Update app.py

app.py

@@ -1,124 +1,17 @@
-# from ultralytics import YOLO
-# import gradio as gr
-# import os
-# import numpy as np
-# from PIL import Image
-
-# # Load the trained model
-# model_path = "best.pt"  # Update this path to your best model
-# model = YOLO(model_path)
-
-# # Define class names
-# class_names = {
-#     0: "bacterial_leaf_blight",
-#     1: "brown_spot",
-#     2: "healthy",
-#     3: "leaf_blast",
-#     4: "leaf_scald",
-#     5: "narrow_brown_spot"
-# }
-
-# # Description for each disease
-# disease_info = {
-#     "bacterial_leaf_blight": "Bacterial Leaf Blight is caused by Xanthomonas oryzae. Water-soaked lesions along leaf margins that turn yellow and eventually grey as they expand.",
-#     "brown_spot": "Brown Spot is caused by Cochliobolus miyabeanus. Oval or round brown lesions with grey or whitish centers and a yellow halo.",
-#     "healthy": "This leaf shows no signs of disease and appears to be healthy.",
-#     "leaf_blast": "Leaf Blast is caused by Magnaporthe oryzae. Diamond-shaped lesions with a grey center and brown to red-brown borders.",
-#     "leaf_scald": "Leaf Scald is caused by Microdochium oryzae. Zonate lesions with alternating light and dark brown bands that run across the leaf.",
-#     "narrow_brown_spot": "Narrow Brown Spot is caused by Cercospora janseana. Narrow, long brown lesions parallel to the leaf veins."
-# }
-
-# # Treatment recommendations
-# treatment_recommendations = {
-#     "bacterial_leaf_blight": "• Use disease-free seeds\n• Apply copper-based bactericides\n• Maintain proper field drainage\n• Rotate crops with non-host plants",
-#     "brown_spot": "• Use fungicides containing azoxystrobin or propiconazole\n• Ensure balanced fertilization\n• Use resistant rice varieties\n• Proper water management",
-#     "healthy": "• Continue regular monitoring\n• Maintain balanced fertilization\n• Follow proper irrigation practices",
-#     "leaf_blast": "• Apply fungicides containing tricyclazole or azoxystrobin\n• Avoid excessive nitrogen fertilization\n• Use resistant rice varieties\n• Maintain proper spacing between plants",
-#     "leaf_scald": "• Apply fungicides containing propiconazole\n• Ensure adequate drainage\n• Use resistant varieties\n• Apply balanced fertilization",
-#     "narrow_brown_spot": "• Apply fungicides containing propiconazole or azoxystrobin\n• Maintain proper field sanitation\n• Use resistant varieties\n• Balanced fertilization"
-# }
-
-# def predict_image(image):
-#     if image is None:
-#         return {class_name: 0 for class_name in class_names.values()}, "", "", ""
-
-#     # Convert to RGB if the image has an alpha channel
-#     if image.shape[2] == 4:
-#         image_rgb = Image.fromarray(image).convert('RGB')
-#         image = np.array(image_rgb)
-
-#     # Run inference
-#     results = model(image)[0]
-
-#     # Get probabilities for all classes
-#     probs = results.probs.data.tolist()
-
-#     # Create a dictionary of class names and their probabilities
-#     class_probs = {class_names[i]: float(prob) for i, prob in enumerate(probs)}
-
-#     # Get the top prediction
-#     top_class_idx = results.probs.top1
-#     top_class_name = class_names[top_class_idx]
-#     confidence = results.probs.top1conf.item()
-
-#     # Get disease info and treatment
-#     info = disease_info[top_class_name]
-#     treatment = treatment_recommendations[top_class_name]
-
-#     # Create result text
-#     result_text = f"Prediction: {top_class_name}\nConfidence: {confidence:.2f}"
-
-#     return class_probs, result_text, info, treatment
-
-# # Create the Gradio interface
-# with gr.Blocks(title="Rice Leaf Disease Classifier") as app:
-#     gr.Markdown("# Rice Leaf Disease Classification")
-#     gr.Markdown("Upload an image of a rice leaf to identify potential diseases.")
-
-#     with gr.Row():
-#         with gr.Column():
-#             input_image = gr.Image(label="Upload Image", type="numpy")
-#             submit_btn = gr.Button("Analyze Leaf", variant="primary")
-
-#         with gr.Column():
-#             result_text = gr.Textbox(label="Prediction Result")
-#             label_probs = gr.Label(label="Disease Probabilities")
-#             disease_description = gr.Textbox(label="Disease Information", lines=4)
-#             treatment_info = gr.Textbox(label="Treatment Recommendations", lines=6)
-
-#     gr.Markdown("### Disease Categories")
-#     gr.Markdown("• Bacterial Leaf Blight\n• Brown Spot\n• Healthy\n• Leaf Blast\n• Leaf Scald\n• Narrow Brown Spot")
-
-#     gr.Markdown("### How to Use")
-#     gr.Markdown("1. Upload a clear image of a rice leaf\n2. Click 'Analyze Leaf'\n3. View the prediction results, disease information, and treatment recommendations")
-
-#     submit_btn.click(
-#         predict_image,
-#         inputs=[input_image],
-#         outputs=[label_probs, result_text, disease_description, treatment_info]
-#     )
-
-# # Launch the app
-# app.launch(share=True)
-
-
-
-
 from ultralytics import YOLO
 import gradio as gr
 import os
 import numpy as np
 from PIL import Image
-import torch
 
