def test_has_dst_no_dst_in_date_range(albuquerque):
sunrise = _get_sunrise(albuquerque, 'America/Denver')
july = sunrise['2020-07-01':'2020-07-31']
february = sunrise['2020-02-01':'2020-03-05']
expected_july = pd.Series(False, index=july.index)
expected_march = pd.Series(False, index=february.index)
assert_series_equal(expected_july, time.has_dst(july, 'America/Denver'))
assert_series_equal(expected_march, time.has_dst(february, 'MST'))
def test_has_dst(tz, observes_dst, albuquerque):
sunrise = _get_sunrise(albuquerque, tz)
dst = time.has_dst(sunrise, 'America/Denver')
expected = pd.Series(False, index=sunrise.index)
expected.loc['2020-03-08'] = observes_dst
expected.loc['2020-11-01'] = observes_dst
assert_series_equal(expected, dst, check_names=False)
def test_has_dst_input_series_not_localized(tz, observes_dst, albuquerque):
sunrise = _get_sunrise(albuquerque, tz)
sunrise = sunrise.tz_localize(None)
expected = pd.Series(False, index=sunrise.index)
expected.loc['2020-03-08'] = observes_dst
expected.loc['2020-11-01'] = observes_dst
dst = time.has_dst(sunrise, 'America/Denver')
assert_series_equal(expected, dst)
def test_has_dst_gaps(albuquerque):
sunrise = _get_sunrise(albuquerque, 'America/Denver')
sunrise.loc['3/5/2020':'3/10/2020'] = pd.NaT
sunrise.loc['7/1/2020':'7/20/2020'] = pd.NaT
sunrise.dropna(inplace=True)
expected = pd.Series(False, index=sunrise.index)
expected['11/1/2020'] = True
assert_series_equal(time.has_dst(sunrise, 'America/Denver'), expected)
def test_has_dst_rounded(tz, freq, observes_dst, albuquerque):
sunrise = _get_sunrise(albuquerque, tz)
# With rounding to 1-hour timestamps we need to reduce how many
# days we look at.
window = 7 if freq != 'H' else 1
expected = pd.Series(False, index=sunrise.index)
expected.loc['2020-03-08'] = observes_dst
expected.loc['2020-11-01'] = observes_dst
dst = time.has_dst(sunrise.dt.round(freq), 'America/Denver', window=window)
assert_series_equal(expected, dst, check_names=False)
def test_has_dst_missing_data(albuquerque):
sunrise = _get_sunrise(albuquerque, 'America/Denver')
sunrise.loc['3/5/2020':'3/10/2020'] = pd.NaT
sunrise.loc['7/1/2020':'7/20/2020'] = pd.NaT
# Doesn't raise since both sides still have some data
expected = pd.Series(False, index=sunrise.index)
expected['3/8/2020'] = True
expected['11/1/2020'] = True
assert_series_equal(time.has_dst(sunrise, 'America/Denver'), expected)
missing_all_before = sunrise.copy()
missing_all_after = sunrise.copy()
missing_all_before.loc['3/1/2020':'3/5/2020'] = pd.NaT
missing_all_after.loc['3/8/2020':'3/14/2020'] = pd.NaT
missing_data_message = r'No data at .*\. ' \
r'Consider passing a larger `window`.'
# Raises for missing data before transition date
with pytest.raises(ValueError, match=missing_data_message):
time.has_dst(missing_all_before, 'America/Denver')
# Raises for missing data after transition date
with pytest.raises(ValueError, match=missing_data_message):
time.has_dst(missing_all_after, 'America/Denver')
# Raises for missing data before and after the shift date
sunrise.loc['3/1/2020':'3/7/2020'] = pd.NaT
sunrise.loc['3/9/2020':'3/14/2020'] = pd.NaT
with pytest.raises(ValueError, match=missing_data_message):
time.has_dst(sunrise, 'America/Denver')
with pytest.warns(UserWarning, match=missing_data_message):
result = time.has_dst(sunrise, 'America/Denver', missing='warn')
expected.loc['3/8/2020'] = False
assert_series_equal(expected, result)
sunrise.loc['3/1/2020':'3/14/2020'] = pd.NaT
with pytest.warns(UserWarning, match=missing_data_message):
result = time.has_dst(sunrise, 'America/Denver', missing='warn')
assert_series_equal(expected, result)
with pytest.raises(ValueError, match=missing_data_message):
time.has_dst(sunrise, 'America/Denver')
