Suggest a Bayesian prior for a probabilistic ellipse slice. The slice should be normally distributed points in the area between two non-intersecting ellipses such that I need to generate points from such an area. It's important that the Kotz distribution is used. Answer with a python code for generating the points and an image that shows the result, as well as equations. The solution should use the covariance matrix between two variables and the mean vector of two entries.