Add param validation to PCA classes (dask#808)

Author: hristog

dask_ml/decomposition/incremental_pca.py

@@ -134,11 +134,6 @@ def __init__(
         iterated_power=0,
         random_state=None,
     ):
-        if center is False:
-            raise NotImplementedError(
-                "IncrementalPCA with center=False is not supported."
-            )
-
         self.n_components = n_components
         self.whiten = whiten
         self.center = center
@@ -148,6 +143,11 @@ def __init__(
         self.iterated_power = iterated_power
         self.random_state = random_state
 
+    def _check_params(self):
+        super()._check_params()
+        if self.center is False:
+            raise ValueError("IncrementalPCA with center=False is not supported.")
+
     def _fit(self, X, y=None):
         """Fit the model with X, using minibatches of size batch_size.
 
@@ -245,6 +245,7 @@ def partial_fit(self, X, y=None, check_input=True):
         self : object
             Returns the instance itself.
         """
+        self._check_params()
         if check_input:
             if sparse.issparse(X):
                 raise TypeError(

dask_ml/decomposition/pca.py

@@ -1,5 +1,4 @@
 import numbers
-import warnings
 
 import dask
 import dask.array as da
@@ -222,16 +221,14 @@ def __init__(
         self.iterated_power = iterated_power
         self.random_state = random_state
 
-        if whiten and not center:
-            warnings.warn(
-                "Whitening requires centering. Please, ensure that your data "
-                "is already centered, in order to avoid unexpected behavior.",
-                RuntimeWarning,
-            )
+    def _check_params(self):
+        pass
 
     def fit(self, X, y=None):
         if not dask.is_dask_collection(X):
             raise TypeError(_TYPE_MSG.format(type(X)))
+
+        self._check_params()
         self._fit(X)
         self.n_features_in_ = X.shape[1]
         return self

tests/test_incremental_pca.py

@@ -478,5 +478,17 @@ def test_incremental_pca_partial_fit_float_division():
 
 
 def test_incremental_pca_no_centering_not_supported():
-    with pytest.raises(NotImplementedError, match="not supported"):
-        IncrementalPCA(n_components=2, center=False)
+    rng = np.random.RandomState(0)
+    A = rng.randn(5, 3) + 2
+    A = da.from_array(A, chunks=[3, -1])
+
+    pca = IncrementalPCA(n_components=2, center=False)
+
+    with pytest.raises(ValueError, match="not supported"):
+        pca.partial_fit(A)
+
+    with pytest.raises(ValueError, match="not supported"):
+        pca.fit(A)
+
+    with pytest.raises(ValueError, match="not supported"):
+        pca.fit_transform(A)

tests/test_pca.py

@@ -183,19 +183,15 @@ def test_whitening():
 
         # Test no centering.
         X_mean_0 = dX_mean_0.copy()
-        with pytest.warns(RuntimeWarning) as record:
-            pca = dd.PCA(
-                n_components=n_components,
-                whiten=True,
-                center=False,
-                copy=copy,
-                svd_solver=solver,
-                random_state=0,
-                iterated_power=4,
-            )
-
-        assert len(record) == 1
-        assert "Whitening requires centering" in record[0].message.args[0]
+        pca = dd.PCA(
+            n_components=n_components,
+            whiten=True,
+            center=False,
+            copy=copy,
+            svd_solver=solver,
+            random_state=0,
+            iterated_power=4,
+        )
 
         X_whitened = pca.fit_transform(X_mean_0.copy())
         assert X_whitened.shape == (n_samples, n_components)
@@ -358,19 +354,14 @@ def test_inverse_transform_no_centering():
     Y_inverse = pca.inverse_transform(Y)
     assert_almost_equal(dX_mean_0.compute(), Y_inverse, decimal=3)
 
-    # Capture warning about employing whitening without centering.
-    with pytest.warns(RuntimeWarning) as record:
-        # As above, but with whitening.
-        pca = dd.PCA(
-            n_components=2,
-            svd_solver="full",
-            random_state=0,
-            whiten=True,
-            center=False,
-        ).fit(dX_mean_0)
-
-    assert len(record) == 1
-    assert "Whitening requires centering" in record[0].message.args[0]
+    # As above, but with whitening.
+    pca = dd.PCA(
+        n_components=2,
+        svd_solver="full",
+        random_state=0,
+        whiten=True,
+        center=False,
+    ).fit(dX_mean_0)
 
     Y = pca.transform(dX_mean_0)
     Y_inverse = pca.inverse_transform(Y)
@@ -818,15 +809,10 @@ def test_pca_score2():
         ll2 = pca.score(dX)
         assert ll1 > ll2
 
-        # Capture warning about employing whitening without centering.
-        with pytest.warns(RuntimeWarning) as record:
-            apca = dd.PCA(n_components=2, whiten=True, center=False, svd_solver=solver)
-            apca.fit(dX_mean_0)
-            all2 = apca.score(dX_mean_0)
-            assert all1 > all2
-
-        assert len(record) == 1
-        assert "Whitening requires centering" in record[0].message.args[0]
+        apca = dd.PCA(n_components=2, whiten=True, center=False, svd_solver=solver)
+        apca.fit(dX_mean_0)
+        all2 = apca.score(dX_mean_0)
+        assert all1 > all2
 
 
 def test_pca_score3():