def test_create_relevant_data_row(self):
set_intervals([5,5,5])
set_target_interval(timedelta(minutes=5))
df = pd.read_csv('tests/resources/dataframe_data.csv').iloc[1998:2018, :]
expected_row = create_expected_row([0.79227, 0.79231, 0.79216, 0.79222, 0.79223, 0.79312, 0.79219, 0.79312, 0.79315, 0.79325, 0.79279, 0.79284], [0.79283, 0.79258])
actual_row = create_relevant_data_row(df, datetime.strptime('2014-07-18 09:04:00', '%Y-%m-%d %H:%M:%S'))
self.assertTrue(np.array_equal(expected_row, actual_row))
def test_get_relevant_data(self, mock_method):
set_intervals([15,15,15,15])
set_target_interval(timedelta(minutes=60))
df = pd.read_csv('tests/resources/dataframe_data.csv')
target_date = datetime.strptime('2014-07-17 00:00:00', '%Y-%m-%d %H:%M:%S')
get_relevant_data(df, target_date)
start_date = datetime.strptime('2014-07-16 23:00:00', '%Y-%m-%d %H:%M:%S')
end_date = datetime.strptime('2014-07-17 01:00:00', '%Y-%m-%d %H:%M:%S')
mock_method.assert_called_with(start_date, end_date, df)
def test_process_input_data_error(self):
set_intervals([5, 5, 5, 60])
df = pd.read_csv('tests/resources/dataframe_data.csv').iloc[1998:2013, :]
expected_error_message = 'Insufficient data to process for this number of intervals'
try:
actual_input_data = process_input_data(df)
except:
exc_type, exc_value, exc_traceback = sys.exc_info()
self.assertEqual(expected_error_message, str(exc_value))
def test_process_input_data(self):
set_intervals([5, 5, 5])
df = pd.read_csv('tests/resources/dataframe_data.csv').iloc[1998:2013, :]
test_data = {
'datetime': ['2014-07-18 08:49:00', '2014-07-18 08:54:00', '2014-07-18 08:59:00'],
'open': [0.79227, 0.79223, 0.79315],
'high': [0.79231, 0.79312, 0.79325],
'low': [0.79216, 0.79219, 0.79279],
'close': [0.79222, 0.79312, 0.79284]
}
expected_input_data = pd.DataFrame(data=test_data)
actual_input_data = process_input_data(df)
self.assertTrue(expected_input_data.equals(actual_input_data))
def test_create_row(self):
set_intervals([5,5,5])
test_data = {
'datetime': ['2014-07-18 08:49:00', '2014-07-18 08:54:00', '2014-07-18 08:59:00'],
'open': [0.79227, 0.79223, 0.79315],
'high': [0.79231, 0.79312, 0.79325],
'low': [0.79216, 0.79219, 0.79279],
'close': [0.79222, 0.79312, 0.79284]
}
input_values = pd.DataFrame(data=test_data)
expected_row = create_expected_row([0.79227, 0.79231, 0.79216, 0.79222, 0.79223, 0.79312, 0.79219, 0.79312, 0.79315, 0.79325, 0.79279, 0.79284], [1, 2])
actual_row = create_row(input_values, [1,2])
self.assertTrue(np.array_equal(expected_row, actual_row))
def test_get_dates(self):
intervals = [5, 5, 5]
set_intervals(intervals)
training_start = datetime.strptime('2020-01-01 00:00:00', '%Y-%m-%d %H:%M:%S')
validation_start = datetime.strptime('2020-01-01 01:00:00', '%Y-%m-%d %H:%M:%S')
test_start = datetime.strptime('2020-01-01 02:00:00', '%Y-%m-%d %H:%M:%S')
test_end = datetime.strptime('2020-01-01 03:00:00', '%Y-%m-%d %H:%M:%S')
actual_training_dates, actual_validation_dates, actual_test_dates = get_dates(training_start, validation_start, test_start, test_end)
expected_training_dates = convert_datestring_array_to_datetime(['2020-01-01 00:00:00', '2020-01-01 00:15:00', '2020-01-01 00:30:00', '2020-01-01 00:45:00'])
expected_validation_dates = convert_datestring_array_to_datetime(['2020-01-01 01:00:00', '2020-01-01 01:15:00', '2020-01-01 01:30:00', '2020-01-01 01:45:00'])
expected_test_dates = convert_datestring_array_to_datetime(['2020-01-01 02:00:00', '2020-01-01 02:15:00', '2020-01-01 02:30:00', '2020-01-01 02:45:00'])
self.assertEqual(expected_training_dates, actual_training_dates)
self.assertEqual(expected_validation_dates, actual_validation_dates)
self.assertEqual(expected_test_dates, actual_test_dates)
def process_raw_data_file(name, intervals, target_interval_minutes, market,
input_minutes_missing_allowance, training_start_str,
validation_start_str, test_start_str, test_end_str,
dataframe_batch_size):
set_name(name)
set_intervals(intervals)
set_target_interval(timedelta(minutes=target_interval_minutes))
set_market(market)
set_max_input_minutes_missing(input_minutes_missing_allowance)
set_df_width(dataframe_batch_size)
training_start = datetime.strptime(training_start_str, '%Y-%m-%d %H:%M:%S')
validation_start = datetime.strptime(validation_start_str,
'%Y-%m-%d %H:%M:%S')
test_start = datetime.strptime(test_start_str, '%Y-%m-%d %H:%M:%S')
test_end = datetime.strptime(test_end_str, '%Y-%m-%d %H:%M:%S')
dates = [training_start, validation_start, test_start, test_end]
timings = create_data(
[training_start, validation_start, test_start, test_end])
meta_data = {
'name': name,
'intervals': intervals,
'target_interval_minutes': target_interval_minutes,
'market': market,
'training_start': training_start_str,
'validation_start': validation_start_str,
'test_start': test_start_str,
'test_end': test_end_str,
'dataframe_batch_size': dataframe_batch_size,
'time_to_write_training': timings[0],
'time_to_write_validation': timings[1],
'time_to_write_test': timings[2]
}
write_meta_data_file(f'models/{name}/data', meta_data)
def test_set_intervals(self):
intervals = [5, 5, 5]
set_intervals(intervals)
self.assertEqual(intervals, get_intervals())
