def tsa_group_by(tsa, datestring, index_keynames, group_func, interval):
"""
group given tsa by subkeys, and use group_func to aggregate data
the individual timeseries are automatically grouped by timestamp and interval
defined in configuration
parameters:
tsa <TimeseriesArray>
datestring <str> datestring to use to aggregate data TODO: get this from tsa
subkey <tuple> could also be empty, to aggregate everything
group_func <func> like lambda a, b : (a + b) / 2 to get averages
interval <int> interval in seconds the timeseries values should appear
returns:
<TimeseriesArray>
"""
# intermediated tsa
tsa2 = TimeseriesArray(index_keynames=index_keynames,
value_keynames=tsa.value_keynames,
ts_key=tsa.ts_key,
datatypes=tsa.datatypes)
start_ts, _ = DataLogger.get_ts_for_datestring(datestring)
ts_keyname = tsa.ts_key
for data in tsa.export():
# align timestamp
nearest_slot = round((data[ts_keyname] - start_ts) / interval)
data[ts_keyname] = int(start_ts + nearest_slot * interval)
tsa2.add(data, group_func)
return tsa2
def test_export(self):
print("testing export, add")
tsa = TimeseriesArray(meta["index_keynames"],
meta["value_keynames"],
"ts",
datatypes=meta["value_keynames"])
for entry in self.app.export():
tsa.add(entry)
assert tsa == self.app
def load_tsa_raw(self):
"""
read data from raw input files and return TimeseriesArray object
parameters:
datestring <str> isodate representation of date like 2015-12-31
timedelta <int> amount second to correct raw input timestamps
returns:
<TimeseriesArray> object wich holds all data of this day
"""
tsa = TimeseriesArray(self.index_keynames, self.value_keynames, datatypes=self.datatypes)
for rowdict in self.__read_raw_dict():
try:
tsa.add(rowdict)
except ValueError as exc:
logging.exception(exc)
logging.error("ValueError by adding this data to TimeseriesArray: %s", rowdict)
raise exc
except AssertionError as exc:
logging.exception(exc)
logging.error("AssertionError by adding this data to TimeseriesArray: %s", rowdict)
raise exc
return tsa