malware_app.py

import streamlit as st
import numpy as np
from tensorflow.keras.models import load_model

# Load the trained model
@st.cache_resource
def load_trained_model():
    model = load_model("phi3_model.h5")
    return model

# Function to preprocess the input
def preprocess_input(timestamp, severity):
    severity_mapping = {"Low": 0, "Medium": 1, "High": 2}
    severity_encoded = severity_mapping.get(severity, 0)
    
    # Prepare the input data as a numpy array
    input_data = np.array([[timestamp, severity_encoded]])
    return input_data

# Main function to run the Streamlit app
def main():
    st.title("Malware Event Prediction")

    # Input fields
    timestamp = st.slider("Timestamp (Hour)", min_value=0, max_value=23, value=12)
    severity = st.selectbox("Severity", ["Low", "Medium", "High"])

    # Predict button
    if st.button("Predict"):
        # Load the model
        model = load_trained_model()
        
        # Preprocess the input
        input_data = preprocess_input(timestamp, severity
        
        # Make prediction
        prediction = model.predict(input_data)[0][0]
        prediction_label = "Unusual" if prediction > 0.5 else "Normal"
        
        # Display the result
        st.write(f"Prediction: {prediction_label}")

if __name__ == "__main__":
    main()