update tests/test_features/test_data_preprocess.py

Author: hakankocakk

tests/test_features/test_data_preprocess.py

@@ -1,6 +1,7 @@
 import pytest
 import pandas as pd
 import os
+import numpy as np
 import src.features.data_preprocess as dp
 
 
@@ -83,3 +84,142 @@ def test_ordinalencoding_test(mock_model_path):
     assert transformed_df_train["EmploymentStatus"].equals(
         transformed_df_test["EmploymentStatus"]
     )
+
+
+def test_onehotencoding_train(mock_model_path):
+
+    data = {
+        "EducationLevel": ["Master", "Associate", "Bachelor", "High School"],
+        "MaritalStatus": ["Married", "Single", "Married", "Single"],
+        "HomeOwnershipStatus": ["Rent", "Own", "Mortgage", "Rent"],
+        "LoanPurpose": ["Auto", "Debt Consolidation", "Home", "Other"],
+        "NumberOfDependents": [0, 2, 1, 3]
+    }
+    sample_dataframe = pd.DataFrame(data)
+
+    transformed_df = dp.onehotencoding(
+        sample_dataframe, mock_model_path, train=True
+    )
+
+    assert "EducationLevel" not in transformed_df.columns
+    assert "MaritalStatus" not in transformed_df.columns
+    assert "HomeOwnershipStatus" not in transformed_df.columns
+    assert "LoanPurpose" not in transformed_df.columns
+    assert "NumberOfDependents" not in transformed_df.columns
+
+    assert any(col.startswith("EducationLevel_")
+               for col in transformed_df.columns)
+    assert any(col.startswith("MaritalStatus_")
+               for col in transformed_df.columns)
+    assert any(col.startswith("HomeOwnershipStatus_")
+               for col in transformed_df.columns)
+    assert any(col.startswith("LoanPurpose_")
+               for col in transformed_df.columns)
+    assert any(col.startswith("NumberOfDependents_")
+               for col in transformed_df.columns)
+
+    model_file = os.path.join(mock_model_path, "one_hot_encoder.pkl")
+    assert os.path.exists(model_file)
+
+
+def test_onehotencoding_test(mock_model_path):
+
+    data = {
+        "EducationLevel": ["Master", "Associate", "Bachelor", "High School"],
+        "MaritalStatus": ["Married", "Single", "Married", "Single"],
+        "HomeOwnershipStatus": ["Rent", "Own", "Mortgage", "Rent"],
+        "LoanPurpose": ["Auto", "Debt Consolidation", "Home", "Other"],
+        "NumberOfDependents": [0, 2, 1, 3]
+    }
+    sample_dataframe = pd.DataFrame(data)
+
+    _ = dp.onehotencoding(
+        sample_dataframe, str(mock_model_path), train=True
+    )
+
+    transformed_df_test = dp.onehotencoding(
+        sample_dataframe, str(mock_model_path), train=False
+    )
+
+    assert "EducationLevel" not in transformed_df_test.columns
+    assert "MaritalStatus" not in transformed_df_test.columns
+    assert "HomeOwnershipStatus" not in transformed_df_test.columns
+    assert "LoanPurpose" not in transformed_df_test.columns
+    assert "NumberOfDependents" not in transformed_df_test.columns
+
+    assert any(col.startswith("EducationLevel_")
+               for col in transformed_df_test.columns)
+    assert any(col.startswith("MaritalStatus_")
+               for col in transformed_df_test.columns)
+    assert any(col.startswith("HomeOwnershipStatus_")
+               for col in transformed_df_test.columns)
+    assert any(col.startswith("LoanPurpose_")
+               for col in transformed_df_test.columns)
+    assert any(col.startswith("NumberOfDependents_")
+               for col in transformed_df_test.columns)
+
+
+def test_normalization_train(mock_model_path):
+
+    num_cols = ['Age', 'AnnualIncome', 'CreditScore',
+                'Experience', 'LoanAmount', 'LoanDuration',
+                'MonthlyDebtPayments', 'CreditCardUtilizationRate',
+                'NumberOfOpenCreditLines', 'NumberOfCreditInquiries',
+                'DebtToIncomeRatio', 'PaymentHistory', 'LengthOfCreditHistory',
+                'SavingsAccountBalance', 'CheckingAccountBalance',
+                'TotalAssets', 'TotalLiabilities', 'MonthlyIncome',
+                'UtilityBillsPaymentHistory', 'JobTenure', 'NetWorth',
+                'BaseInterestRate', 'InterestRate', 'MonthlyLoanPayment',
+                'TotalDebtToIncomeRatio', 'AnIncomeToAssetsRatio',
+                'AnExperienceToAnIncomeRatio', 'LoantoAnIncomeRatio',
+                'DependetToAnIncomeRatio', 'LoansToAssetsRatio',
+                'LoanPaymentToIncomeRatio', 'AnIncomeToDepts', 'AssetsToLoan']
+
+    data = np.random.rand(10, len(num_cols)) * 100
+    df_train = pd.DataFrame(data, columns=num_cols)
+
+    df_normalized = dp.normalization(
+        df_train, str(mock_model_path), train=True
+    )
+
+    model_file = os.path.join(mock_model_path, 'standardscaler.pkl')
+    assert os.path.exists(model_file)
+
+    assert np.all(
+        np.isclose(df_normalized.mean(), 0, atol=1e-1)
+    )
+    assert np.all(
+        np.isclose(df_normalized.std(), 1, atol=1e-1)
+    )
+
+
+def test_normalization_test(mock_model_path):
+
+    num_cols = ['Age', 'AnnualIncome', 'CreditScore',
+                'Experience', 'LoanAmount', 'LoanDuration',
+                'MonthlyDebtPayments', 'CreditCardUtilizationRate',
+                'NumberOfOpenCreditLines', 'NumberOfCreditInquiries',
+                'DebtToIncomeRatio', 'PaymentHistory', 'LengthOfCreditHistory',
+                'SavingsAccountBalance', 'CheckingAccountBalance',
+                'TotalAssets', 'TotalLiabilities', 'MonthlyIncome',
+                'UtilityBillsPaymentHistory', 'JobTenure', 'NetWorth',
+                'BaseInterestRate', 'InterestRate', 'MonthlyLoanPayment',
+                'TotalDebtToIncomeRatio', 'AnIncomeToAssetsRatio',
+                'AnExperienceToAnIncomeRatio', 'LoantoAnIncomeRatio',
+                'DependetToAnIncomeRatio', 'LoansToAssetsRatio',
+                'LoanPaymentToIncomeRatio', 'AnIncomeToDepts', 'AssetsToLoan']
+
+    data = np.random.rand(10, len(num_cols)) * 100
+    df_train = pd.DataFrame(data, columns=num_cols)
+    df_test = pd.DataFrame(data, columns=num_cols)
+
+    df_train_normalized = dp.normalization(
+        df_train, str(mock_model_path), train=True
+    )
+    df_test_normalized = dp.normalization(
+        df_test, (mock_model_path), train=False
+    )
+
+    np.testing.assert_array_almost_equal(
+        df_train_normalized.values, df_test_normalized, decimal=5
+    )