test_class_unittest.py

import unittest
import pandas as pd
import numpy as np
import sys
import re
import logging
import statsmodels.api as sm

sys.path.append("refactored_pd/")

import data
from class_modelperf import ModelPerfomance

diagnostics_logger = logging.getLogger("class_logistic_unittest")
diagnostics_logger.setLevel(logging.DEBUG)
console_handler = logging.StreamHandler()
console_handler.setFormatter(
    logging.Formatter(fmt="{levelname}:{name}:{message}", style="{")
)
diagnostics_logger.addHandler(console_handler)
diagnostics_logger.info("LOGISTIC REGRESSION UNIT TESTS")


class TestDataframe(unittest.TestCase):
    def __str__(self):
        """This executed when calling print on this object"""

        pattern = re.compile(r"^_")
        method_names = []
        for name, func in TestDataframe.__dict__.items():
            if not pattern.match(name) and callable(func):
                method_names.append(name)

        return f"This is Class {self.__class__.__name__} with methods {method_names}"

    def test_no_missing_values(self):
        """Here we are testing if our Imputation methods worked form the Imputation class"""

        dataframe = data.imputer_cat
        self.assertFalse(
            dataframe.isnull().values.any(), "Dataframe contains missing values"
        )

    def test_zeros_and_ones_values(self):
        """Here we are testing if Onehot Encoding worked from OneHotEncoding class"""

        dataframe = data.instance_stats
        self.assertTrue(
            (dataframe.onehot_encoding().applymap(lambda x: x in [0, 1])).all().all(),
            "Dataframe contains values other than zero or one",
        )


class TestProbability(unittest.TestCase, object):
    def __str__(self):
        """This executed when calling print on this object"""

        pattern = re.compile(r"^_")
        method_names = []
        for name, func in TestProbability.__dict__.items():
            if not pattern.match(name) and callable(func):
                method_names.append(name)
        return f"This is Class {self.__class__.__name__} with methods {method_names}"

    def test_no_values_less_than_one(self):
        """Here we are testing if our Logistic regression does not return nonsensical values,
        e.g. values less than zero"""
        x_test = sm.add_constant(data.m.x_test_glm.values, has_constant="add")
        values = np.array(data.m.glm_probability_prediction(x_test))
        self.assertFalse(
            (values < 0).any(), "Prediction contains missing values less than 0"
        )

    def test_no_values_greater_than_one(self):
        """Here we are testing if our Logistic regression does not return nonsensical values,
        e.g. values greater than one"""

        x_test = sm.add_constant(data.m.x_test_glm.values, has_constant="add")
        values = np.array(data.m.glm_probability_prediction(x_test))
        self.assertFalse(
            (values > 1).any(), "Prediction contains missing values less than 0"
        )


if __name__ == "__main__":
    unittest.main()