Computation of PDF integral

[m-aguena]

In the current computation of the weights for pdfs with shared bins:

        mask = (pzbins >= zmin) * (pzbins <= zmax)
        z_grid = pzbins[mask]
        pz_matrix = np.array(pzpdf)[:, mask]
        kernel_matrix = kernel(z_grid)

    return simpson(pz_matrix * kernel_matrix, x=z_grid, axis=1)

the precision is limited to the pdf bin sizes (pzbins). This can have a significant impact on the integral. For instance:

and dz=0.05 is similar to what we might get from the PZ WG. So I propose to interpolate the pdf in finer bins before integration