Enhance prediction function with detailed logging and error handling for improved user feedback

utils/prediction.py

@@ -73,24 +73,30 @@ def predict_sentence(model, sentence, tokenizer, label_encoder):
     """
     Make prediction for a single sentence with label validation.
     """
+    import time
+    start_time = time.time()
+    
     # Validation checks
+    st.write("🔄 Starting prediction process...")
     if model is None:
-        print("Error: Model not loaded")
+        st.error("Error: Model not loaded")
         return "Error: Model not loaded", 0.0
     if tokenizer is None:
-        print("Error: Tokenizer not loaded")
+        st.error("Error: Tokenizer not loaded")
         return "Error: Tokenizer not loaded", 0.0
     if label_encoder is None:
-        print("Error: Label encoder not loaded")
+        st.error("Error: Label encoder not loaded")
         return "Error: Label encoder not loaded", 0.0
         
     # Force CPU device
+    st.write("⚙️ Preparing model...")
     device = torch.device('cpu')
     model = model.to(device)
     model.eval()
     
     # Tokenize
     try:
+        st.write(f"📝 Processing text: {sentence[:50]}...")
         encoding = tokenizer(
             sentence,
             add_special_tokens=True,
@@ -100,15 +106,19 @@ def predict_sentence(model, sentence, tokenizer, label_encoder):
             return_tensors='pt'
         ).to(device)
         
+        st.write("🤖 Running model inference...")
         with torch.no_grad():
             outputs = model(encoding['input_ids'], encoding['attention_mask'])
             probabilities = torch.softmax(outputs, dim=1)
             prob, pred_idx = torch.max(probabilities, dim=1)
             predicted_label = label_encoder.classes_[pred_idx.item()]
-            return predicted_label, prob.item()
+            
+        elapsed_time = time.time() - start_time
+        st.write(f"✅ Prediction completed in {elapsed_time:.2f} seconds")
+        return predicted_label, prob.item()
             
     except Exception as e:
-        print(f"Prediction error: {str(e)}")
+        st.error(f"❌ Prediction error: {str(e)}")
         return f"Error: {str(e)}", 0.0
     
 def print_labels(label_encoder, show_counts=False):