Use an arbitrary function to set up radial gradient

[mogres]

Use Case

This functionality will be useful to recreate observed radial distributions of punctate structures such as peroxisomes and endosomes.

Solution

This could be implemented as a new gradient mode

1. Get user input as a single valued function of distance from a surface. This input should ideally vary between 0 and 1 indicating the scaled distance between the nucleus and the membrane
2. Calculate weights for grid points based on this function
3. Use these weights to set up a gradient and pack objects based on these weights

Considerations

We need to account for differences in number of grid points as a function of radial distance. Even if the weight for all grid points is the same, points farther away from the nucleus are more likely to be selected simply because there are more grid points available.

Perhaps we can correct for this by adding a factor to the original weights

Another alternative would be to select a weight first, then select a random grid point with that given weight

[mogres]

Added a point picking mode which implements the considerations above