def _mod_time_dim(time_dim: pd.DatetimeIndex, freq: str):
"""
Modify Time dimension so it fits the requirements of the "resample_for_gsee" function
Parameters
----------
time_dim: array
with datetime entries
freq: string
representing data frequency of na_time
Returns
-------
array
modified time dimension
"""
if freq == 'A':
# Annual data is set to the beginning of the year
return time_dim.map(lambda x: pd.Timestamp(
year=x.year, month=1, day=1, hour=0, minute=0))
elif freq in ['S', 'M']:
# Seasonal data is set to middle of month, as it is often represented with the day in the middle of the season.
# Monthly data is set to middle of month
return time_dim.map(
lambda x: pd.Timestamp(year=x.year,
month=x.month,
day=int(monthrange(x.year, x.month)[1] / 2),
hour=0,
minute=0))
elif freq == 'D':
# Daily data is set to 00:00 hours of the day
return time_dim.map(lambda x: pd.Timestamp(
year=x.year, month=x.month, day=x.day, hour=0, minute=0))
else:
return time_dim
def test_map_bug_1677(self):
index = DatetimeIndex(['2012-04-25 09:30:00.393000'])
f = index.asof
result = index.map(f)
expected = Index([f(index[0])])
tm.assert_index_equal(result, expected)
def test_map_bug_1677(self):
index = DatetimeIndex(['2012-04-25 09:30:00.393000'])
f = index.asof
result = index.map(f)
expected = Index([f(index[0])])
tm.assert_index_equal(result, expected)
def get_index_by_days_dict(timestamps: pd.DatetimeIndex, dates = None):
"""Create dictionary with dates as index an intraday timestamps as values
"""
if(dates is None):
dates = timestamps.map(pd.Timestamp.date)
indexes_by_day_dict = timestamps.groupby(dates)
return indexes_by_day_dict
def get_timestamps_by_trading_session_dict(self, timestamps: pd.DatetimeIndex):
"""Create dictionary with dates as index an intraday timestamps as values
"""
## TODO: Take iso into consideration here.
dates = timestamps.map(pd.Timestamp.date)
indexes_by_day_dict = timestamps.groupby(dates)
return indexes_by_day_dict
