Update app.py

app.py

@@ -1,93 +1,87 @@
+import h5py
 import gradio as gr
-from tensorflow.keras.utils import img_to_array,load_img
+from tensorflow.keras.utils import img_to_array, load_img
 from keras.models import load_model
 import numpy as np
+from deep_translator import GoogleTranslator
 
 # Load the pre-trained model from the local path
 model_path = 'peach.h5'
-model = load_model(model_path)  # Load the model here
-
-def predict_disease(image_file, model, all_labels):
 
+# Check if the model is loading correctly
+try:
+    with h5py.File(model_path, 'r+') as f:
+        if 'groups' in f.attrs['model_config']:  
+            model_config_string = f.attrs['model_config'] 
+            model_config_string = model_config_string.replace('"groups": 1,', '')
+            model_config_string = model_config_string.replace('"groups": 1}', '}')
+            f.attrs['model_config'] = model_config_string.encode('utf-8')
+
+    model = load_model(model_path)
+    print("Model loaded successfully.")
+except Exception as e:
+    print(f"Error loading model: {e}")
+
+def predict_disease(image_file, model, all_labels, target_language):
     try:
         # Load and preprocess the image
-        img = load_img(image_file, target_size=(224, 224))  # Use load_img from tensorflow.keras.utils
+        print(f"Received image file: {image_file}")
+        img = load_img(image_file, target_size=(224, 224))  # Ensure image size matches model input
         img_array = img_to_array(img)
         img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
         img_array = img_array / 255.0  # Normalize the image
 
         # Predict the class
-        predictions = model.predict(img_array)  # Use the loaded model here
+        predictions = model.predict(img_array)
         predicted_class = np.argmax(predictions[0])
-
+        
         # Get the predicted class label
         predicted_label = all_labels[predicted_class]
-
-        # Print the predicted label to the console
-
-        if predicted_label=='Peach Healthy':
-            predicted_label = predicted_label = """<h3 align="center">Peach Healthy</h3><br><br>
-          <center>No need use Pesticides</center>"""
-        elif predicted_label=='Peach Bacterial Spot':
-            predicted_label = """
-            <style>
-                      li{
-                        font-size: 15px;
-                        margin-left: 90px;
-                        margin-top: 15px;
-                        margin-bottom: 15px;
-                      }
-                      h4{
-                        font-size: 17px;
-                        margin-top: 15px;
-                      }
-                      h4:hover{
-                        cursor: pointer;
-                      }
-
-                      h3:hover{
-                        cursor: pointer;
-                        color: blue;
-                        transform: scale(1.3);
-                      }
-                      .note{
-                        text-align: center;
-                        font-size: 16px;
-                      }
-                      p{
-                        font-size: 13px;
-                        text-align: center;
-                      }
-
-                </style>
-                <h3><center><b>Peach Bacterial Spot</b></center></h3>
-                <h4>PESTICIDES TO BE USED:</h4>
-                <ul>
+        
+        # Translate the predicted label to the selected language
+        translated_label = GoogleTranslator(source='en', target=target_language).translate(predicted_label)
+        
+        # Provide pesticide information based on the predicted label
+        if predicted_label == 'Peach Bacterial Spot':
+            pesticide_info = """
+                <h2><center><b>Peach Bacterial Spot</b></center></h2>
+                <h4>PESTICIDES TO BE USED:</h4><br>
+                
+                <ul style="font-size:17px;margin-left:40px;">
                     <li>1. Copper oxychloride (Kocide)</li>
                     <li>2. Streptomycin (Streptomycin sulfate)</li>
                     <li>3. Tetracycline (Agrimycin)</li>
                     <li>4. Oxytetracycline (Terramycin)</li>
-
-
-                </ul>
-                <p class="note"><b>* * * IMPORTANT NOTE * * *</b></p>
-                <p>Be sure to follow local regulations and guidelines for application</p>
-
-
+                </ul><br>
+                <center><p class="note" style="font-size:15px;"><b>* * * IMPORTANT NOTE * * *</b></p></center><br>
+                <center><p style="font-size:13px;">Be sure to follow local regulations and guidelines for application</p></center>
             """
-
-
-
+        
+        
+            
+        elif predicted_label == 'Peach Healthy':
+            pesticide_info = """<h2><center><b>Peach Healthy</b></center></h2>
+            <h5> No pesticides needed"""
+        
         else:
-          predicted_label = """<h3 align="center">Choose Correct image</h3><br><br>
-          """
+            pesticide_info = 'No pesticide information available.'
+
+        print(f"Pesticide Info (Before Translation): {pesticide_info}")
 
-        return predicted_label
+        # Translate the pesticide information to the selected language
+        translated_pesticide_info = GoogleTranslator(source='en', target=target_language).translate(pesticide_info)
+        print(f"Translated Pesticide Info: {translated_pesticide_info}")
 
+        # Return translated label and pesticide information with associated styling
+        predicted_label_html = f"""
+        
+        {translated_pesticide_info}
+        """
+        return predicted_label_html
 
     except Exception as e:
-        print(f"Error: {e}")
-        return None
+        print(f"Error during prediction: {e}")
+        return f"<h3>Error: {e}</h3>"
 
 # List of class labels
 all_labels = [
@@ -95,18 +89,42 @@ all_labels = [
     'Peach Bacterial Spot'
 ]
 
+# Language codes and their full names (display full names in dropdown)
+language_choices = {
+    'hi': 'Hindi',
+    'te': 'Telugu',
+    'en': 'English',
+    'ml': 'Malayalam',
+    'ta': 'Tamil',
+    'bn': 'Bengali',
+    'gu': 'Gujarati',
+    'kn': 'Kannada',
+    'mr': 'Marathi'
+}
+
+# Mapping full names back to their corresponding language code
+full_to_code = {value: key for key, value in language_choices.items()}
+
+# Create a dropdown of full language names, using the full name in the UI
+languages = list(language_choices.values())  # List of full language names
+
 # Define the Gradio interface
-def gradio_predict(image_file):
-    return predict_disease(image_file, model, all_labels)  # Pass the model to the function
+def gradio_predict(image_file, target_language):
+    # Map full name back to language code for translation
+    language_code = full_to_code.get(target_language, 'en')
+    return predict_disease(image_file, model, all_labels, language_code)
 
-# Create a Gradio interface
+# Create the Gradio interface
 gr_interface = gr.Interface(
-    fn=gradio_predict,  # Function to call for predictions
-    inputs=gr.Image(type="filepath"),  # Upload image as file path
-    outputs="html",  # Output will be the class label as text
+    fn=gradio_predict,
+    inputs=[
+        gr.Image(type="filepath"),  # Image input for disease prediction
+        gr.Dropdown(label="Select language", choices=languages, value='English')  # Language selection dropdown with full names
+    ],
+    outputs="html",  # Output will be in HTML (translated text)
     title="Peach Disease Predictor",
-    description="Upload an image of a plant to predict the disease.",
+    description="Upload an image of a plant to predict the disease and get the translated label and pesticide information in the selected language."
 )
 
 # Launch the Gradio app
-gr_interface.launch(debug=True)
+gr_interface.launch()