Incorrect usage of numpy.setdiff1d

[JATC1024]

I assumle that line 349 of the file deepimpute/multinet.py genes_not_in_target = np.setdiff1d(targets, covariance_matrix.columns) tries to get the genes that are in the covariance matrix but not in "targets". However, the line of code above does the opposite, i.e. getting the genes that are in targets but not in the matrix's columns. Is this a bug?