Reformat test_pca.py with black==19.10b0

Author: hristog

tests/test_pca.py

@@ -218,11 +218,7 @@ def test_whitening():
 
 
 def assert_fit_results_almost_equal(
-    pca_baseline,
-    pca,
-    ev_decimal=1,
-    evr_decimal=3,
-    nv_decimal=3,
+    pca_baseline, pca, ev_decimal=1, evr_decimal=3, nv_decimal=3,
 ):
     assert_array_almost_equal(
         pca_baseline.explained_variance_, pca.explained_variance_, decimal=ev_decimal
@@ -252,9 +248,7 @@ def assert_explained_and_empirical_var_almost_eq(X, dX_mean_0, apca, rpca):
 
     X_rpca = rpca.transform(dX_mean_0)
     assert_array_almost_equal(
-        rpca.explained_variance_,
-        np.var(X_rpca, ddof=1, axis=0),
-        decimal=1,
+        rpca.explained_variance_, np.var(X_rpca, ddof=1, axis=0), decimal=1,
     )
     assert_array_almost_equal(rpca.explained_variance_, empirical_variances, decimal=1)
 
@@ -272,21 +266,15 @@ def test_no_centering():
 
     # Correlated data.
     X_corr = datasets.make_classification(
-        n_samples,
-        n_features,
-        n_informative=n_features - 2,
-        random_state=rng,
+        n_samples, n_features, n_informative=n_features - 2, random_state=rng,
     )[0]
     dX_corr = da.from_array(X_corr, chunks=(50, n_features))
     dX_corr_mean_0 = dX_corr - dX_corr.mean(axis=0)
 
     # Test fitting pseudo-random data.
     pca = sd.PCA(n_components=n_components, svd_solver="full", random_state=0).fit(X)
     apca = dd.PCA(
-        n_components=n_components,
-        svd_solver="full",
-        random_state=0,
-        center=False,
+        n_components=n_components, svd_solver="full", random_state=0, center=False,
     ).fit(dX_mean_0)
 
     assert_fit_results_almost_equal(pca, apca)
@@ -306,10 +294,7 @@ def test_no_centering():
         X_corr
     )
     apca_corr = dd.PCA(
-        n_components=n_components,
-        svd_solver="full",
-        random_state=0,
-        center=False,
+        n_components=n_components, svd_solver="full", random_state=0, center=False,
     ).fit(dX_corr_mean_0)
 
     assert_fit_results_almost_equal(pca_corr, apca_corr)
@@ -342,12 +327,9 @@ def test_inverse_transform_no_centering():
     dX = da.from_array(X, chunks=(n // 2, p))
     dX_mean_0 = dX - dX.mean(axis=0)
 
-    pca = dd.PCA(
-        n_components=2,
-        svd_solver="full",
-        random_state=0,
-        center=False,
-    ).fit(dX_mean_0)
+    pca = dd.PCA(n_components=2, svd_solver="full", random_state=0, center=False,).fit(
+        dX_mean_0
+    )
 
     # Test inverse transformation of artificial data, with strongly expressed mean.
     Y = pca.transform(dX_mean_0)
@@ -356,11 +338,7 @@ def test_inverse_transform_no_centering():
 
     # As above, but with whitening.
     pca = dd.PCA(
-        n_components=2,
-        svd_solver="full",
-        random_state=0,
-        whiten=True,
-        center=False,
+        n_components=2, svd_solver="full", random_state=0, whiten=True, center=False,
     ).fit(dX_mean_0)
 
     Y = pca.transform(dX_mean_0)
@@ -374,26 +352,17 @@ def test_sample_scoring_no_centering():
     n_features = 80
 
     X = datasets.make_classification(
-        n_samples,
-        n_features,
-        n_informative=n_features - 2,
-        random_state=rng,
+        n_samples, n_features, n_informative=n_features - 2, random_state=rng,
     )[0]
     dX = da.from_array(X, chunks=(50, n_features))
     dX_mean_0 = dX - dX.mean(axis=0)
 
-    pca = dd.PCA(
-        n_components=2,
-        svd_solver="full",
-        random_state=0,
-        center=True,
-    ).fit(dX)
+    pca = dd.PCA(n_components=2, svd_solver="full", random_state=0, center=True,).fit(
+        dX
+    )
 
     pca_mean_0 = dd.PCA(
-        n_components=2,
-        svd_solver="full",
-        random_state=0,
-        center=False,
+        n_components=2, svd_solver="full", random_state=0, center=False,
     ).fit(dX_mean_0)
 
     assert_almost_equal(pca.score(dX), pca_mean_0.score(dX_mean_0), decimal=6)