Trending EE plot normalization issue

[mperrin]

Reported by @mlibralato:

I'm studying the temporal variations of the ePSFs. I took the code in WebbPSF that show how the encircled energy varies over time and I correlated it with by how much my ePSFs seem to vary over time. I do not see any significant change, as I was told should be at this wavelengths. The comment here is that the code that makes this plot here (https://webbpsf.readthedocs.io/en/latest/jwst_measured_opds.html#Wavefront-Trending-throughout-a-given-month) computes the change in the encircled energy as EE_ref - EE_date but this is not a % change, so the shadowed box that highlight the +/-3% variations is not correct. Do you agree? I believe that to be consistent the plot should show on the y axis something like (EE_ref-EE_date)/EE_ref