Merge pull request #577 from Labelbox/ms/classification-metric-cleanup

update classification metric name

Author: msokoloff1

labelbox/data/metrics/confusion_matrix/confusion_matrix.py

@@ -40,9 +40,8 @@ def confusion_matrix_metric(ground_truths: List[Union[
     if value is None:
         return []
 
-    return [
-        ConfusionMatrixMetric(metric_name=f"{int(iou*100)}pct_iou", value=value)
-    ]
+    metric_name = _get_metric_name(ground_truths, predictions, iou)
+    return [ConfusionMatrixMetric(metric_name=metric_name, value=value)]
 
 
 def feature_confusion_matrix_metric(
@@ -74,8 +73,34 @@ def feature_confusion_matrix_metric(
                                          include_subclasses, iou)
         if value is None:
             continue
+
+        metric_name = _get_metric_name(annotation_pairs[key][0],
+                                       annotation_pairs[key][1], iou)
         metrics.append(
-            ConfusionMatrixMetric(metric_name=f"{int(iou*100)}pct_iou",
+            ConfusionMatrixMetric(metric_name=metric_name,
                                   feature_name=key,
                                   value=value))
     return metrics
+
+
+def _get_metric_name(ground_truths: List[Union[ObjectAnnotation,
+                                               ClassificationAnnotation]],
+                     predictions: List[Union[ObjectAnnotation,
+                                             ClassificationAnnotation]],
+                     iou: float):
+
+    if _is_classification(ground_truths, predictions):
+        return "classification"
+
+    return f"{int(iou*100)}pct_iou"
+
+
+def _is_classification(ground_truths: List[Union[ObjectAnnotation,
+                                                 ClassificationAnnotation]],
+                       predictions: List[Union[ObjectAnnotation,
+                                               ClassificationAnnotation]]):
+    # Check if either the prediction or label contains a classification annotation
+    return (len(predictions) and
+            isinstance(predictions[0], ClassificationAnnotation) or
+            len(ground_truths) and
+            isinstance(ground_truths[0], ClassificationAnnotation))