test(Evaluation): added edge case support of non-complete data for co…

…nfusion matrix

tests/test_evaluation.py

@@ -49,8 +49,7 @@ def setUpClass(self):
             class_index = np.random.randint(0, 4)
             self.labels_ohe[i][class_index] = 1
         # Create predictions
-        for i in range(0, 50):
-            self.preds = np.random.rand(50, 4)
+        self.preds = np.random.rand(50, 4)
 
         # Create imaging data indices
         self.sample_list = []
@@ -100,17 +99,17 @@ def test_evaluate_fitting_standard_tl(self):
         self.assertTrue(os.path.exists(os.path.join(self.tmp_plot.name,
                                       "plot.fitting_course.standard.png")))
 
-    # def test_evaluate_fitting_standard_tl_noft(self):
-    #     # Create artificial history data - transfer learning
-    #     hist_standard_tl = {"tl_loss": [], "tl_val_loss": []}
-    #     for i in range(0, 150):
-    #         hist_standard_tl["tl_loss"].append(random.uniform(0, 1))
-    #         hist_standard_tl["tl_val_loss"].append(random.uniform(0, 1))
-    #     # Apply fitting evaluation
-    #     evaluate_fitting(hist_standard_tl, out_path=self.tmp_plot.name,
-    #                     monitor=["loss"], suffix="standard")
-    #     self.assertTrue(os.path.exists(os.path.join(self.tmp_plot.name,
-    #                                   "plot.fitting_course.standard.png")))
+    def test_evaluate_fitting_standard_tl_noft(self):
+        # Create artificial history data - transfer learning
+        hist_standard_tl = {"tl_loss": [], "tl_val_loss": []}
+        for i in range(0, 150):
+            hist_standard_tl["tl_loss"].append(random.uniform(0, 1))
+            hist_standard_tl["tl_val_loss"].append(random.uniform(0, 1))
+        # Apply fitting evaluation
+        evaluate_fitting(hist_standard_tl, out_path=self.tmp_plot.name,
+                        monitor=["loss"], suffix="standard")
+        self.assertTrue(os.path.exists(os.path.join(self.tmp_plot.name,
+                                      "plot.fitting_course.standard.png")))
 
     def test_evaluate_fitting_advanced(self):
         # Create artificial history data - advanced
@@ -304,6 +303,33 @@ def test_evaluate_performance_confusionmatrix(self):
                                  "plot.performance.confusion_matrix.test.png")
         self.assertTrue(os.path.exists(path_plot))
 
+    def test_evaluate_performance_confusionmatrix_edgecases(self):
+        # Create classification labels
+        labels_ohe = np.zeros((50, 4), dtype=np.uint8)
+        for i in range(0, 50):
+            class_index = np.random.randint(0, 3)
+            labels_ohe[i][class_index] = 1
+        # Confusion Mat
+        evaluate_performance(self.preds, labels_ohe,
+                             out_path=self.tmp_plot.name,
+                             multi_label=False,
+                             store_csv=False,
+                             plot_barplot=False,
+                             plot_confusion_matrix=True,
+                             plot_roc_curve=False)
+        # Create predictions
+        preds = np.random.rand(50, 4)
+        for i in range(0, 50):
+            preds[i][3] = 0.0
+        # Confusion Mat
+        evaluate_performance(preds, self.labels_ohe,
+                             out_path=self.tmp_plot.name,
+                             multi_label=False,
+                             store_csv=False,
+                             plot_barplot=False,
+                             plot_confusion_matrix=True,
+                             plot_roc_curve=False)
+
     def test_evaluate_performance_roc(self):
         evaluate_performance(self.preds, self.labels_ohe,
                              out_path=self.tmp_plot.name,