def helper_test_drop(test_case, test_series: TimeSeries):
seriesA = test_series.drop_after(pd.Timestamp('20130105'))
test_case.assertEqual(seriesA.end_time(),
pd.Timestamp('20130105') - test_series.freq())
test_case.assertTrue(
np.all(seriesA.time_index() < pd.Timestamp('20130105')))
seriesB = test_series.drop_before(pd.Timestamp('20130105'))
test_case.assertEqual(seriesB.start_time(),
pd.Timestamp('20130105') + test_series.freq())
test_case.assertTrue(
np.all(seriesB.time_index() > pd.Timestamp('20130105')))
test_case.assertEqual(test_series.freq_str(), seriesA.freq_str())
test_case.assertEqual(test_series.freq_str(), seriesB.freq_str())
def helper_test_append(test_case, test_series: TimeSeries):
# reconstruct series
seriesA, seriesB = test_series.split_after(pd.Timestamp('20130106'))
test_case.assertEqual(seriesA.append(seriesB), test_series)
test_case.assertEqual(
seriesA.append(seriesB).freq(), test_series.freq())
# Creating a gap is not allowed
seriesC = test_series.drop_before(pd.Timestamp('20130107'))
with test_case.assertRaises(ValueError):
seriesA.append(seriesC)
# Changing frequence is not allowed
seriesM = TimeSeries.from_times_and_values(
pd.date_range('20130107', '20130507', freq='30D'), range(5))
with test_case.assertRaises(ValueError):
seriesA.append(seriesM)
def helper_test_append_values(test_case, test_series: TimeSeries):
# reconstruct series
seriesA, seriesB = test_series.split_after(pd.Timestamp('20130106'))
test_case.assertEqual(
seriesA.append_values(seriesB.values(), seriesB.time_index()),
test_series)
test_case.assertEqual(seriesA.append_values(seriesB.values()),
test_series)
# test for equality
test_case.assertEqual(
test_series.drop_after(pd.Timestamp('20130105')).append_values(
test_series.drop_before(pd.Timestamp('20130104')).values()),
test_series)
test_case.assertEqual(seriesA.append_values([]), seriesA)
# randomize order
rd_order = np.random.permutation(range(len(seriesB.values())))
test_case.assertEqual(
seriesA.append_values(seriesB.values()[rd_order],
seriesB.time_index()[rd_order]), test_series)
# add non consecutive index
with test_case.assertRaises(ValueError):
test_case.assertEqual(
seriesA.append_values(seriesB.values(),
seriesB.time_index() + seriesB.freq()),
test_series)
# add existing indices
with test_case.assertRaises(ValueError):
test_case.assertEqual(
seriesA.append_values(
seriesB.values(),
seriesB.time_index() - 3 * seriesB.freq()), test_series)
# other frequency
with test_case.assertRaises(ValueError):
test_case.assertEqual(
seriesA.append_values(
seriesB.values(),
pd.date_range('20130107', '20130113', freq='2d')),
test_series)
def _crop_to_match_seasons(series: TimeSeries,
required_matches: Optional[set]) -> TimeSeries:
"""Crops TimeSeries instance to contain full periods.
Crops a given TimeSeries `series` that will be used as a training set in such
a way that its first entry has a timestamp that matches the first timestamp
right after the end of `series` in all attributes given in `required_matches`.
If no such timestamp can be found, the original TimeSeries instance is returned.
If the value of `required_matches` is `None`, the original TimeSeries instance is returned.
Parameters
----------
series
TimeSeries instance to be cropped.
required_matches
A set of pd.Timestamp attributes which will be used to choose the cropping point.
Returns
-------
TimeSeries
New TimeSeries instance that is cropped as described above.
"""
if required_matches is None or len(required_matches) == 0:
return series
first_ts = series.time_index[0]
freq = series.freq
pred_ts = series.time_index[-1] + freq
# start at first timestamp of given series and move forward until a matching timestamp is found
curr_ts = first_ts
while curr_ts < pred_ts - 4 * freq:
curr_ts += freq
if _compare_timestamps_on_attributes(pred_ts, curr_ts,
required_matches):
new_series = series.drop_before(curr_ts)
return new_series
logger.warning(
"No matching timestamp could be found, returning original TimeSeries.")
return series