Upload folder using huggingface_hub

app.py

@@ -197,19 +197,55 @@ def predict_ripeness(audio, image, model_path):
         else:
             return "Error: Failed to process inputs. Please check the debug logs."
         
-        # Format the result
+        # Format the result with a range display
         if ripeness is not None:
-            result = f"Predicted Ripeness: {ripeness.item():.2f}/13"
+            ripeness_value = ripeness.item()
             
-            # Add a qualitative description
-            if ripeness.item() < 9:
-                result += "\n\nThis watermelon is not very ripe. You might want to choose another one."
-            elif ripeness.item() < 10:
-                result += "\n\nThis watermelon has moderate ripeness."
-            elif ripeness.item() < 11:
-                result += "\n\nThis watermelon is ripe! A good choice."
+            # Create a header with the numerical result
+            result = f"🍉 Predicted Ripeness Score: {ripeness_value:.2f}/13 🍉\n\n"
+            
+            # Add ripeness scale visualization
+            result += "Ripeness Scale based on Sugar Content:\n"
+            result += "──────────────────────────────────\n"
+            
+            # Create the scale display
+            scale_ranges = [
+                (0, 8, "Underripe"),
+                (8, 9, "Slightly Ripe"),
+                (9, 10, "Moderately Ripe"),
+                (10, 11, "Ripe"),
+                (11, 13, "Very Ripe")
+            ]
+            
+            # Find which category the prediction falls into
+            user_category = None
+            for min_val, max_val, category_name in scale_ranges:
+                if min_val <= ripeness_value < max_val:
+                    user_category = category_name
+                    break
+            if ripeness_value >= scale_ranges[-1][0]:  # Handle edge case
+                user_category = scale_ranges[-1][2]
+            
+            # Display the scale with the user's result highlighted
+            for min_val, max_val, category_name in scale_ranges:
+                if category_name == user_category:
+                    result += f"▶ {min_val}-{max_val}: {category_name} ◀ (YOUR WATERMELON)\n"
+                else:
+                    result += f"  {min_val}-{max_val}: {category_name}\n"
+            
+            result += "──────────────────────────────────\n\n"
+            
+            # Add assessment of the watermelon's ripeness
+            if ripeness_value < 8:
+                result += "Assessment: This watermelon is underripe. It may not have developed full flavor yet."
+            elif ripeness_value < 9:
+                result += "Assessment: This watermelon is slightly ripe. You might want to wait a few more days."
+            elif ripeness_value < 10:
+                result += "Assessment: This watermelon has moderate ripeness. It should have decent flavor."
+            elif ripeness_value < 11:
+                result += "Assessment: This watermelon is properly ripe! It should be sweet and juicy."
             else:
-                result += "\n\nThis watermelon is perfectly ripe! Excellent choice!"
+                result += "Assessment: This watermelon is perfectly ripe! Excellent choice for maximum sweetness and flavor."
                 
             return result
         else:
@@ -250,7 +286,7 @@ def create_app(model_path):
                 submit_btn = gr.Button("Predict Ripeness", variant="primary")
             
             with gr.Column():
-                output = gr.Textbox(label="Prediction Results", lines=6)
+                output = gr.Textbox(label="Prediction Results", lines=12)
                 
         submit_btn.click(
             fn=predict_fn,
@@ -259,6 +295,14 @@ def create_app(model_path):
         )
         
         gr.Markdown("""
+        ## How it works
+        
+        The app uses a deep learning model that combines:
+        - Audio analysis using MFCC features and LSTM neural network
+        - Image analysis using ResNet-50 convolutional neural network
+        
+        The model evaluates watermelons on a scale from 0-13, where higher numbers indicate greater ripeness.
+        
         ## Tips for best results
         - For audio: Tap the watermelon with your knuckle and record the sound
         - For image: Take a clear photo of the whole watermelon in good lighting