"This means we are ready to calculate quasi-normalized structure factors by computing f_sq / <f_sq/epsilon> in resolution bins:

all_e_values = miller_array.quasi_normalize_structure_factors().sort(by_value="data")

This statement performs two steps at once. First, the quasi_normalize_structure_factors() method creates a new cctbx.miller.array instance with the same unit cell, space group, anomalous flag and Miller indices as the input miller_array, but with a new data array containing the normalized structure factors. The sort() method is used immediately on this intermediate instance to sort the E-values by magnitude. By default, the data are sorted in descending order (largest first, smallest last). This is exactly what we want here. To convince yourself it is correct, insert all_e_values.show_array()."