create tests/test_features/test_data_preprocess.py

Author: hakankocakk

tests/test_features/test_data_preprocess.py

@@ -0,0 +1,85 @@
+import pytest
+import pandas as pd
+import os
+import src.features.data_preprocess as dp
+
+
+def test_load_data_valid_file(tmp_path):
+    test_file = tmp_path / "test.csv"
+    test_file.write_text("col1,col2\n1,2\n3,4")
+
+    df = dp.load_data(str(test_file))
+
+    assert isinstance(df, pd.DataFrame)
+    assert df.shape == (2, 2)
+
+
+def test_feature_engineering_valid():
+
+    data = {
+        "AnnualIncome": [50000, 60000],
+        "TotalAssets": [200000, 300000],
+        "Experience": [5, 10],
+        "LoanAmount": [20000, 25000],
+        "NumberOfDependents": [2, 3],
+        "TotalLiabilities": [50000, 60000],
+        "MonthlyLoanPayment": [1000, 1200],
+        "MonthlyIncome": [4000, 5000],
+        "MonthlyDebtPayments": [500, 700]
+    }
+    df = pd.DataFrame(data)
+    transformed_df = dp.feature_engineering(df)
+    expected_columns = [
+        "AnIncomeToAssetsRatio", "AnExperienceToAnIncomeRatio",
+        "LoantoAnIncomeRatio", "DependetToAnIncomeRatio",
+        "LoansToAssetsRatio", "LoanPaymentToIncomeRatio",
+        "AnIncomeToDepts", "AssetsToLoan"
+    ]
+
+    for col in expected_columns:
+        assert col in transformed_df.columns, f"{col} is missing!"
+
+    assert isinstance(transformed_df, pd.DataFrame)
+
+
+@pytest.fixture
+def mock_model_path(tmp_path):
+    return tmp_path
+
+
+def test_ordinalencoding_train(mock_model_path):
+
+    data = {
+        "EmploymentStatus": ["Employed", "Self-Employed",
+                             "Unemployed", "Employed"]
+    }
+    df1 = pd.DataFrame(data)
+
+    transformed_df = dp.ordinalencoding(df1, str(mock_model_path), train=True)
+
+    assert transformed_df["EmploymentStatus"].iloc[0] == 0
+    assert transformed_df["EmploymentStatus"].iloc[1] == 1
+    assert transformed_df["EmploymentStatus"].iloc[2] == 2
+
+    assert os.path.exists(os.path.join(mock_model_path, 'ordinal_encoder.pkl'))
+
+
+def test_ordinalencoding_test(mock_model_path):
+
+    data = {
+        "EmploymentStatus": ["Employed", "Self-Employed",
+                             "Unemployed", "Employed"]
+    }
+    df_train = pd.DataFrame(data)
+    df_test = pd.DataFrame(data)
+
+    transformed_df_train = dp.ordinalencoding(
+        df_train, str(mock_model_path), train=True
+    )
+    transformed_df_test = dp.ordinalencoding(
+        df_test, str(mock_model_path), train=False
+    )
+
+    assert transformed_df_train["EmploymentStatus"].equals(
+        transformed_df_test["EmploymentStatus"]
+    )