def test_bidding_zones_returned_are_accurate_when_only_half_a_year_before_end_of_DST_is_considered(self): start = dt.date(2015, 1, 1) end = dt.date(2015, 8, 1) german_local_data = Germany() assert german_local_data.get_DST_datetime(start, end) == ( [dt.datetime(2015, 3, 29, 3, 0)], [] )
def test_get_zones_for_Germany_for_dates_after_split(self):
start = pd.Timestamp('20181001', tz='Europe/Berlin')
end = pd.Timestamp('20190930', tz='Europe/Berlin')
german_local_data = Germany()
bidding_zones = german_local_data.get_bidding_zones(start, end)
assert bidding_zones == [
('DE-LU', pd.Timestamp('20181001', tz='Europe/Berlin'), pd.Timestamp('20190930', tz='Europe/Berlin'))
]
def test_bidding_zones_returned_are_accurate_when_only_half_a_year_after_start_of_DST_is_considered(self): start = dt.date(2015, 4, 1) end = dt.date(2015, 12, 1) german_local_data = Germany() assert german_local_data.get_DST_datetime(start, end) == ( [], [dt.datetime(2015, 10, 25, 2, 0)] )
def test_bidding_zones_returned_are_empty_when_non_DST_datetime_ranges_are_considered(self): start = dt.date(2015, 1, 1) end = dt.date(2015, 2, 1) german_local_data = Germany() assert german_local_data.get_DST_datetime(start, end) == ( [], [] )
def test_bidding_zones_returned_are_accurate_from_01_2015_to_12_2015(self): start = dt.date(2015, 1, 1) end = dt.date(2015, 12, 1) german_local_data = Germany() assert german_local_data.get_DST_datetime(start, end) == ( [dt.datetime(2015, 3, 29, 3, 0)], [dt.datetime(2015, 10, 25, 2, 0)] )
def test_daily_dataframe_fix_DST_does_not_add_nan_values_outside_of_dataframe_range(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
with pytest.raises(KeyError):
ddf.dataframe.loc[dt.datetime(2020, 3, 29, 2, 0), 'Day Ahead Price']
def test_daily_dataframe_fixes_DST_for_real_dataset_with_local_DST_included_time(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
assert ddf.start == dt.date(2015,1,1)
assert ddf.end == dt.date(2020,2,6)
def test_daily_dataframe_initialises_correctly_from_ml_ready_df(self):
test_df = pd.read_hdf('tests/test_data/ml_ready.hdf', key='da_prices')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_ml_ready_df(test_df, german_info, time_slots)
assert ddf.start == dt.date(2015,1,6)
assert ddf.end == dt.date(2020,4,29)
def test_daily_dataframe_retrieves_only_23_times_for_beginning_of_DST(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
assert len(ddf.retrieve(
pd.Timestamp('20150329 00:00:0000', tz="Europe/Berlin"),
pd.Timestamp('20150329 23:00:0000', tz="Europe/Berlin"),
['Day Ahead Price']
)) == 23
def test_daily_dataframe_fix_DST_adds_nan_values_for_March_where_there_is_a_missing_hour(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
assert str(ddf.dataframe.loc[dt.datetime(2015, 3, 29, 2, 0), 'Day Ahead Price']) == 'nan'
assert str(ddf.dataframe.loc[dt.datetime(2016, 3, 27, 2, 0), 'Day Ahead Price']) == 'nan'
assert str(ddf.dataframe.loc[dt.datetime(2017, 3, 26, 2, 0), 'Day Ahead Price']) == 'nan'
assert str(ddf.dataframe.loc[dt.datetime(2018, 3, 25, 2, 0), 'Day Ahead Price']) == 'nan'
assert str(ddf.dataframe.loc[dt.datetime(2019, 3, 31, 2, 0), 'Day Ahead Price']) == 'nan'
