add unit tests for epsilon in normal dist and var_smoothing in generalnb

tests/test_dist.py

@@ -62,6 +62,21 @@ def test_normal_pdf():
     assert_array_almost_equal(norm_wnb(X), norm_scipy.pdf(X), decimal=10)
 
 
+@pytest.mark.parametrize("epsilon", [1e-10, 1e-9, 1e-6, 1e-3])
+def test_normal_with_epsilon(epsilon: float):
+    """
+    Test whether epsilon is correctly applied for `NormalDist`.
+    """
+    norm_1 = NormalDist(mu=1, sigma=0)
+    norm_2 = NormalDist(mu=1, sigma=0, epsilon=epsilon)
+    norm_3 = NormalDist(mu=1, sigma=np.sqrt(epsilon))
+    assert norm_1.sigma == norm_2.sigma == 0
+    assert norm_3.sigma == np.sqrt(epsilon)
+    X = np.random.uniform(-100, 100, size=10000)
+    assert np.isnan(norm_1(X)).all()
+    assert_array_almost_equal(norm_2(X), norm_3(X), decimal=10)
+
+
 def test_lognormal_pdf(random_uniform):
     """
     Test whether pdf method of `LognormalDist` returns the same result as pdf method of `scipy.stats.lognorm`.

tests/test_gnb.py

@@ -247,3 +247,37 @@ def test_gnb_invalid_dist():
     msg = r"Distribution .* is not supported"
     with pytest.raises(ValueError, match=msg):
         clf.fit(X, y)
+
+
+def test_gnb_var_smoothing():
+    """
+    Test whether var_smoothing parameter properly affects the variances of normal distributions.
+    """
+    X = np.array([[1, 0], [2, 0], [3, 0], [4, 0], [5, 0]])  # First feature has variance 2.0
+    y = np.array([1, 1, 2, 2, 2])
+
+    clf1 = GeneralNB(var_smoothing=0.0)
+    clf1.fit(X, y)
+
+    clf2 = GeneralNB(var_smoothing=1.0)
+    clf2.fit(X, y)
+
+    test_point = np.array([[2.5, 0]])
+    prob1 = clf1.predict_proba(test_point)
+    prob2 = clf2.predict_proba(test_point)
+
+    assert not np.allclose(prob1, prob2)
+    assert clf1.epsilon_ == 0.0
+    assert clf2.epsilon_ > clf1.epsilon_
+
+
+def test_gnb_var_smoothing_non_numeric():
+    """
+    Test that var_smoothing is ignored for non-numeric features.
+    """
+    X = np.array([["a", 1], ["b", 2], ["a", 2], ["b", 1]])
+    y = np.array([1, 1, 2, 2])
+
+    clf = GeneralNB(distributions=[D.CATEGORICAL, D.CATEGORICAL], var_smoothing=1e-6)
+    clf.fit(X, y)
+    assert clf.epsilon_ == 0