How to retrieve the subset of n least correlated series from a covariance matrix in Python?

Question

Given a 25x25 covariance matrix, how can I extract the 5 least intercorrelated indexes?

My quick guess is that I should sum each row of the covariance matrix in a new