REF: Fix extended tests

bashtage · Sep 18, 2024 · 1c6b1b6 · 1c6b1b6
1 parent c42b361
commit 1c6b1b6
Showing 1 changed file with 47 additions and 47 deletions.
diff --git a/randomgen/tests/test_extended_generator.py b/randomgen/tests/test_extended_generator.py
@@ -41,7 +41,7 @@ def mv_seed():
 
 @pytest.fixture(scope="function")
 def extended_gen():
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     return ExtendedGenerator(pcg)
 
 
@@ -155,7 +155,7 @@ def test_multivariate_normal_method(seed, method):
 
 @pytest.mark.parametrize("method", ["svd", "eigh", "cholesky"])
 def test_multivariate_normal_basic_stats(seed, method):
-    random = ExtendedGenerator(MT19937(seed))
+    random = ExtendedGenerator(MT19937(seed, mode="sequence"))
     n_s = 1000
     mean = np.array([1, 2])
     cov = np.array([[2, 1], [1, 2]])
@@ -184,24 +184,24 @@ def test_multivariate_normal_bad_size(mean, size):
 
 
 def test_multivariate_normal(seed):
-    random.bit_generator.seed(seed)
+    random = ExtendedGenerator(MT19937(seed, mode="sequence"))
     mean = (0.123456789, 10)
     cov = [[1, 0], [0, 1]]
     size = (3, 2)
     actual = random.multivariate_normal(mean, cov, size)
     desired = np.array(
         [
             [
-                [-3.34929721161096100, 9.891061435770858],
-                [-0.12250896439641100, 9.295898449738300],
+                [0.8032359104382388, 9.020376052272443],
+                [-0.549265600883053, 7.859449222248867],
             ],
             [
-                [0.48355927611635563, 10.127832101772366],
-                [3.11093021424924300, 10.283109168794352],
+                [0.4159548290083927, 10.569454865174936],
+                [-1.5353359555824029, 9.318305704280354],
             ],
             [
-                [-0.20332082341774727, 9.868532121697195],
-                [-1.33806889550667330, 9.813657233804179],
+                [0.950123388279074, 10.055774500309264],
+                [0.13457049786723752, 9.803457241505443],
             ],
         ]
     )
@@ -210,7 +210,7 @@ def test_multivariate_normal(seed):
 
     # Check for default size, was raising deprecation warning
     actual = random.multivariate_normal(mean, cov)
-    desired = np.array([-1.097443117192574, 10.535787051184261])
+    desired = np.array([0.3668305896561763, 10.345231541085731])
     assert_array_almost_equal(actual, desired, decimal=15)
 
     # Check that non positive-semidefinite covariance warns with
@@ -244,21 +244,21 @@ def test_multivariate_normal(seed):
 
 
 def test_complex_normal(seed):
-    random.bit_generator.seed(seed)
+    random = ExtendedGenerator(MT19937(seed, mode="sequence"))
     actual = random.complex_normal(loc=1.0, gamma=1.0, relation=0.5, size=(3, 2))
     desired = np.array(
         [
             [
-                -2.007493185623132 - 0.05446928211457126j,
-                0.7869874090977291 - 0.35205077513085050j,
+                1.5887059881277819 - 0.4898119738637788j,
+                0.41740532066669644 - 1.0702753888755665j,
             ],
             [
-                1.3118579018087224 + 0.06391605088618339j,
-                3.5872278793967554 + 0.14155458439717636j,
+                1.253310733204425 + 0.2847274325874684j,
+                -0.4365566564216725 - 0.3408471478598232j,
             ],
             [
-                0.7170022862582056 - 0.06573393915140235j,
-                -0.26571837106621987 - 0.0931713830979103j,
+                1.7159142754357688 + 0.027887250154631615j,
+                1.009624754209292 - 0.09827137924727837j,
             ],
         ]
     )
@@ -321,7 +321,7 @@ def test_set_get_state(seed):
 
 
 def test_complex_normal_size(mv_seed):
-    random = ExtendedGenerator(MT19937(mv_seed))
+    random = ExtendedGenerator(MT19937(mv_seed, mode="sequence"))
     state = random.state
     loc = np.ones((1, 2))
     gamma = np.ones((3, 1))
@@ -330,16 +330,16 @@ def test_complex_normal_size(mv_seed):
     desired = np.array(
         [
             [
-                1.393937478212015 - 0.31374589731830593j,
-                0.9474905694736895 - 0.16424530802218726j,
+                1.2093667943035076 + 0.13793154570353136j,
+                1.5769647969244578 + 0.6158732529417138j,
             ],
             [
-                1.119247463119766 + 0.023956373851168843j,
-                0.8776366291514774 + 0.2865220655803411j,
+                0.5900374596315505 + 0.5238612206671325j,
+                1.2024921591940638 - 0.5370509227680859j,
             ],
             [
-                0.5515508326417458 - 0.15986016780453596j,
-                -0.6803993941303332 + 1.1782711493556892j,
+                1.088126576683737 + 0.5544754121023469j,
+                0.8485995861109008 - 0.047090051743345455j,
             ],
         ]
     )
@@ -368,7 +368,7 @@ def test_default_pcg64():
 @pytest.mark.parametrize("dim", [2, 5, 10])
 @pytest.mark.parametrize("size", [None, 5, (3, 7)])
 def test_standard_wishart_reproduce(df, dim, size):
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     w = eg.standard_wishart(df, dim, size)
     if size is not None:
@@ -378,7 +378,7 @@ def test_standard_wishart_reproduce(df, dim, size):
         assert w.ndim == 2
     assert w.shape[-2:] == (dim, dim)
 
