Update app.py

app.py

@@ -6,41 +6,48 @@ import google.generativeai as genai
 import os
 import markdown2
 
-# Load TensorFlow model & configure Gemini API
+# Load TensorFlow model
 model = tf.saved_model.load('model')
-genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
-
-# Disease labels & inline prompt generation for AI response
 labels = ['cataract', 'diabetic_retinopathy', 'glaucoma', 'normal']
-prompt_map = lambda disease: (
-    "Provide a congratulatory message for healthy eyes with tips for maintenance." if disease == "normal"
-    else f"Diagnosis: {disease}\nDescription, causes, and prevention advice for {disease}."
-)
 
-# Function for AI-based disease explanation
-def generate_explanation(disease):
+# Configure Gemini API
+genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
+
+# Generate AI-based explanation for the predicted disease
+def get_disease_detail(disease):
+    prompt = (
+        "Create a text congratulating on healthy eyes with tips to keep them healthy."
+        if disease == "normal" else
+        f"Diagnosis: {disease}\n\n"
+        f"What is {disease}?\nCauses and suggestions to prevent {disease}."
+    )
     try:
-        response = genai.GenerativeModel("gemini-1.5-flash").generate_content(prompt_map(disease))
-        return markdown2.markdown(response.text.strip()) if response and response.text else "No response available."
+        response = genai.GenerativeModel("gemini-1.5-flash").generate_content(prompt)
+        return markdown2.markdown(response.text.strip() if response and response.text else "No response.")
     except Exception as e:
         return f"Error: {e}"
 
-# Image processing & prediction
+# Process and predict uploaded image
 def predict_image(image):
-    img_array = np.expand_dims(np.array(image.resize((224, 224))) / 255.0, axis=0).astype(np.float32)
-    predictions = model.signatures['serving_default'](tf.convert_to_tensor(img_array))['output_0']
+    img_array = np.expand_dims(np.array(image.resize((224, 224))).astype(np.float32) / 255.0, axis=0)
+    predictions = model.signatures['serving_default'](tf.convert_to_tensor(img_array, dtype=tf.float32))['output_0']
 
     top_label = labels[np.argmax(predictions.numpy())]
-    return {top_label: predictions.numpy().max()}, generate_explanation(top_label)
+    explanation = get_disease_detail(top_label)
+
+    return {top_label: predictions.numpy().max()}, explanation
+
+# Example images
+example_images = [[f"exp_eye_images/{img}"] for img in ["0_right_h.png", "03fd50da928d_dr.png", "108_right_h.png", "1062_right_c.png", "1084_right_c.png", "image_1002_g.jpg"]]
 
-# Gradio Interface with minimalist style
+# Gradio Interface
 interface = gr.Interface(
     fn=predict_image,
     inputs=gr.Image(type="pil"),
     outputs=[gr.Label(num_top_classes=1, label="Prediction"), gr.HTML(label="Explanation", elem_classes=["scrollable-html"])],
-    examples=[[f"exp_eye_images/{img}"] for img in ["0_right_h.png", "03fd50da928d_dr.png", "108_right_h.png", "1062_right_c.png", "1084_right_c.png", "image_1002_g.jpg"]],
+    examples=example_images,
     title="DR Predictor",
-    description="Upload an eye fundus image to receive a condition prediction. *For educational use only*.",
+    description=("Upload an eye fundus image, and the model predicts the condition. This model is for educational use only."),
     allow_flagging="never",
     css=".scrollable-html {height: 206px; overflow-y: auto; border: 1px solid #ccc; padding: 10px; box-sizing: border-box;}"
 )