Fixing decision_function shape for PCovC, writing corresponding Z test

Author: rvasav26

src/skmatter/decomposition/pcovc_new.py

@@ -288,20 +288,19 @@ class likelihoods, :math:`{\mathbf{Z}}`.
             else:
                 W = self.z_classifier_.coef_.T.reshape(X.shape[1], -1)
                 Z = self.z_classifier_.decision_function(X).reshape(X.shape[0], -1) 
-                #computes Z this will throw an error since pxz and ptz aren't defined yet
 
         else:
-            Z = y.copy()  
+            #Z = y.copy()  
+            Z = X @ W
             if W is None:
                 W = np.linalg.lstsq(X, Z, self.tol)[0]  #W = weights for Pxz
         # print("Z: "+str(Z[:4]))
         # print("W: "+str(W[:4]))
         self._label_binarizer = LabelBinarizer(neg_label=-1, pos_label=1)
         Y = self._label_binarizer.fit_transform(y) #check if we need this
-
         if not self._label_binarizer.y_type_.startswith("multilabel"):
             y = column_or_1d(y, warn=True)
-  
+
         if self.space_ == "feature":
             self._fit_feature_space(X, Y.reshape(Z.shape), Z)
         else:
@@ -332,10 +331,9 @@ class likelihoods, :math:`{\mathbf{Z}}`.
             self.ptz_ = np.hstack(
                 [est_.coef_.T for est_ in self.classifier_.estimators_]
             ) 
-
             self.pxz_ = self.pxt_ @ self.ptz_
         else:
-            self.ptz_ = self.classifier_.coef_.T #this is actually of shape (n_features, 1) when we have binary classification, but we need it to be shape (n_features, n_classes)
+            self.ptz_ = self.classifier_.coef_.T 
             self.pxz_ = self.pxt_ @ self.ptz_ 
 
         if len(Y.shape) == 1:
@@ -425,10 +423,16 @@ def decision_function(self, X=None, T=None):
 
         if X is not None:
             X = check_array(X)
-            return X @ self.pxz_
+            scores = X @ self.pxz_ 
         else:
             T = check_array(T)
-            return T @ self.ptz_
+            scores = T @ self.ptz_ 
+        
+        return (
+                np.reshape(scores, (-1, ))
+                if (scores.ndim > 1 and scores.shape[1] == 1)
+                else scores
+        )  
 
     def predict(self, X=None, T=None):
         """Predicts the property labels using classification on T."""

tests/test_kernel_pcovc.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 from sklearn import exceptions
+from sklearn.calibration import LinearSVC
 from sklearn.datasets import load_breast_cancer as get_dataset
 from sklearn.kernel_ridge import KernelRidge
 from sklearn.linear_model import Ridge, RidgeCV
@@ -193,6 +194,8 @@ def test_centerer(self):
 
     def test_prefit_classifier(self):
         classifier = SVC(kernel="rbf", gamma=0.1)
+        #this fails since we are trying to call decision_function(K) on a classifier fitted with X
+        #see line 340 of kernel_pcovr
         classifier.fit(self.X, self.Y)
         print(classifier.n_features_in_)
         kpcovc = self.model(mixing=0.5, classifier=classifier, kernel="rbf", gamma=0.1)
@@ -336,7 +339,7 @@ def test_linear_matches_pcovc(self):
         """Check that KernelPCovC returns the same results as PCovC when using a linear
         kernel.
         """
-        svc = SVC(kernel="linear", gamma="scale", coef0=0)
+        svc = LinearSVC()
         svc.fit(self.X, self.Y)
 
         # common instantiation parameters for the two models

tests/test_pcovc.py

@@ -416,6 +416,32 @@ def test_T_shape(self):
         T = pcovc.transform(self.X)
         self.assertTrue(check_X_y(self.X, T, multi_output=True))
         self.assertTrue(T.shape[-1] == n_components)
+    
+    def test_Z_shape(self):
+        """Check that PCovC returns an evidence matrix consistent with the shape of the input
+         matrix and the number of classes.
+        """
+        n_components = 5
+        pcovc = self.model(n_components=n_components, tol=1e-12)
+        
+        pcovc.fit(self.X, self.Y)
+
+        # Shape (n_samples, ) for binary classifcation
+        Z = pcovc.decision_function(self.X) 
+        
+        self.assertTrue(Z.ndim == 1)
+        self.assertTrue(Z.shape[0] == self.X.shape[0])
+
+        Y_multiclass = self.Y.copy()
+        Y_multiclass[0] = 2
+
+        pcovc.fit(self.X, Y_multiclass)
+
+        # Shape (n_samples, n_classes) for multiclass classification
+        Z = pcovc.decision_function(self.X) 
+        
+        self.assertTrue(Z.ndim == 2)
+        self.assertTrue(Z.shape[0] == self.X.shape[0])
 
     def test_default_ncomponents(self):
         pcovc = PCovC(mixing=0.5)
@@ -451,6 +477,7 @@ def test_prefit_classifier(self):
     def test_prefit_classification(self):
         classifier = LogisticRegression()
         classifier.fit(self.X, self.Y)
+        #Yhat = classifier.predict(self.X)
         Yhat = classifier.predict(self.X)
         W = classifier.coef_.reshape(self.X.shape[1], -1)
         pcovc1 = self.model(mixing=0.5, classifier="precomputed", n_components=1)
@@ -525,7 +552,7 @@ def test_incompatible_coef_shape(self):
         #     "The coefficients have dimension %d and the targets "
         #     "have dimension %d" % (classifier.coef_.ndim, self.Y.squeeze().ndim),
         # )
-
+        
         with self.assertRaises(ValueError) as cm:
             pcovc.fit(self.X, np.column_stack((self.Y, self.Y)))
         self.assertEqual(