[BUG] Test fails with newer scikit-learn

[ckastner]

Test test_classification_report_imbalanced_multiclass_with_unicode_label fails with scikit-learn >= 0.23.2 because the installed NumPy version representation has changed, see MNT define a parse_version function in sklearn.utils.fixes.

=================================== FAILURES ===================================
_____ test_classification_report_imbalanced_multiclass_with_unicode_label ______

    def test_classification_report_imbalanced_multiclass_with_unicode_label():
        y_true, y_pred, _ = make_prediction(binary=False)
    
        labels = np.array(["blue\xa2", "green\xa2", "red\xa2"])
        y_true = labels[y_true]
        y_pred = labels[y_pred]
    
        expected_report = (
            "pre rec spe f1 geo iba sup blue¢ 0.83 0.79 0.92 0.81 "
            "0.85 0.72 24 green¢ 0.33 0.10 0.86 0.15 0.29 0.08 31 "
            "red¢ 0.42 0.90 0.55 0.57 0.70 0.51 20 avg / total "
            "0.51 0.53 0.80 0.47 0.58 0.40 75"
        )
>       if np_version[:3] < (1, 7, 0):
E       TypeError: 'Version' object is not subscriptable

I will submit a PR with a proposed change of bumping the dependency to 0.23.2, and updating the comparison.

[ckastner]

Closing, as the solution in #745 is superior.