def test_hits_at_n_score():
y_pred_true = np.array([[[0, 1, 0], [.32, .84, .73]],
[[0, 1, 0], [.66, .11, .33]]])
rankings = []
for y_pred_true_k in y_pred_true:
rankings.append(rank_score(y_pred_true_k[0], y_pred_true_k[1]))
hits_actual = hits_at_n_score(rankings, n=2)
assert hits_actual == 0.5
def test_mrr_score():
y_pred_true = np.array([[[0, 1, 0], [.32, .84, .73]],
[[0, 1, 0], [.66, .11, .33]]])
rankings = []
for y_pred_true_k in y_pred_true:
rankings.append(rank_score(y_pred_true_k[0], y_pred_true_k[1]))
mrr_actual = mrr_score(rankings)
np.testing.assert_almost_equal(mrr_actual, 0.66666, decimal=5)
def test_rank_score():
y_pred = np.array([.434, .65, .21, .84])
y_true = np.array([0, 0, 1, 0])
rank_actual = rank_score(y_true, y_pred)
assert rank_actual == 4