stocksymbolscraper/tests/test_scrape.py

from unittest import TestCase
from unittest.mock import patch, MagicMock, call

from bs4 import BeautifulSoup

from stocksymbolscraper import scrape


class ScrapeTestCase(TestCase):

    @patch.object(scrape, 'get_single_tr_data')
    def test_extract_row_data(self, mock_get_single_tr_data: MagicMock):
        foo = 'foo'
        mock_get_single_tr_data.return_value = foo
        input1, input2, input3 = MagicMock(), MagicMock(), MagicMock()
        # Although the function expects BS4 Tag objects as arguments, we substitute with Mocks here
        # because those arguments are immediately passed into another function, which we mock out anyway.
        output = scrape.extract_row_data(input1, input2, input3)
        expected_output = [foo, foo, foo]
        self.assertListEqual(expected_output, output)
        mock_get_single_tr_data.assert_has_calls([call(input1), call(input2), call(input3)])

    @patch.object(scrape, 'get_str_from_td')
    def test_get_single_tr_data(self, mock_get_str_from_td: MagicMock):
        a, b, x = 'a', 'b', 'x'
        mock_get_str_from_td.return_value = x
        test_html = f'<tr> <td><a>{a}<small>({b})</small></a></td>' + \
                    f'<td>foo</td> <td>bar</td> <td>baz</td> </tr>'
        test_tr = BeautifulSoup(test_html, scrape.HTML_PARSER).tr
        test_tds = test_tr.find_all('td')
        output = scrape.get_single_tr_data(test_tr)
        expected_output = (a, b, x, x, x)
        self.assertTupleEqual(expected_output, output)
        mock_get_str_from_td.assert_has_calls([call(test_tds[1]), call(test_tds[2]), call(test_tds[3])])

        test_html = f'<tr> <td><a>{a}<small>***{b}***</small></a></td>' + \
                    f'<td>foo</td> <td>bar</td> <td>baz</td> </tr>'
        test_tr = BeautifulSoup(test_html, scrape.HTML_PARSER).tr
        test_tds = test_tr.find_all('td')
        output = scrape.get_single_tr_data(test_tr)
        expected_output = (a, f'***{b}***', x, x, x)
        self.assertTupleEqual(expected_output, output)
        mock_get_str_from_td.assert_has_calls([call(test_tds[1]), call(test_tds[2]), call(test_tds[3])])

    def test_get_str_from_td(self):
        expected_output = 'foo bar'
        test_td = BeautifulSoup(f'<td>   {expected_output} </td>', scrape.HTML_PARSER).td
        output = scrape.get_str_from_td(test_td)
        self.assertEqual(expected_output, output)

        expected_output = ''
        test_td = BeautifulSoup('<td></td>', scrape.HTML_PARSER).td
        output = scrape.get_str_from_td(test_td)
        self.assertEqual(expected_output, output)

    def test_trs_from_page(self):
        # Tested function takes URL as argument (GET request is issued)
        # HTML to be parsed could be substituted
        #
        pass