Fix multiclass PRC test text removal for matplotlib compatibility

In test_multiclass_probability_with_class_labels, fix text annotation
removal by calling remove() on the Text object itself rather than trying
to modify the ArtistList collection. This works with matplotlib's current
Artist management system.

Before:
    oz.ax.texts.remove(child)

After:
    child.remove()

This fixes the AttributeError: 'ArtistList' object has no attribute 'remove'
error and subsequent TypeError from attempted list slice assignment, while
maintaining the test's original functionality of cleaning up ISO F1 curve
annotations before image comparison.

Author: lwgray

tests/baseline_images/test_classifier/test_prcurve/test_multiclass_probability_with_class_labels.png

@@ -324,7 +324,7 @@ def test_multiclass_probability_with_class_labels(self):
         # Will not check for these as they appears okay in other test images.
         for child in oz.ax.get_children():
             if isinstance(child, matplotlib.text.Annotation):
-                oz.ax.texts.remove(child)
+                child.remove()
 
         # Compare the images
         tol = (
@@ -438,9 +438,11 @@ def test_quick_method_with_test_set(self):
 
         viz = precision_recall_curve(
             RandomForestClassifier(random_state=72),
-            X_train, y_train,
-            X_test, y_test,
-            show=False
+            X_train,
+            y_train,
+            X_test,
+            y_test,
+            show=False,
         )
         self.assert_images_similar(viz)
 
@@ -487,24 +489,31 @@ def test_within_pipeline(self):
         classes = ["unoccupied", "occupied"]
 
         X_train, X_test, y_train, y_test = tts(
-                    X, y, test_size=0.2, shuffle=True, random_state=42
-                )
-
-        model = Pipeline([
-            ('minmax', MinMaxScaler()), 
-            ('prc', PrecisionRecallCurve(SVC(random_state=42),
-                                            per_class=True,
-                                            micro=False,
-                                            fill_area=False,
-                                            iso_f1_curves=True,
-                                            ap_score=False,
-                                            classes=classes))
-        ])
+            X, y, test_size=0.2, shuffle=True, random_state=42
+        )
+
+        model = Pipeline(
+            [
+                ("minmax", MinMaxScaler()),
+                (
+                    "prc",
+                    PrecisionRecallCurve(
+                        SVC(random_state=42),
+                        per_class=True,
+                        micro=False,
+                        fill_area=False,
+                        iso_f1_curves=True,
+                        ap_score=False,
+                        classes=classes,
+                    ),
+                ),
+            ]
+        )
 
         model.fit(X_train, y_train)
         model.score(X_test, y_test)
-        model['prc'].finalize()
-        self.assert_images_similar(model['prc'], tol=5.5)
+        model["prc"].finalize()
+        self.assert_images_similar(model["prc"], tol=5.5)
 
     def test_within_pipeline_quickmethod(self):
         """
@@ -514,22 +523,32 @@ def test_within_pipeline_quickmethod(self):
         X, y = load_occupancy(return_dataset=True).to_pandas()
 
         X_train, X_test, y_train, y_test = tts(
-                    X, y, test_size=0.2, shuffle=True, random_state=42
-                )
-
-        model = Pipeline([
-            ('minmax', MinMaxScaler()), 
-            ('prc', precision_recall_curve(SVC(random_state=42),
-                                            X_train, y_train, X_test, y_test,
-                                            per_class=True,
-                                            micro=False,
-                                            fill_area=False,
-                                            iso_f1_curves=True,
-                                            ap_score=False,
-                                            classes=["unoccupied", "occupied"],
-                                            show=False))
-            ])
-        self.assert_images_similar(model['prc'], tol=5.5)
+            X, y, test_size=0.2, shuffle=True, random_state=42
+        )
+
+        model = Pipeline(
+            [
+                ("minmax", MinMaxScaler()),
+                (
+                    "prc",
+                    precision_recall_curve(
+                        SVC(random_state=42),
+                        X_train,
+                        y_train,
+                        X_test,
+                        y_test,
+                        per_class=True,
+                        micro=False,
+                        fill_area=False,
+                        iso_f1_curves=True,
+                        ap_score=False,
+                        classes=["unoccupied", "occupied"],
+                        show=False,
+                    ),
+                ),
+            ]
+        )
+        self.assert_images_similar(model["prc"], tol=5.5)
 
     def test_pipeline_as_model_input(self):
         """
@@ -539,21 +558,20 @@ def test_pipeline_as_model_input(self):
         classes = ["unoccupied", "occupied"]
 
         X_train, X_test, y_train, y_test = tts(
-                    X, y, test_size=0.2, shuffle=True, random_state=42
-                )
-
-        model = Pipeline([
-            ('minmax', MinMaxScaler()), 
-            ('svc', SVC(random_state=42))
-        ])
-
-        oz = PrecisionRecallCurve(model,
-                                  per_class=True,
-                                  micro=False,
-                                  fill_area=False,
-                                  iso_f1_curves=True,
-                                  ap_score=False,
-                                  classes=classes)
+            X, y, test_size=0.2, shuffle=True, random_state=42
+        )
+
+        model = Pipeline([("minmax", MinMaxScaler()), ("svc", SVC(random_state=42))])
+
+        oz = PrecisionRecallCurve(
+            model,
+            per_class=True,
+            micro=False,
+            fill_area=False,
+            iso_f1_curves=True,
+            ap_score=False,
+            classes=classes,
+        )
         oz.fit(X_train, y_train)
         oz.score(X_test, y_test)
         oz.finalize()
@@ -567,20 +585,23 @@ def test_pipeline_as_model_input_quickmethod(self):
         X, y = load_occupancy(return_dataset=True).to_pandas()
 
         X_train, X_test, y_train, y_test = tts(
-                    X, y, test_size=0.2, shuffle=True, random_state=42
-                )
-
-        model = Pipeline([
-            ('minmax', MinMaxScaler()), 
-            ('svc', SVC(random_state=42))
-        ])
-
-        oz = precision_recall_curve(model, X_train, y_train, X_test, y_test, 
-                     per_class=True,
-                     micro=False,
-                     fill_area=False,
-                     iso_f1_curves=True,
-                     ap_score=False,
-                     classes=["unoccupied", "occupied"],
-                     show=False)
-        self.assert_images_similar(oz, tol=5.5)
+            X, y, test_size=0.2, shuffle=True, random_state=42
+        )
+
+        model = Pipeline([("minmax", MinMaxScaler()), ("svc", SVC(random_state=42))])
+
+        oz = precision_recall_curve(
+            model,
+            X_train,
+            y_train,
+            X_test,
+            y_test,
+            per_class=True,
+            micro=False,
+            fill_area=False,
+            iso_f1_curves=True,
+            ap_score=False,
+            classes=["unoccupied", "occupied"],
+            show=False,
+        )
+        self.assert_images_similar(oz, tol=5.5)