def test_get_date_ranges_to_download():
date_ranges = utils.get_date_ranges_to_download(TEST_HDF, PV_SYSTEM,
'2018-01-01', '2019-01-10')
# 2018-01-02 and 2018-01-03 are already known to be missing.
np.testing.assert_array_equal(date_ranges, [
DateRange(start_date=date(2018, 1, 1), end_date=date(2018, 12, 31)),
DateRange(start_date=date(2019, 1, 4), end_date=date(2019, 1, 10))
])
def _filter_date_range(
self,
store_filename: str,
system_id: int,
date_ranges: Iterable[DateRange],
min_data_availability: Optional[float] = 0.5) -> List[DateRange]:
"""Check getstatistic to see if system_id has data for all date ranges.
Args:
system_id: PV system ID.
store_filename: HDF5 filename to cache statistics to / from.
date_ranges: List of DateRange objects.
min_data_availability: A float in the range [0, 1]. 1 means only
accept PV systems which have no days of missing data. 0 means
accept all PV systems, no matter if they have missing data.
"""
if not date_ranges:
return date_ranges
stats = self._get_statistic_with_cache(
store_filename,
system_id,
date_to=date_ranges[-1].end_date,
wait_if_rate_limit_exceeded=True).squeeze()
if (pd.isnull(stats['actual_date_from'])
or pd.isnull(stats['actual_date_to'])):
_LOG.info('system_id %d: Stats say there is no data!', system_id)
return []
timeseries_date_range = DateRange(stats['actual_date_from'],
stats['actual_date_to'])
data_availability = (stats['num_outputs'] /
(timeseries_date_range.total_days() + 1))
if data_availability < min_data_availability:
_LOG.info(
'system_id %d: Data availability too low! Only %.0f %%.',
system_id, data_availability * 100)
return []
new_date_ranges = []
for date_range in date_ranges:
new_date_range = date_range.intersection(timeseries_date_range)
if new_date_range:
new_date_ranges.append(new_date_range)
return new_date_ranges
def test_merge_date_ranges_to_years():
jan = DateRange("2018-01-01", "2018-02-01")
multiyear = DateRange("2017-01-01", "2018-02-01")
old_multiyear = DateRange("2014-01-01", "2016-02-01")
ancient_jan = DateRange("2010-01-01", "2010-02-01")
for date_ranges, merged in [([], []),
([jan],
[DateRange("2017-02-01", "2018-02-01")]),
([multiyear], [
DateRange("2017-02-01", "2018-02-01"),
DateRange("2016-02-02", "2017-02-01")
]),
([old_multiyear, multiyear], [
DateRange("2017-02-01", "2018-02-01"),
DateRange("2016-02-02", "2017-02-01"),
DateRange("2015-02-01", "2016-02-01"),
DateRange("2014-02-01", "2015-02-01"),
DateRange("2013-02-01", "2014-02-01")
]),
([ancient_jan, old_multiyear, multiyear], [
DateRange("2017-02-01", "2018-02-01"),
DateRange("2016-02-02", "2017-02-01"),
DateRange("2015-02-01", "2016-02-01"),
DateRange("2014-02-01", "2015-02-01"),
DateRange("2013-02-01", "2014-02-01"),
DateRange("2009-02-01", "2010-02-01")
])]:
assert merge_date_ranges_to_years(date_ranges) == merged
def test_split_into_years():
short_dr = DateRange("2019-01-01", "2019-01-10")
assert short_dr.split_into_years() == [short_dr]
one_year = DateRange("2019-01-01", "2020-01-01")
assert one_year.split_into_years() == [one_year]
year_and_half = DateRange("2019-01-01", "2020-06-01")
assert year_and_half.split_into_years() == [
DateRange("2019-06-02", "2020-06-01"),
DateRange("2019-01-01", "2019-06-02")
]
def test_total_days():
assert DateRange("2019-01-01", "2019-01-10").total_days() == 9
def test_intersection():
assert DateRange("2019-01-01", "2019-01-02").intersection(
DateRange("2020-01-01", "2020-01-02")) is None
assert DateRange("2019-01-01", "2019-01-10").intersection(
DateRange("2019-01-01",
"2019-01-02")) == DateRange("2019-01-01", "2019-01-02")
assert DateRange("2019-01-01", "2019-01-10").intersection(
DateRange("2019-01-05",
"2019-01-20")) == DateRange("2019-01-05", "2019-01-10")
year = DateRange("2018-01-1", "2019-01-01")
dec = DateRange("2018-12-01", "2019-01-01")
assert year.intersection(dec) == dec
june = DateRange("2018-06-01", "2018-07-01")
assert year.intersection(june) == june
incomplete_overlap = DateRange("2017-07-01", "2018-02-01")
assert year.intersection(incomplete_overlap) != incomplete_overlap