Update app.py

app.py

@@ -10,8 +10,12 @@ model.eval()
 
 # Sigmoid-based threshold adjustment function
 def calculate_threshold(base_relevance, min_threshold=0.02, max_threshold=0.5, k=10):
-    return min_threshold + (max_threshold - min_threshold) * (1 / (1 + torch.exp(-k * (base_relevance - 0.5))))
-
+    base_relevance_tensor = torch.tensor(base_relevance)  # Ensure input is a tensor
+    threshold = min_threshold + (max_threshold - min_threshold) * (
+        1 / (1 + torch.exp(-k * (base_relevance_tensor - 0.5)))
+    )
+    return threshold.item()  # Convert tensor back to float for use in other functions
+    
 # Function to compute relevance score and dynamically adjust threshold
 def get_relevance_score_and_excerpt(query, paragraph, threshold_weight):
     if not query.strip() or not paragraph.strip():