dump the bar plot

app.py

@@ -21,41 +21,26 @@ classifier = pipeline(
 )
 logger.info("Pipeline initialized successfully")
 
-# Default DataFrame for BarPlot
-emotion_labels = [pred['label'] for pred in classifier("test input")]
-default_scores = [0] * len(emotion_labels)
-default_df = pd.DataFrame({'label': emotion_labels, 'score': default_scores})
-
-# Error DataFrame for BarPlot
-error_df = pd.DataFrame({'label': ['error'], 'score': [1.0]})
-
 def predict_emotion(text):
     """Predict emotions from text and return formatted results."""
     try:
         if not text:
             logger.warning("Empty text received")
-            # Return default DataFrame directly
-            return default_df
-        
+            return {}
+
         # Get predictions and handle result structure
         logger.info("Running prediction...")
         predictions = classifier(text)
         logger.debug(f"Raw predictions output: {predictions}")
         
-        # Ensure predictions are a list of dictionaries
-        if isinstance(predictions, list) and len(predictions) > 0 and all(isinstance(item, dict) for item in predictions):
-            df = pd.DataFrame(predictions)  # Convert list of dictionaries to DataFrame
-        else:
-            logger.error(f"Unexpected predictions format: {predictions}")
-            return error_df
-
-        logger.info(f"Processed scores:\n{df}")
-        return df
+        # Process predictions into a dict for display
+        scores = {item['label']: item['score'] for item in predictions}
+        logger.info(f"Processed scores: {scores}")
+        return scores
         
     except Exception as e:
         logger.error(f"Error in prediction: {str(e)}")
-        # Return error DataFrame directly
-        return error_df
+        return {"error": "An error occurred during emotion prediction"}
 
 # Create the Gradio interface
 demo = gr.Interface(
@@ -65,16 +50,8 @@ demo = gr.Interface(
         label="Input Text",
         lines=4
     ),
-    outputs=gr.BarPlot(
-        value=default_df,
-        x="label",
-        y="score",
-        title="Emotion Probabilities",
-        color=["#ff6f61", "#6b5b95", "#88b04b", "#f7cac9", "#92a8d1", "#f7786b"],  # Dynamic colors
-        height=400,
-        vertical=True
-    ),
-    title="Emotion Detection with DistilBERT",
+    outputs=gr.JSON(),  # Display the scores in a JSON format
+    title="CREATIVE MACHINES: Emotion Detection with DistilBERT",
     description="This app uses the DistilBERT model fine-tuned for emotion detection. Enter any text to analyze its emotional content.",
     examples=[
         "I am so happy to see you!",