app.py

import streamlit as st
import tensorflow as tf
import numpy as np
from PIL import Image
import os

# Page configuration
st.set_page_config(
    page_title="Pneumonia Detection AI",
    page_icon="🫁",
    layout="wide"
)

# Custom CSS
st.markdown("""
<style>
    .main-header {
        text-align: center;
        color: #1f77b4;
        margin-bottom: 2rem;
    }
    .result-box {
        padding: 1rem;
        border-radius: 10px;
        margin: 1rem 0;
        text-align: center;
    }
    .pneumonia-result {
        background-color: #f8d7da;
        border: 2px solid #dc3545;
        color: #721c24;
    }
    .normal-result {
        background-color: #d4edda;
        border: 2px solid #28a745;
        color: #155724;
    }
</style>
""", unsafe_allow_html=True)

# Header
st.markdown('<h1 class="main-header">🫁 Pneumonia Detection AI</h1>', unsafe_allow_html=True)
st.markdown('<p style="text-align: center; font-size: 1.2rem; color: #666;">Upload a chest X-ray image for AI-powered pneumonia detection</p>', unsafe_allow_html=True)

@st.cache_resource
def load_model():
    """Load the Keras model - much more reliable!"""
    try:
        model_path = 'Fast_VGG19_3epoch.keras'  # Changed to .keras format!
        
        if os.path.exists(model_path):
            model = tf.keras.models.load_model(model_path)
            st.success("✅ Keras model loaded successfully!")
            return model
        else:
            st.error(f"Model file not found: {model_path}")
            return None
            
    except Exception as e:
        st.error(f"Error loading model: {str(e)}")
        return None

def preprocess_image(image):
    """Preprocess image for model prediction"""
    try:
        img = np.array(image.convert('RGB'))
        img = tf.image.resize(img, [64, 64])
        img = tf.cast(img, tf.float32) / 255.0
        
        mean = [0.485, 0.456, 0.406]
        std = [0.229, 0.224, 0.225]
        img = (img - mean) / std
        
        return tf.expand_dims(img, 0)
    
    except Exception as e:
        st.error(f"Error preprocessing image: {str(e)}")
        return None

# Sidebar information
with st.sidebar:
    st.header("📊 Model Information")
    st.info("""
    **Architecture:** VGG19 Transfer Learning
    **Accuracy:** 77.0%
    **AUC Score:** 0.850
    **Format:** Keras (.keras) ✅
    **TensorFlow:** 2.15.0
    """)
    
    st.header("📋 How to Use")
    st.markdown("""
    1. **Upload Image:** Click browse to select X-ray
    2. **Supported Formats:** JPG, JPEG, PNG
    3. **Click Analyze:** Get AI prediction
    4. **Review Results:** Check confidence levels
    """)

# Load model
model = load_model()

if model is not None:
    uploaded_file = st.file_uploader(
        "Choose a chest X-ray image",
        type=['jpg', 'jpeg', 'png'],
        help="Upload a chest X-ray image in JPG, JPEG, or PNG format"
    )
    
    if uploaded_file is not None:
        col1, col2 = st.columns([1, 1])
        
        with col1:
            st.subheader("📸 Uploaded Image")
            image = Image.open(uploaded_file)
            st.image(image, caption="Chest X-ray Image", use_column_width=True)
            
            st.write(f"**Filename:** {uploaded_file.name}")
            st.write(f"**Size:** {image.size}")
        
        with col2:
            st.subheader("🤖 AI Analysis")
            
            if st.button("🔍 Analyze X-ray", type="primary", use_container_width=True):
                with st.spinner("🧠 AI is analyzing the X-ray..."):
                    processed_img = preprocess_image(image)
                    
                    if processed_img is not None:
                        try:
                            prediction = model.predict(processed_img, verbose=0)[0][0]
                            
                            if prediction > 0.5:
                                confidence = prediction * 100
                                pneumonia_html = f"""
                                <div class="result-box pneumonia-result">
                                    <h3>🚨 PNEUMONIA DETECTED</h3>
                                    <h4>Confidence: {confidence:.1f}%</h4>
                                </div>
                                """
                                st.markdown(pneumonia_html, unsafe_allow_html=True)
                                
                                if confidence >= 80:
                                    st.error("🚨 **HIGH CONFIDENCE** - Immediate medical attention recommended")
                                elif confidence >= 60:
                                    st.warning("⚠️ **MODERATE CONFIDENCE** - Medical consultation advised")
                                else:
                                    st.info("ℹ️ **LOW CONFIDENCE** - Consider additional testing")
                            
                            else:
                                confidence = (1 - prediction) * 100
                                normal_html = f"""
                                <div class="result-box normal-result">
                                    <h3>✅ NORMAL</h3>
                                    <h4>Confidence: {confidence:.1f}%</h4>
                                </div>
                                """
                                st.markdown(normal_html, unsafe_allow_html=True)
                                
                                if confidence >= 80:
                                    st.success("✅ **HIGH CONFIDENCE** - X-ray appears normal")
                                elif confidence >= 60:
                                    st.info("ℹ️ **MODERATE CONFIDENCE** - Likely normal")
                                else:
                                    st.warning("⚠️ **LOW CONFIDENCE** - Review recommended")
                            
                            # Detailed probabilities
                            st.subheader("📊 Detailed Analysis")
                            
                            normal_prob = (1 - prediction) * 100
                            pneumonia_prob = prediction * 100
                            
                            st.write("**Probability Breakdown:**")
                            st.progress(normal_prob/100, text=f"Normal: {normal_prob:.1f}%")
                            st.progress(pneumonia_prob/100, text=f"Pneumonia: {pneumonia_prob:.1f}%")
                            
                            col_a, col_b = st.columns(2)
                            with col_a:
                                st.metric("Normal", f"{normal_prob:.1f}%")
                            with col_b:
                                st.metric("Pneumonia", f"{pneumonia_prob:.1f}%")
                        
                        except Exception as e:
                            st.error(f"❌ Prediction failed: {str(e)}")
                    else:
                        st.error("❌ Failed to process image")
    
    else:
        st.info("👆 Please upload a chest X-ray image to begin analysis")

else:
    st.error("❌ Failed to load the AI model. Please check the deployment.")

# Footer
footer_html = """
<div style="text-align: center; padding: 2rem; background-color: #f8f9fa; border-radius: 10px; margin-top: 2rem;">
    <h4>⚠️ Medical Disclaimer</h4>
    <p>This AI tool is for <strong>educational and research purposes only</strong>. 
    It should not be used as a substitute for professional medical advice, diagnosis, or treatment. 
    Always seek the advice of qualified healthcare providers with any questions regarding medical conditions.</p>
    
    <hr style="margin: 1rem 0;">
    
    <p><em>🐳 Powered by Docker • TensorFlow 2.15.0 • Streamlit</em></p>
    <p><em>🤖 VGG19 Transfer Learning Model (.keras format)</em></p>
</div>
"""

st.markdown("---")
st.markdown(footer_html, unsafe_allow_html=True)