ValueError: Contour levels must be increasing

[Cameron307]

I'm going through the tutorial, but I'm stuck on the section "Simple Bayesian Modeling" because running the solution code for the Breakout exercise gives an error "ValueError: Countour levels must be increasing". I'm not an experienced programmer, so I could use some help getting this demonstration code to run.

[flothesof]

Hi Cameron,

you can easily fix this if you reverse the values returned by the function that computes the contour levels.
A fixed version reads:

def contour_levels(grid):
    """Compute 1, 2, 3-sigma contour levels for a gridded 2D posterior"""
    sorted_ = np.sort(grid.ravel())[::-1]
    pct = np.cumsum(sorted_) / np.sum(sorted_)
    cutoffs = np.searchsorted(pct, np.array([0.68, 0.95, 0.997]) ** 2)
    return sorted_[cutoffs][::-1]

Notice the [::-1] on the last line.

Cheers,

Florian