import unittest
import importlib
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
import seaborn as sns

def _dynamic_test(test_case, condition, success_msg, failure_msg):
    if condition:
        test_case._testMethodName = success_msg
        test_case.assertTrue(True, success_msg)
    else:
        test_case._testMethodName = failure_msg
        test_case.fail(failure_msg)

class TestTitanicTransformations(unittest.TestCase):
    @classmethod
    def setUpClass(cls):
        cls.user_code = importlib.import_module("user_code")
        cls.df = cls.user_code.transformed_data
        cls.original = sns.load_dataset("titanic")

    def test_missing_values_filled(self):
        """Check missing values in 'age' and 'embarked' were handled."""
        df = self.df
        cond = ("age" in df.columns) and not df["age"].isnull().any()
        _dynamic_test(
            self,
            cond,
            "Missing values in 'age' and 'embarked' successfully filled.",
            "Missing values still exist in 'age' or 'embarked'."
        )

    def test_get_dummies_applied(self):
        """Check that get_dummies() was used for 'sex' and 'embarked'."""
        df = self.df
        cond = any(col.startswith("sex_") for col in df.columns) and any(
            col.startswith("embarked_") for col in df.columns
        )
        _dynamic_test(
            self,
            cond,
            "Categorical columns 'sex' and 'embarked' successfully encoded using get_dummies().",
            "'sex' and 'embarked' were not properly one-hot encoded."
        )

    def test_standard_scaler(self):
        """Check scaling of 'age' and 'fare' with StandardScaler."""
        df = self.df
        cond = (
            abs(df["age_scaled"].mean()) < 1e-6 and
            abs(df["fare_scaled"].mean()) < 1e-6 and
            np.isclose(df["age_scaled"].std(), 1, atol=0.1) and
            np.isclose(df["fare_scaled"].std(), 1, atol=0.1)
        )
        _dynamic_test(
            self,
            cond,
            "'age' and 'fare' successfully standardized (mean≈0, std≈1).",
            "'age_scaled' or 'fare_scaled' not properly standardized."
        )

    def test_family_size(self):
        """Check that 'family_size' column exists and was correctly calculated."""
        df = self.df
        cond = "family_size" in df.columns and (
            df["family_size"] == df["sibsp"] + df["parch"] + 1
        ).all()
        _dynamic_test(
            self,
            cond,
            "'family_size' successfully created and computed.",
            "'family_size' missing or incorrectly computed."
        )

if __name__ == "__main__":
    unittest.main(argv=["first-arg-is-ignored"], exit=False)

test_code.py

Apprenez des techniques pratiques pour nettoyer, transformer et concevoir des données pour l'apprentissage automatique avec Python. Ce cours couvre les étapes essentielles de prétraitement, la création de variables et des exercices pratiques pour préparer les données à la modélisation.

Maîtriser les étapes fondamentales du nettoyage et de la préparation des données brutes pour l'analyse et l'apprentissage automatique.

Transformation des données brutes en caractéristiques exploitables pour les modèles d'apprentissage automatique.

Conception et sélection de caractéristiques pour améliorer la performance et l'interprétabilité du modèle.

Défi : Grille de Transformation

Solution