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     w2 = eg.standard_wishart(df, dim, size)
     assert_allclose(w, w2)
@@ -388,7 +388,7 @@ def test_standard_wishart_reproduce(df, dim, size):
 @pytest.mark.parametrize("df", [8, [10], [[5], [6]]])
 def test_wishart_broadcast(df, scale_dim):
     dim = 5
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     scale = np.eye(dim)
     for _ in range(scale_dim):
@@ -402,7 +402,7 @@ def test_wishart_broadcast(df, scale_dim):
         assert w.shape[:-2] == np.broadcast(df, z).shape
 
     size = w.shape[:-2]
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     w2 = eg.wishart(df, scale, size=size)
     assert_allclose(w, w2)
@@ -417,7 +417,7 @@ def test_wishart_broadcast(df, scale_dim):
 def test_wishart_reduced_rank(method):
     scale = np.eye(3)
     scale[0, 1] = scale[1, 0] = 1.0
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     w = eg.wishart(10, scale, method=method, rank=2)
     assert w.shape == (3, 3)
@@ -428,7 +428,7 @@ def test_wishart_reduced_rank(method):
 def test_missing_scipy_exception():
     scale = np.eye(3)
     scale[0, 1] = scale[1, 0] = 1.0
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     with pytest.raises(ImportError):
         eg.wishart(10, scale, method="cholesky", rank=2)
@@ -490,7 +490,7 @@ def test_broadcast_both_paths():
 
 
 def test_factor_wishart():
-    pcg = PCG64(0)
+    pcg = PCG64(0, mode="sequence")
     eg = ExtendedGenerator(pcg)
     w = eg.wishart([3, 5], 2 * np.eye(4), size=(10000, 2), method="factor")
     assert_allclose(np.diag((w[:, 0] / 3).mean(0)).mean(), 4, rtol=1e-2)
@@ -648,36 +648,36 @@ def test_random_other_type(extended_gen):
         extended_gen.random(dtype=f16)
 
 
-def test_random(extended_gen_legacy):
-    extended_gen_legacy.bit_generator.seed(SEED)
-    actual = extended_gen_legacy.random((3, 2))
+def test_random():
+    random = ExtendedGenerator(MT19937(0, mode="sequence"))
+    actual = random.random((3, 2))
     desired = np.array(
         [
-            [0.61879477158567997, 0.59162362775974664],
-            [0.88868358904449662, 0.89165480011560816],
-            [0.4575674820298663, 0.7781880808593471],
+            [0.840045643478751, 0.526812305612705],
+            [0.390476663670696, 0.373221178199718],
+            [0.144160402211855, 0.255532529630851],
         ]
     )
     assert_array_almost_equal(actual, desired, decimal=15)
 
-    extended_gen_legacy.bit_generator.seed(SEED)
-    actual = extended_gen_legacy.random()
+    random = ExtendedGenerator(MT19937(0, mode="sequence"))
+    actual = random.random()
     assert_array_almost_equal(actual, desired[0, 0], decimal=15)
 
 
-def test_random_float(extended_gen_legacy):
-    extended_gen_legacy.bit_generator.seed(SEED)
-    actual = extended_gen_legacy.random((3, 2))
+def test_random_float(seed):
+    random = ExtendedGenerator(MT19937(seed, mode="sequence"))
+    actual = random.random((3, 2))
     desired = np.array(
-        [[0.6187948, 0.5916236], [0.8886836, 0.8916548], [0.4575675, 0.7781881]]
+        [[0.7165936, 0.6035045], [0.4473828, 0.359537], [0.7954794, 0.1942982]]
     )
     assert_array_almost_equal(actual, desired, decimal=7)
 
 
-def test_random_float_scalar(extended_gen_legacy):
-    extended_gen_legacy.bit_generator.seed(SEED)
-    actual = extended_gen_legacy.random(dtype=np.float32)
-    desired = 0.6187948
+def test_random_float_scalar(seed):
+    random = ExtendedGenerator(MT19937(seed, mode="sequence"))
+    actual = random.random(dtype=np.float32)
+    desired = 0.7165936
     assert_array_almost_equal(actual, desired, decimal=7)