def test_daily_dataframe_retrieves_correct_dates_for_end_of_DST_where_there_are_duplicates(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
assert len(ddf.retrieve(
pd.Timestamp('20151025 00:00:0000', tz="Europe/Berlin"),
pd.Timestamp('20151025 23:00:0000', tz="Europe/Berlin"),
['Day Ahead Price']
)) == 25
assert len(ddf.retrieve(
pd.Timestamp('20151025 02:00:0000+02:00', tz="Europe/Berlin"),
pd.Timestamp('20151025 02:00:0000+01:00', tz="Europe/Berlin"),
['Day Ahead Price']
)) == 2
def test_daily_dataframe_fix_DST_deletes_duplicate_values_for_October(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
assert len(test_df.loc[dt.datetime(2015, 10, 25, 2, 0)]) == 2
assert len(test_df.loc[dt.datetime(2016, 10, 30, 2, 0)]) == 2
assert len(test_df.loc[dt.datetime(2017, 10, 29, 2, 0)]) == 2
assert len(test_df.loc[dt.datetime(2018, 10, 28, 2, 0)]) == 2
assert len(test_df.loc[dt.datetime(2019, 10, 27, 2, 0)]) == 2
assert ddf.dataframe.loc[dt.datetime(2015, 10, 25, 2, 0), 'Day Ahead Price'] == 25.07
assert ddf.dataframe.loc[dt.datetime(2016, 10, 30, 2, 0), 'Day Ahead Price'] == 31.55
assert ddf.dataframe.loc[dt.datetime(2017, 10, 29, 2, 0), 'Day Ahead Price'] == -83.06
assert ddf.dataframe.loc[dt.datetime(2018, 10, 28, 2, 0), 'Day Ahead Price'] == 41.62
assert ddf.dataframe.loc[dt.datetime(2019, 10, 27, 2, 0), 'Day Ahead Price'] == -29.97
def test_daily_dataframe_retrieves_correct_dates(self):
test_df = pd.read_hdf('tests/test_data/manually_processed_dataframes', key='raw_df')
german_info = Germany()
time_slots = [dt.time(i,0) for i in range(24)]
ddf = daily_dataframe.from_tz_aware_df(test_df, german_info, time_slots)
date_index = ddf.datetimes()
assert min(date_index) == pd.Timestamp('20150101 00:00:0000', tz="Europe/Berlin")
assert max(date_index) == pd.Timestamp('20200206 23:00:0000', tz="Europe/Berlin")
assert np.all(
date_index == pd.date_range(
start='20150101 00:00:0000',
end='20200206 23:00:0000',
freq='H',
tz="Europe/Berlin"
)
)
def test_daily_dataframe_len_works_as_expected(self):
test_df = pd.DataFrame(
data=[
{'v': 1},
{'v': 1},
{'v': 1},
{'v': 1},
],
index=[
dt.datetime(2020,1,1,0,0),
dt.datetime(2020,1,1,2,0),
dt.datetime(2020,1,2,0,0),
dt.datetime(2020,1,2,2,0),
]
)
german_info = Germany()
ddf = daily_dataframe.from_ml_ready_df(test_df, german_info, [dt.time(0,0), dt.time(2,0)])
assert len(ddf) == 4
def test_daily_dataframe_instantiates_for_correct_dataframe(self):
test_df = pd.DataFrame(
data=[
{'v': 1},
{'v': 1},
{'v': 1},
{'v': 1},
{'v': 1},
{'v': 1},
],
index=[
dt.datetime(2020,1,1,0,0),
dt.datetime(2020,1,1,1,0),
dt.datetime(2020,1,2,0,0),
dt.datetime(2020,1,2,1,0),
dt.datetime(2020,1,3,0,0),
dt.datetime(2020,1,3,1,0),
]
)
german_info = Germany()
ddf = daily_dataframe(test_df, german_info, [dt.time(0,0), dt.time(1,0)])
assert ddf.start == dt.date(2020,1,1)
assert ddf.end == dt.date(2020,1,3)
def test_date_range_for_Germany_is_timezone_aware(self): german_local_data = Germany() start, end = german_local_data.get_range() assert start.tzname() == 'CET' assert end.tzname() == 'CEST'
def test_local_info_for_Germany_is_correct(self): german_local_data = Germany() assert german_local_data.timezone_name == 'Europe/Berlin' assert german_local_data.hour_offset_to_UTC == 1