klib.tests.test_preprocess

Complexity

Total Complexity

9

Size/Duplication

Total Lines	58
Duplicated Lines	56.9 %

Importance

Changes

0

3 Methods

Rating	Name	Duplication	Size	Complexity
A	Test_train_dev_test_split.setUpClass()	0	14	1
A	Test_train_dev_test_split.test_train_dev_test_split_series()	17	17	4
A	Test_train_dev_test_split.test_train_dev_test_split_col()	16	17	4

import pandas as pd
import unittest
from ..preprocess import train_dev_test_split


class Test_train_dev_test_split(unittest.TestCase):

    @classmethod
    def setUpClass(cls):
        cls.data_split = pd.DataFrame([[1, 2, 3, 4, 'a'],
                                       [2, 4, 5, 6, 'b'],
                                       [3, 4, 2, 1, 'c'],
                                       [4, 0, 3, 4, 'a'],
                                       [5, 4, 5, 6, 'b'],
                                       [6, 4, 2, 1, 'c'],
                                       [7, 0, 3, 4, 'a'],
                                       [8, 4, 5, 6, 'b'],
                                       [9, 4, 2, 1, 'c'],
                                       [10, 2, 1, 5, 'b']],
                                      columns=['Col1', 'Col2', 'Col3', 'Col4', 'Col5'])
        cls.data_target = pd.Series([1, 0, 1, 0, 0, 1, 1, 0, 1, 1])

    def test_train_dev_test_split_col(self):
        # Test the propper splitting in train, dev and test sets

        expected_results = [(8, 4), (1, 4), (1, 4), (8,), (1,), (1,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, 'Col2',
                                                  random_state=1234)[i].shape, expected_results[i])

        expected_results = [(8, 4), (2, 4), (8,), (2,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, target='Col2',
                                                  dev_size=0, test_size=0.2)[i].shape, expected_results[i])

        expected_results = [(5, 4), (5, 4), (5,), (5,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, target='Col2',
                                                  dev_size=0.5, test_size=0)[i].shape, expected_results[i])

    def test_train_dev_test_split_series(self):
        # Test the propper splitting in train, dev and test sets

        expected_results = [(6, 5), (2, 5), (2, 5), (6,), (2,), (2,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, target=self.data_target,
                                                  dev_size=0.2, test_size=0.2)[i].shape, expected_results[i])

        expected_results = [(8, 5), (2, 5), (8,), (2,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, target=self.data_target,
                                                  dev_size=0, test_size=0.2)[i].shape, expected_results[i])

        expected_results = [(5, 5), (5, 5), (5,), (5,)]
        for i, _ in enumerate(expected_results):
            self.assertEqual(train_dev_test_split(self.data_split, target=self.data_target,
                                                  dev_size=0.5, test_size=0)[i].shape, expected_results[i])