Update norm function in transform.py

[NolanBrb]

New feature / Bug Fix

Added default behaviour when values to normalize are all equal; previously holoviews would attempt division by zero and return a runtime warning.
Proposed change : if all values are equal to zero, set all values to 0; otherwise, set all values to 1.

[hoxbro]

Can you explain why this is needed?

[NolanBrb]

Sure !

I ran into this error because I'm working on xarray datasets with multiple variables and procedurally creating vectorfields for each variable. I'm controlling the scale of the vectors using .opts(magnitude=hv.dim("Magnitude").norm())
Some of the variables I use have fixed magnitude, and when displaying the plot the line above returned a warning because since min==max, norm() was attempting to divide by zero.

[hoxbro]

I'm not opposed to this PR, but currently, it breaks CI.

Did you try to pass in limits to hv.dim("Magnitude").norm(limits=(0, None))? Or would you like it to be dynamically updated?

Would you happen to have a small example I could play around with?