 # Load the trained model
-model_path = "best.pt"  # Update this path to your best model
+model_path = "best.pt"  # Ensure this file is uploaded to your Hugging Face Space
 model = YOLO(model_path)
 
 # Define class names
 class_names = {
     0: "bacterial_leaf_blight",
-    1: "brown_spot", 
+    1: "brown_spot",
     2: "healthy",
     3: "leaf_blast",
     4: "leaf_scald",
@@ -127,97 +20,57 @@ class_names = {
 
 # Description for each disease
 disease_info = {
-    "bacterial_leaf_blight": "Bacterial Leaf Blight is caused by Xanthomonas oryzae. Water-soaked lesions along leaf margins that turn yellow and eventually grey as they expand.",
-    "brown_spot": "Brown Spot is caused by Cochliobolus miyabeanus. Oval or round brown lesions with grey or whitish centers and a yellow halo.",
+    "bacterial_leaf_blight": "Bacterial Leaf Blight is caused by Xanthomonas oryzae...",
+    "brown_spot": "Brown Spot is caused by Cochliobolus miyabeanus...",
     "healthy": "This leaf shows no signs of disease and appears to be healthy.",
-    "leaf_blast": "Leaf Blast is caused by Magnaporthe oryzae. Diamond-shaped lesions with a grey center and brown to red-brown borders.",
-    "leaf_scald": "Leaf Scald is caused by Microdochium oryzae. Zonate lesions with alternating light and dark brown bands that run across the leaf.",
-    "narrow_brown_spot": "Narrow Brown Spot is caused by Cercospora janseana. Narrow, long brown lesions parallel to the leaf veins."
+    "leaf_blast": "Leaf Blast is caused by Magnaporthe oryzae...",
+    "leaf_scald": "Leaf Scald is caused by Microdochium oryzae...",
+    "narrow_brown_spot": "Narrow Brown Spot is caused by Cercospora janseana..."
 }
 
 # Treatment recommendations
 treatment_recommendations = {
-    "bacterial_leaf_blight": "• Use disease-free seeds\n• Apply copper-based bactericides\n• Maintain proper field drainage\n• Rotate crops with non-host plants",
-    "brown_spot": "• Use fungicides containing azoxystrobin or propiconazole\n• Ensure balanced fertilization\n• Use resistant rice varieties\n• Proper water management",
-    "healthy": "• Continue regular monitoring\n• Maintain balanced fertilization\n• Follow proper irrigation practices",
-    "leaf_blast": "• Apply fungicides containing tricyclazole or azoxystrobin\n• Avoid excessive nitrogen fertilization\n• Use resistant rice varieties\n• Maintain proper spacing between plants",
-    "leaf_scald": "• Apply fungicides containing propiconazole\n• Ensure adequate drainage\n• Use resistant varieties\n• Apply balanced fertilization",
-    "narrow_brown_spot": "• Apply fungicides containing propiconazole or azoxystrobin\n• Maintain proper field sanitation\n• Use resistant varieties\n• Balanced fertilization"
+    "bacterial_leaf_blight": "• Use disease-free seeds\n• Apply copper-based bactericides...",
+    "brown_spot": "• Use fungicides containing azoxystrobin or propiconazole...",
+    "healthy": "• Continue regular monitoring...",
+    "leaf_blast": "• Apply fungicides containing tricyclazole or azoxystrobin...",
+    "leaf_scald": "• Apply fungicides containing propiconazole...",
+    "narrow_brown_spot": "• Apply fungicides containing propiconazole or azoxystrobin..."
 }
 
