Update app.py

app.py

@@ -13,16 +13,16 @@ if torch.cuda.is_available():
     model.cuda()
 
 validation_results = json.load(open('validation_results.json'))
-scores, thresholds, precisions = validation_results['scores'], validation_results['thresholds'], validation_results['precisions']
+scores, thresholds, precisions, recalls = validation_results['scores'], validation_results['thresholds'], validation_results['precisions'], validation_results['recalls']
 
 def get_threshold_precision(score_):
-    for score, threshold, precision in zip(scores, thresholds, precisions):
+    for score, threshold, precision, recall in zip(scores, thresholds, precisions, recalls):
         if score_ < score:
             break
-        prev_score, prev_threshold, prev_precision = score, threshold, precision
+        prev_score, prev_threshold, prev_precision, prev_recall = score, threshold, precision, recall
     if prev_threshold == prev_score:
         prev_threshold = score_
-    return prev_threshold, prev_precision
+    return prev_threshold, prev_precision, prev_recall
 
 def normalize_spaces(text):
     return re.sub(r'\s+', ' ', text).strip()
@@ -57,9 +57,9 @@ def predict(title, authors, abstract):
     #selected = [d for d in validation_data if d['score'] >= score]
     #true_positives = sum(1 for d in selected if d['label'] == 1)
     #precision = true_positives / len(selected) if selected else 0
-    threshold, precision = get_threshold_precision(score)
+    threshold, precision, recall = get_threshold_precision(score)
     
-    result = f"Your score: {score:.2f}.\nFor papers with score >= {threshold:.2f}, {precision * 100:.2f}% are selected by AK."
+    result = f"Your score: {score:.2f}.\nFor papers with score>={threshold:.2f}, {precision * 100:.2f}% are selected by AK.\nFor papers selected by AK, {recall * 100:.2f}% have score>={threshold:.2f}"
     
     return score, result