Merge remote-tracking branch 'origin/main'

metrics/calculator.py

@@ -167,7 +167,7 @@ def F1(self):
         """
 
         precision = self.metrics['precision']
-        recall = self.metrics['recall']
+        recall = self.TPR(log=False)
         if precision + recall == 0:
             f1 = 0
         else:
@@ -207,17 +207,21 @@ def TPR(self, log=True):
             tpr = 0
         else:
             tpr = self.metrics['TP'] / (self.metrics['TP'] + self.metrics['FN'])
+
         if log:
             self.log(f"{self.tool}: TPR: ", tpr)
         return tpr
 
-    def FNR(self):
+    def FNR(self) -> float:
         """
-        FNR = 1- TPR
+        FNR = FN / (FN + TP)
         prints the false negative rate of a given tool
-        :return: float
         """
-        fnr = 1 - self.TPR(log=False)
+        try:
+            fnr = self.metrics["FN"] / (self.metrics["FN"] + self.metrics["TP"])
+        except ZeroDivisionError:
+            fnr = 0
+
         self.log(f"{self.tool}: FNR: ", fnr)
         return fnr
 
@@ -262,14 +266,11 @@ def calc_all_metrics(self) -> Dict[str, float]:
             self.FNR,
             self.TPR,
             self.TNR,
-            self.recall,
             self.precision,
             self.F1,
             self.accuracy,
             self.MCC,
         ):
             res.update({metric.__name__ : metric()})
         return res
-    # def __del__(self):
-        # if hasattr(self, 'db'):
-        #     self.db.close()
+

scripts/slips_metrics_getter.py

@@ -11,8 +11,8 @@
 from argparse import ArgumentParser
 
 from plot.plot import Plot
-from scripts.extracted_levels import extracted_threat_levels
-from scripts.extracted_gt_tw_labels import gt_tw_labels
+from scripts.extracted_scores.extracted_levels import extracted_threat_levels
+from scripts.extracted_scores.extracted_gt_tw_labels import gt_tw_labels
 from metrics.calculator import Calculator
 
 
@@ -69,7 +69,7 @@ def print_extremes(threshold_with_min_max: Dict[str, Dict]):
     """
     Print the extreme values for each metric and their corresponding thresholds.
     """
-    print("Best/Worst thresholds so far")
+    print("Below are the minimum and maximum of the 4 error metrics with the threshold that resulted in them.")
     for metric, info in threshold_with_min_max.items():
         print(f"{metric}:")
         print(f"  Min value: {info['min_value']}, Threshold: {info['min_threshold']}")
@@ -232,14 +232,13 @@ def main():
             # calc.calc_all_metrics()
             experiment_metrics = {
                     'MCC': calc.MCC(),
-                    'recall': calc.recall(),
                     'precision': calc.precision(),
-                    'F1': calc.F1(),
                     'FPR': calc.FPR(),
                     'TPR': calc.TPR(),
                     'FNR': calc.FNR(),
                     'TNR': calc.TNR(),
                     'accuracy': calc.accuracy(),
+                    'F1': calc.F1(),
                 }
             experiment_metrics.update(confusion_matrix)
             metrics[threshold].update({exp: experiment_metrics})