daniil-berg/mwfin
2
1
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

implementation to pass tests

Browse Source
This commit is contained in:
Daniil Fajnberg 2021-12-03 16:28:12 +01:00
parent d46148a510
commit 94568c6165
3 changed files with 15 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
src/mwfin/constants.py
View File

@ -1,3 +1,4 @@
DEV_MODE = False
MAIN_LOGGER_NAME = 'mwfin' MAIN_LOGGER_NAME = 'mwfin'
HTML_PARSER = 'html.parser' HTML_PARSER = 'html.parser'

13
src/mwfin/functions.py
View File

@ -9,6 +9,7 @@ from webutils import in_async_session, gather_in_batches
from .constants import (HTML_PARSER, BASE_URL, END_DATE, BS, IS, CF, FIN_STMT_URL_SUFFIX, FIN_STMT_ITEMS, from .constants import (HTML_PARSER, BASE_URL, END_DATE, BS, IS, CF, FIN_STMT_URL_SUFFIX, FIN_STMT_ITEMS,
DEFAULT_CONCURRENT_BATCH_SIZE) DEFAULT_CONCURRENT_BATCH_SIZE)
from .exceptions import UnknownFinancialStatementItem
log = logging.getLogger(__name__) log = logging.getLogger(__name__)
@ -48,15 +49,21 @@ def is_relevant_table_row(tr: Tag) -> bool:
return FIN_STMT_ITEMS[item_name] return FIN_STMT_ITEMS[item_name]
except KeyError: except KeyError:
log.warning(f"Unknown item name '{item_name}' found in financial statement.") log.warning(f"Unknown item name '{item_name}' found in financial statement.")
return False raise UnknownFinancialStatementItem
def find_relevant_table_rows(soup: BeautifulSoup) -> List[Tag]: def find_relevant_table_rows(soup: BeautifulSoup) -> List[Tag]:
""" """
Returns the table rows containing the data of interest. Returns the table rows containing the data of interest.
""" """
trs = soup.find('div', attrs={'class': 'financials'}).tbody.find_all('tr') trs = []
return [tr for tr in trs if is_relevant_table_row(tr)] for tr in soup.find('div', attrs={'class': 'financials'}).tbody.find_all('tr'):
try:
if is_relevant_table_row(tr):
trs.append(tr)
except UnknownFinancialStatementItem:
pass
return trs
def extract_row_data(tr: Tag) -> tuple[str, tuple[int]]: def extract_row_data(tr: Tag) -> tuple[str, tuple[int]]:

5
tests/test_functions.py
View File

@ -70,14 +70,17 @@ class FunctionsTestCase(IsolatedAsyncioTestCase):
def test_find_relevant_table_rows(self, mock_is_relevant_table_row): def test_find_relevant_table_rows(self, mock_is_relevant_table_row):
mock_is_relevant_table_row.return_value = True mock_is_relevant_table_row.return_value = True
expected_output = self.test_soup.find('div', attrs={'class': 'financials'}).tbody.find_all('tr') expected_output = self.test_soup.find('div', attrs={'class': 'financials'}).tbody.find_all('tr')
tr0, tr1 = expected_output
output = functions.find_relevant_table_rows(self.test_soup) output = functions.find_relevant_table_rows(self.test_soup)
self.assertListEqual(expected_output, output) self.assertListEqual(expected_output, output)
mock_is_relevant_table_row.assert_has_calls([call(expected_output[0]), call(expected_output[1])]) mock_is_relevant_table_row.assert_has_calls([call(tr0), call(tr1)])
mock_is_relevant_table_row.reset_mock()
mock_is_relevant_table_row.side_effect = UnknownFinancialStatementItem() mock_is_relevant_table_row.side_effect = UnknownFinancialStatementItem()
expected_output = self.test_soup.find_all('thistagdoesntexist') expected_output = self.test_soup.find_all('thistagdoesntexist')
output = functions.find_relevant_table_rows(self.test_soup) output = functions.find_relevant_table_rows(self.test_soup)
self.assertListEqual(expected_output, output) self.assertListEqual(expected_output, output)
mock_is_relevant_table_row.assert_has_calls([call(tr0), call(tr1)])
def test_extract_row_data(self): def test_extract_row_data(self):
test_row = self.test_soup.find('div', attrs={'class': 'financials'}).tbody.tr test_row = self.test_soup.find('div', attrs={'class': 'financials'}).tbody.tr