preprocessing.py

import pandas as pd
import numpy as np
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import StandardScaler, OneHotEncoder


class AdvancedPreprocessor:
    """Pipeline de pré-processamento avançado.

    Esta classe lida com:
    - Imputação de dados faltantes
    - Escalonamento de features numéricas
    - Codificação de variáveis categóricas
    """

    def __init__(self, numeric_features=None, categorical_features=None):
        self.numeric_features = numeric_features
        self.categorical_features = categorical_features
        self.pipeline = None

    def build_pipeline(self):
        numeric_transformer = Pipeline(steps=[
            ('imputer', SimpleImputer(strategy='median')),
            ('scaler', StandardScaler())
        ])

        categorical_transformer = Pipeline(steps=[
            ('imputer', SimpleImputer(strategy='most_frequent')),
            ('onehot', OneHotEncoder(handle_unknown='ignore'))
        ])

        self.pipeline = ColumnTransformer(
            transformers=[
                ('num', numeric_transformer, self.numeric_features),
                ('cat', categorical_transformer, self.categorical_features)
            ]
        )
        return self.pipeline

    def fit_transform(self, X):
        if self.pipeline is None:
            self.build_pipeline()
        return self.pipeline.fit_transform(X)

    def transform(self, X):
        if self.pipeline is None:
            raise ValueError("Pipeline não construído. Execute fit_transform primeiro.")
        return self.pipeline.transform(X)