app.py

# import libraries
import streamlit as st
import predict
import logging

# Logging configuration
logging.basicConfig(level=logging.DEBUG, format='%(asctime)s - %(levelname)s - %(message)s')

# Load necessary components
df = predict.load_dataset()
model, tokenizer = predict.load_model_and_tokenizer()

if df is not None and model is not None and tokenizer is not None:
    # Preprocess dataset and prepare stopwords and stemmer
    manual_stopwords = ["di", "ke", "dari", "yang", "dan", "atau", "dengan", "untuk", "ini", "itu", "aja", "saja", "lah", "bri", "brimo", "aplikasi", "rekening", "coba", "yg", "ke", "untuk", "nya", "saya", "dia", "dan", "sangat", "video", "login", "apk", "jadi", "akun", "malah", "uang", "banget", "dalam", "atm", "padahal"]
    stop_words = set(predict.stopwords.words('indonesian'))
    stop_words.update(manual_stopwords)
    factory = predict.StemmerFactory()
    stemmer = factory.create_stemmer()
    df['label'] = df['score'].apply(predict.map_labels)

    navigation = st.sidebar.selectbox('Choose Page:', ('predictor'))

    if navigation == 'predictor':
        predict.run(model, tokenizer, stop_words, stemmer)
    # elif navigation == 'eda':
    #     eda.run()
else:
    if df is None:
        logging.error("Failed to load dataset.")
        st.error("Failed to load the dataset. Please check the dataset file.")
    if model is None or tokenizer is None:
        logging.error("Failed to load model or tokenizer.")
        st.error("Failed to load the model or tokenizer. Please check the model file.")