preprocessing.py

import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.impute import SimpleImputer
from sklearn.model_selection import train_test_split

class DataPreprocessor:
    def __init__(self, data_path='data/credit.csv', random_state=42):
        self.data_path = data_path
        self.random_state = random_state
        self.data = None
        self.scaler = None
        self.X_train = None
        self.X_test = None
        self.y_train = None
        self.y_test = None

    def load_data(self):
        print("Loading data...")
        self.data = pd.read_csv(self.data_path)
        print(f"✓ Loaded {len(self.data)} records with {len(self.data.columns)} features")

    def handle_missing_values(self):
        print("Handling missing values...")
        imputer = SimpleImputer(strategy='median')
        numeric_cols = self.data.select_dtypes(include=[np.number]).columns
        self.data[numeric_cols] = imputer.fit_transform(self.data[numeric_cols])
        print(f"✓ Missing values handled")

    def encode_categorical(self):
        print("Encoding categorical variables...")
        categorical_cols = self.data.select_dtypes(include=['object']).columns
        for col in categorical_cols:
            if col != 'default':
                encoder = LabelEncoder()
                self.data[col] = encoder.fit_transform(self.data[col])
        print(f"✓ Encoded {len(categorical_cols)} columns")

    def engineer_features(self):
        print("Creating new features...")
        if 'debt' in self.data.columns and 'income' in self.data.columns:
            self.data['debt_to_income_ratio'] = self.data['debt'] / self.data['income']
            print(f"✓ Created DTI ratio (mean: {self.data['debt_to_income_ratio'].mean():.4f})")
        else:
            print("⚠ Debt/Income columns not found")

    def split_data(self):
        print("Splitting data into train/test...")
        y = self.data['default'].astype(int)
        X = self.data.drop('default', axis=1)
        
        self.X_train, self.X_test, self.y_train, self.y_test = train_test_split(
            X, y,
            test_size=0.33,
            random_state=self.random_state,
            stratify=y
        )
        
        print(f"✓ Train: {len(self.X_train)} | Test: {len(self.X_test)}")
        print(f"✓ Train classes: {np.bincount(self.y_train.astype(int))}")
        print(f"✓ Test classes: {np.bincount(self.y_test.astype(int))}")

    def scale_features(self):
        print("Scaling features...")
        self.scaler = StandardScaler()
        self.X_train = self.scaler.fit_transform(self.X_train)
        self.X_test = self.scaler.transform(self.X_test)
        print(f"✓ Features scaled")

    def run(self):
        """Execute the full preprocessing pipeline"""
        try:
            self.load_data()
            self.handle_missing_values()
            self.encode_categorical()
            self.engineer_features()
            self.split_data()
            self.scale_features()
            print("\nPreprocessing complete!\n")
            return self.X_train, self.X_test, self.y_train, self.y_test
        except Exception as e:
            print(f"Error: {e}")
            raise