config.py

# Đường dẫn file
DATA_DIR = 'dataset'
MODEL_DIR = 'models'
CSV_FILE = 'enron_spam_data.csv'
DEFAULT_MODEL_FILE = 'best_spam_detector.pkl'

# URL dataset
DATASET_URL = 'https://github.com/MWiechmann/enron_spam_data/raw/refs/heads/master/enron_spam_data.zip'

# Cấu hình model
TRAIN_TEST_SPLIT = 0.3
RANDOM_STATE = 42

# Cấu hình TF-IDF
TFIDF_CONFIG = {
    'stop_words': 'english',
    'ngram_range': (1, 2),
    'max_features': 50000,
    'min_df': 2,
    'max_df': 0.95
}

# Cấu hình các thuật toán
MODEL_CONFIG = {
    'naive_bayes': {
        'class': 'MultinomialNB',
        'params': {}
    },
    'logistic_regression': {
        'class': 'LogisticRegression', 
        'params': {
            'max_iter': 1000,
            'class_weight': 'balanced'
        }
    }
}

# Email test mẫu
SAMPLE_EMAILS = [
    "Meeting scheduled for tomorrow at 10 AM.",  # ham
    "URGENT: Your account has been compromised. Click the link below to secure it immediately.",  # spam
    "Here are the quarterly sales reports.",  # ham
    "Limited-time offer! Buy now and get 50% off.",  # spam
    "Congratulations! You've won a free lottery, click here to claim.",  # spam
    "Can you send me the project files?",  # ham
]