def test_format_civil_time(self):
ct = time.CivilTime(1970, 1, 1)
self.assertEqual("1970-01-01T00:00:00", time.FormatCivilTime(ct))
ct = time.CivilTime(2015, 1, 2, 3, 4, 5)
self.assertEqual("2015-01-02T03:04:05", time.FormatCivilTime(ct))
ct = time.ToCivil(time.Time(1234567890 * 1e9))
self.assertEqual("2009-02-13T23:31:30", time.FormatCivilTime(ct))
est: time.Timezone = time.timezone('US/Eastern')
ct = time.ToCivil(time.Time(1234567890 * 1e9), est)
self.assertEqual("2009-02-13T18:31:30", time.FormatCivilTime(ct))
def _get_filter(self, lookup_key: store.DataStore.LookupKey) -> Dict:
res = {}
if lookup_key.data_space:
res['dataSpace'] = lookup_key.data_space
if lookup_key.symbol:
res['symbol'] = lookup_key.symbol
if lookup_key.data_type:
res['dataType'] = lookup_key.data_type
if lookup_key.timestamp:
res['timestamp'] = time.ToCivil(lookup_key.timestamp)
return res
def get_close_prices_for_ema20(t: time.Time, symbol: str) -> List[DataEntry]:
# INTC 2019-12-03 to 2019-12-31
price_data = [
56.07, 56.02, 56.08, 56.81, 56.53, 56.59, 57.07, 57.55, 57.79, 57.70,
57.23, 57.17, 57.96, 58.95, 59.23, 59.41, 59.82, 60.08, 59.62, 59.85,
] # yapf: disable
close_prices = [
_make_close_price(symbol, value, time.FromCivil(d))
for value, d in zip(price_data,
trading_days.get_last_n(time.ToCivil(t), 20))
]
random.shuffle(close_prices)
return close_prices
def get_data(
self,
symbol: str,
start_date: Optional[time.CivilTime] = None,
end_date: Optional[time.CivilTime] = None,
) -> Dict[int, List[_DataEntry]]:
if not symbol:
raise errors.InvalidArgumentError('Symbol not specified')
if not end_date:
end_date = time.ToCivil(time.Now())
if not start_date:
start_date = end_date
if start_date > end_date:
raise errors.InvalidArgumentError(
f'start_date {start_date} is later than end_date {end_date}')
if end_date > time.ToCivil(time.Now()):
raise errors.InvalidArgumentError(
f'end_date {end_date} must be in the past')
data = self._batch_api.get_batch_quotes(symbol, start_date, end_date)
chart_series = data[symbol]['chart']
now = time.Now()
data_types_to_import = _DATA_TYPE_TO_CHART_FIELD_MAP.keys()
results = {data_type: [] for data_type in data_types_to_import}
for chart_data in chart_series:
data_date = time.ParseCivilTime(chart_data['date'])
for data_type in data_types_to_import:
data = self._make_data_base(symbol, data_date, now)
data.data_type = data_type
data.value = chart_data[
_DATA_TYPE_TO_CHART_FIELD_MAP[data_type]]
results[data_type].append(data)
return results
def recursive_inputs_shape(
base_calc_type: DataType.Enum, incr_calc_type: DataType.Enum,
t: time.Time, time_spec: calc.CalcTimeSpecs) -> calc.RecursiveInputs:
# TODO: same here, think about how to get it work with period other than
# days. e.g week-based moving averages.
the_day_before = trading_days.get_last_n(time.ToCivil(t),
1,
include_input_date=False)[0]
return (DataEntry(
data_type=base_calc_type,
timestamp=time_util.from_time(time.FromCivil(the_day_before)),
), DataEntry(
data_type=incr_calc_type,
timestamp=time_util.from_time(t),
))
def series_source_inputs_shape(
source_calc_type: DataType.Enum, t: time.Time,
time_spec: calc.CalcTimeSpecs) -> calc.SourceInputs:
# TODO: right now this is still no properly generalized. For time span
# longer than a day, they should all work. However for day and shorter span,
# they need to be checked against trading days and trading hours.
# NOTE: both FromCivil and ToCivil assumes UTC, which later might
# be updated to take into account the exchange timezone of a
# security.
timestamps = [
time.FromCivil(day) for day in trading_days.get_last_n(
time.ToCivil(t), time_spec.num_periods)
]
return [
DataEntry(
data_type=source_calc_type,
timestamp=time_util.from_time(timestmap),
) for timestmap in timestamps
]
def _get_shortest_historical_ranges(start_date: time.CivilTime) -> str:
"""Get the shortest API range string that covers the requested number of
days for the time period ending today.
See: https://iexcloud.io/docs/api/#historical-prices.
"""
num_days = 1 + time.ToCivil(time.Now()).diff(start_date).in_days()
if num_days <= 5:
return '5d'
if num_days <= 28: # Shortest month has 28 days
return '1m'
if num_days <= 90:
return '3m'
if num_days <= 182:
return '6m'
if num_days <= 365:
return '1y'
if num_days <= 730:
return '2y'
if num_days <= 365 * 5 + 1:
return '5y'
if num_days <= 365 * 15 + 3:
return 'max'
raise errors.InvalidArgumentError(
'IEX does not support historical data older than 15 years')
def to_civil(pb: timestamp_pb2.Timestamp) -> time.CivilTime:
return time.ToCivil(to_time(pb))