Smoothing of correlated variables

The smoothing implemented in corner.py does not handle strongly correlated variables very well.

The smoothing for 2D histograms is done with the scipy gaussian_filter function, which is isotropic. If your variables are highly correlated this is pretty distorting, since it smooths points in the wrong directions. See the Delta-M versus h subplot in the attached image, where the contour is enlarged a lot in one direction.

The easiest way to solve this in general is to either supply a covariance matrix to use an elliptical smoothing kernel, or transform to variables with unit covariance before smoothing. Neither of those is a simple change here, unfortunately.

Start by reading corner.py and locating the 2D histogram smoothing path that calls scipy's gaussian_filter. Compare the current isotropic behavior with the correlated Delta-M versus h example, then determine whether covariance-based smoothing or a unit-covariance transformation fits the design. Done means strongly correlated variables no longer produce contours enlarged in the wrong direction.