def test_returns_tables_that_span_multiple_pages(self, page_overflow) -> None:
result = extract_table(page_overflow[0:3])
assert result.shape == (100, 5)
pd.testing.assert_index_equal(
result.columns,
pd.Index(['IdField', 'NameField', 'TestField1', 'TestField2', 'TestField3'])
)
def test_returns_correct_table_data_when_working_with_borders(self, borders) -> None:
actual = extract_table(borders)
expected = pd.DataFrame(
columns=['IdField', 'NameField', 'TestField1', 'TestField2', 'TestField3'],
data=[["1", 'Name1', 'Value1', 'Value2', 'Value3'],
["2", 'Name2', 'Value4', 'Value5', 'Value6'],
["3", 'Name3', 'Value7', 'Value8', 'Value9'],
["4", 'Name4', 'Value10', 'Value11', 'Value12'],
["5", 'Name5', 'Value13', 'Value14', 'Value15']]
)
pd.testing.assert_frame_equal(actual, expected)
def test_returns_correct_table_when_missing_most_values(self, missing_values) -> None:
actual = extract_table(missing_values[4])
expected = pd.DataFrame(
columns=['IdField', 'NameField', 'TestField1', 'TestField2', 'TestField3'],
data=[["1", 'Name1', 'Value1', '', ''],
["2", 'Name2', '', '', 'Value2'],
["3", 'Name3', '', '', ''],
["4", 'Name4', '', 'Value3', ''],
["5", 'Name5', '', '', '']]
)
pd.testing.assert_frame_equal(actual, expected)
def test_returns_dataframe(self, table_borders) -> None:
result = extract_table(table_borders[0])
assert isinstance(result, pd.DataFrame)