 def predict_image(image):
-    # Default response for no image
-    default_response = {
-        'class_probs': {class_name: 0 for class_name in class_names.values()},
-        'result_text': "No image uploaded",
-        'info': "",
-        'treatment': ""
-    }
-
     if image is None:
-        return default_response['class_probs'], default_response['result_text'], default_response['info'], default_response['treatment']
-
-    try:
-        # Convert to RGB if the image has an alpha channel
-        if image.shape[2] == 4:
-            image_rgb = Image.fromarray(image).convert('RGB')
-            image = np.array(image_rgb)
-
-        # Ensure image is in the right format for YOLO
-        # Typically, YOLO expects images in RGB format and normalized
-        if not isinstance(image, torch.Tensor):
-            # Convert numpy array to torch tensor if needed
-            image = torch.from_numpy(image).permute(2, 0, 1).float() / 255.0
-            image = image.unsqueeze(0)  # Add batch dimension
-
-        # Run inference
-        results = model(image)
-
-        # Validate results
-        if not results:
-            return default_response['class_probs'], "No results from model", default_response['info'], default_response['treatment']
-
-        # Get results based on the model's output
-        if hasattr(results[0], 'probs'):
-            # Classification model
-            probs = results[0].probs.data.tolist()
-            class_probs = {class_names[i]: float(prob) for i, prob in enumerate(probs)}
-            
-            # Get top prediction
-            top_class_idx = results[0].probs.top1
-            top_class_name = class_names[top_class_idx]
-            confidence = results[0].probs.top1conf.item()
-
-        elif hasattr(results[0], 'boxes'):
-            # Detection model (fallback)
-            class_probs = {class_name: 0 for class_name in class_names.values()}
-            top_class_name = "Unable to classify"
-            confidence = 0
-        else:
-            return default_response['class_probs'], "Unexpected model output", default_response['info'], default_response['treatment']
-
-        # Get disease info and treatment
-        info = disease_info.get(top_class_name, "No information available")
-        treatment = treatment_recommendations.get(top_class_name, "No treatment recommendations")
-
-        # Create result text
-        result_text = f"Prediction: {top_class_name}\nConfidence: {confidence:.2f}"
-
-        return class_probs, result_text, info, treatment
-
-    except Exception as e:
-        # Comprehensive error handling
-        error_msg = f"Error in prediction: {str(e)}"
-        return (
-            default_response['class_probs'], 
-            error_msg, 
-            default_response['info'], 
-            default_response['treatment']
-        )
+        return {class_name: 0 for class_name in class_names.values()}, "", "", ""
+
+    # Convert to RGB if needed
+    if image.shape[2] == 4:
+        image_rgb = Image.fromarray(image).convert('RGB')
+        image = np.array(image_rgb)
+
+    # Run inference
+    results = model(image)[0]
+
+    # Get probabilities
+    probs = results.probs.data.tolist()
+    class_probs = {class_names[i]: float(prob) for i, prob in enumerate(probs)}
+
+    # Get top prediction
+    top_class_idx = results.probs.top1
+    top_class_name = class_names[top_class_idx]
+    confidence = results.probs.top1conf.item()
+
+    # Get disease info and treatment
+    info = disease_info[top_class_name]
+    treatment = treatment_recommendations[top_class_name]
+
+    # Create result text
+    result_text = f"Prediction: {top_class_name}\nConfidence: {confidence:.2f}"
+
+    return class_probs, result_text, info, treatment
 
 # Create the Gradio interface
 with gr.Blocks(title="Rice Leaf Disease Classifier") as app:
-    gr.Markdown("# Rice Leaf Disease Classification")
+    gr.Markdown("# 🌱 Rice Leaf Disease Classification")
     gr.Markdown("Upload an image of a rice leaf to identify potential diseases.")
 
     with gr.Row():
@@ -244,5 +97,4 @@ with gr.Blocks(title="Rice Leaf Disease Classifier") as app:
     )
 
 # Launch the app
-if __name__ == "__main__":
-    app.launch(share=False)
+app.launch(server_name="0.0.0.0")