=2, vector='x,y')), color=wrap(["#000000", "#FF0000"], channel('colors')))

Does that solve your issue?

[nathanLaubeuf]

No luck there. I'm trying to plot 3 populations of points with different sizes.

rock = math.random_uniform(instance(x=10), channel(vector='x,y'))
paper = math.random_uniform(instance(x=5), channel(vector='x,y'))
scissors = math.random_uniform(instance(x=7), channel(vector='x,y'))

particles = math.stack((rock, paper, scissors), channel(colors=3))

vis.plot(particles, color=wrap(["#0000FF", "#FF0000", "#00FF00"], channel(colors=3)))

This snippet breaks with a NotImplementedError from particles being a non-uniform tensor.

[holl-]

Okay, I got it to work by replacing math.stack by math.layout.
This runs on the latest develop version at least.

The stack version unfortunately results in non-uniform tensors for which not all operations are supported yet.
The layout is basically just like a list with a named dimension.

[nathanLaubeuf]

That fixed it. Thanks!