import streamlit as st
import numpy as np
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.preprocessing.text import Tokenizer

# Constants
MAX_LENGTH = 100
TOKENIZER_PATH = "tokenizer.json"  # Assuming the tokenizer is saved separately.

# Load pre-trained model
@st.cache_resource
def load_trained_model():
    return load_model("deep_learning_model.h5")

# Load tokenizer
@st.cache_resource
def load_tokenizer():
    import json
    from tensorflow.keras.preprocessing.text import tokenizer_from_json
    with open(TOKENIZER_PATH, "r") as f:
        tokenizer_data = json.load(f)
    return tokenizer_from_json(tokenizer_data)

# Preprocessing function
def preprocess_prompt(prompt, tokenizer, max_length):
    sequence = tokenizer.texts_to_sequences([prompt])
    padded_sequence = pad_sequences(sequence, maxlen=max_length)
    return padded_sequence

# Predict function
def detect_prompt(prompt, model, tokenizer, max_length):
    processed_prompt = preprocess_prompt(prompt, tokenizer, max_length)
    prediction = model.predict(processed_prompt)[0][0]
    class_label = "Malicious" if prediction >= 0.5 else "Valid"
    confidence_score = prediction * 100 if prediction >= 0.5 else (1 - prediction) * 100
    return class_label, confidence_score

# Streamlit App
st.title("Prompt Injection Detection App")
st.write("Detect and prevent prompt injection attacks using a deep learning model.")

# Load model and tokenizer
model = load_trained_model()
tokenizer = load_tokenizer()

# Input Section
user_input = st.text_area("Enter a prompt to test:", "")
if st.button("Detect"):
    if user_input:
        label, confidence = detect_prompt(user_input, model, tokenizer, MAX_LENGTH)
        st.write(f"**Predicted Class:** {label}")
        st.write(f"**Confidence Score:** {confidence:.2f}%")
    else:
        st.warning("Please enter a prompt to test.")
        
import os

if st.button("Train Model"):
    os.system("python train_model.py")
    st.success("Model training complete. Saved as deep_learning_model.h5")

if not os.path.exists("deep_learning_model.h5"):
    st.info("Training the model for the first time...")
    os.system("python train_model.py")
    st.success("Model trained successfully and saved as deep_learning_model.h5")