def log_attributes_stats()

in src/loader.py [0:0]

• 11 lines of code
• 5 McCabe index (conditional complexity)

def log_attributes_stats(train_attributes, valid_attributes, test_attributes, params):
    """
    Log attributes distributions.
    """
    k = 0
    for (attr_name, n_cat) in params.attr:
        logger.debug('Train %s: %s' % (attr_name, ' / '.join(['%.5f' % train_attributes[:, k + i].mean() for i in range(n_cat)])))
        logger.debug('Valid %s: %s' % (attr_name, ' / '.join(['%.5f' % valid_attributes[:, k + i].mean() for i in range(n_cat)])))
        logger.debug('Test  %s: %s' % (attr_name, ' / '.join(['%.5f' % test_attributes[:, k + i].mean() for i in range(n_cat)])))
        assert train_attributes[:, k:k + n_cat].sum() == train_attributes.size(0)
        assert valid_attributes[:, k:k + n_cat].sum() == valid_attributes.size(0)
        assert test_attributes[:, k:k + n_cat].sum() == test_attributes.size(0)
        k += n_cat
    assert k == params.n_attr