app.py

import streamlit as st
import pickle
import string
import nltk
from nltk.corpus import stopwords
from nltk.stem.porter import PorterStemmer

nltk.download('stopwords')
nltk.download('punkt')

# Initialize the PorterStemmer
ps = PorterStemmer()

# Load models and resources
def load_resources():
    tfidf = pickle.load(open('vectorizer.pkl', 'rb'))
    model = pickle.load(open('model.pkl', 'rb'))
    return tfidf, model

# Text preprocessing function
def transform_text(text):
    text = text.lower()
    tokens = nltk.word_tokenize(text)

    # Remove non-alphanumeric tokens and stopwords, and apply stemming
    filtered_tokens = [ps.stem(word) for word in tokens if word.isalnum() and word not in stopwords.words('english')]
    return " ".join(filtered_tokens)

# Predict whether a message is spam or not
def predict_spam(input_text, tfidf, model):
    transformed_text = transform_text(input_text)
    vector_input = tfidf.transform([transformed_text])
    result = model.predict(vector_input)[0]
    return result

# Display result in Streamlit
def display_prediction(result):
    if result == "spam":
        st.success("This is spam 🚫")
    elif result == "ham":
        st.success("This is not spam 👍")

# Main Streamlit app function
def main():
    # Load resources
    tfidf, model = load_resources()

    # Set the app title
    st.title("Email/SMS Spam Classifier")

    # Input text area for user message
    input_sms = st.text_area("Enter your message here:")

    # Placeholder for prediction result
    prediction_placeholder = st.empty()

    # Predict button
    if st.button('Predict'):
        if input_sms.strip() == "":
            prediction_placeholder.markdown(
                "<h3 style='color: #f24b4b; font-size: 1.75rem;'>Please enter a message first ⚠️</h3>",
                unsafe_allow_html=True)
        else:
            result = predict_spam(input_sms, tfidf, model)
            with prediction_placeholder:
                display_prediction(result)

# Run the app
if __name__ == "__main__":
    main()