add status bar

app/main.py

@@ -174,8 +174,16 @@ if uploaded_file:
             # Dynamically create the Pydantic model for validation
             ClassificationOutput = generate_classification_model(list(label_descriptions.keys()))
 
+            # Create a placeholder for the progress bar
+            progress_bar = st.progress(0)
+            progress_text = st.empty()
+
             # Function to classify a single row
-            def classify_row(row):
+            def classify_row(row, index, total_rows):
+                # Update progress bar
+                progress_bar.progress((index + 1) / total_rows)
+                progress_text.text(f"Processing row {index + 1}/{total_rows}...")
+
                 # Generate system and user prompts
                 system_prompt, user_prompt = generate_prompts(
                     row=row.to_dict(),
@@ -206,25 +214,33 @@ if uploaded_file:
                 )
 
             # Apply the classification to each row in the limited data
-            limited_data[prediction_column] = limited_data.apply(classify_row, axis=1)
+            total_rows = len(limited_data)
+            predictions = []
+
+            for index, row in limited_data.iterrows():
+                prediction = classify_row(row, index, total_rows)
+                predictions.append(prediction)
+
+            # Add predictions to the DataFrame
+            limited_data[prediction_column] = predictions
+
+            # Reset progress bar and text
+            progress_bar.empty()
+            progress_text.empty()
 
             # Display Predictions
             st.write(f"### Predictions ({prediction_column})", limited_data)
 
+            # Evaluate if ground truth is available
             if label_column in limited_data.columns:
-                # Evaluation Mode: Target column exists
                 from utils.evaluation import evaluate_predictions
                 report = evaluate_predictions(limited_data[label_column], limited_data[prediction_column])
                 st.write("### Evaluation Metrics")
                 display_metrics_as_table(report)
             else:
-                # Inference Mode: No target column
-                st.warning("Inference mode: No target column provided, so no evaluation metrics are available.")
-
+                st.warning(f"Inference mode: No target column provided, so no evaluation metrics are available.")
                 # Count predictions
                 label_counts = limited_data[prediction_column].value_counts().reset_index()
                 label_counts.columns = ["Label", "Count"]
-
-                # Display prediction stats
                 st.subheader("Prediction Statistics")
-                st.table(label_counts)
+                st.table(label_counts)