Fixing PCovR to not mess with data shape

Author: cajchristian

examples/pcovr/PCovR_Regressors.py

@@ -29,7 +29,7 @@
 X_scaled = X_scaler.fit_transform(X)
 
 y_scaler = StandardScaler()
-y_scaled = y_scaler.fit_transform(y.reshape(-1, 1)).ravel()
+y_scaled = y_scaler.fit_transform(y.reshape(-1, 1))
 
 
 # %%

examples/pcovr/PCovR_Scaling.py

@@ -57,7 +57,7 @@
 
 pcovr_scaled = PCovR(mixing=0.5, n_components=4).fit(X_scaled, y_scaled)
 T_scaled = pcovr_scaled.transform(X_scaled)
-Yp_scaled = y_scaler.inverse_transform(pcovr_scaled.predict(X_scaled).reshape(-1, 1))
+Yp_scaled = y_scaler.inverse_transform(pcovr_scaled.predict(X_scaled))
 
 fig, ((ax1_T, ax2_T), (ax1_Y, ax2_Y)) = plt.subplots(2, 2, figsize=(8, 10))
 

src/skmatter/decomposition/_pcovr.py

@@ -310,7 +310,7 @@ def fit(self, X, Y, W=None):
             self._fit_sample_space(X, Y.reshape(Yhat.shape), Yhat, W)
 
         self.pxy_ = self.pxt_ @ self.pty_
-        if len(Y.shape) == 1 or Y.shape[1] == 1:
+        if len(Y.shape) == 1:
             self.pxy_ = self.pxy_.reshape(
                 X.shape[1],
             )
@@ -637,10 +637,6 @@ def score(self, X, y, T=None):
         """
         X, y = validate_data(self, X, y, reset=False)
 
-        if y.ndim == 2:
-            if y.shape[1] == 1:
-                y = y.ravel()
-
         if T is None:
             T = self.transform(X)
 

src/skmatter/utils/_pcovr_utils.py

@@ -41,12 +41,18 @@ def check_lr_fit(regressor, X, y):
         # Check compatibility with X
         validate_data(fitted_regressor, X, y, reset=False, multi_output=True)
 
-        if y.ndim == 2:
-            if y.shape[1] == 1:
-                y = y.ravel()
-
         # Check compatibility with y
+
+        # TO DO: This if statement is a band-aid for the case when we pass in a
+        # prefitted Ridge() or RidgeCV(), which, as of sklearn 1.6, will create
+        # coef_ with shape (n_features, ) even if fitted on a 2-D y with one target.
+        # In the future, we can optimize this block if LinearRegression() also changes.
+
         if fitted_regressor.coef_.ndim != y.ndim:
+            if y.ndim == 2:
+                if fitted_regressor.coef_.ndim == 1 and y.shape[1] == 1:
+                    return fitted_regressor
+
             raise ValueError(
                 "The regressor coefficients have a dimension incompatible with the "
                 "supplied target space. The coefficients have dimension "

tests/test_pcovr.py

@@ -111,15 +111,14 @@ def test_lr_with_t_errors(self):
         projection and that the prediction error increases with `mixing`.
         """
         prev_error = -1.0
-        Ytrue = self.Y.ravel()
 
         for mixing in np.linspace(0, 1, 11):
             pcovr = self.model(mixing=mixing, n_components=2, tol=1e-12)
             pcovr.fit(self.X, self.Y)
 
             T = pcovr.transform(self.X)
             Yp = pcovr.predict(T=T)
-            error = np.linalg.norm(Ytrue - Yp) ** 2.0 / np.linalg.norm(Ytrue) ** 2.0
+            error = np.linalg.norm(self.Y - Yp) ** 2.0 / np.linalg.norm(self.Y) ** 2.0
 
             with self.subTest(error=error):
                 self.assertFalse(np.isnan(error))
@@ -483,21 +482,22 @@ def test_none_regressor(self):
         self.assertTrue(pcovr.regressor_ is not None)
 
     def test_incompatible_coef_dim(self):
-        # self.Y is 2D with one target
+        # self.Y is 1D with one target
         # Don't need to test X shape, since this should
         # be caught by sklearn's validate_data
+        Y_2D = np.column_stack((self.Y, self.Y))
         regressor = Ridge(alpha=1e-8, fit_intercept=False, tol=1e-12)
-        regressor.fit(self.X, self.Y)
+        regressor.fit(self.X, Y_2D)
         pcovr = self.model(mixing=0.5, regressor=regressor)
 
         # Dimension mismatch
         with self.assertRaises(ValueError) as cm:
-            pcovr.fit(self.X, np.zeros((self.Y.shape[0], 2)))
+            pcovr.fit(self.X, self.Y)
         self.assertEqual(
             str(cm.exception),
             "The regressor coefficients have a dimension incompatible with the "
-            "supplied target space. The coefficients have dimension 1 and the targets "
-            "have dimension 2",
+            "supplied target space. The coefficients have dimension 2 and the targets "
+            "have dimension 1",
         )
 
     def test_incompatible_coef_shape(self):