def time_series_aggregated(
self, aggregation, aggregation_period,
dt_start=None, dt_end=None):
"""
Aggregated time series.
Result is a dictionary with timeseries. The keys are: (location,
parameter, option), where option is (GraphItemMixin.)TIME_SERIES_ALL,
TIME_SERIES_POSITIVE or TIME_SERIES_NEGATIVE
aggregation: see PredefinedGraph
'avg', 'sum'
aggregation_period: see PredefinedGraph
'day', 'month', 'quarter', 'year'
"""
source = self.fews_norm_source
series = self.series()
result = {}
for single_series in series:
obj = {
GraphItemMixin.TIME_SERIES_ALL: timeseries.TimeSeries(),
GraphItemMixin.TIME_SERIES_POSITIVE: timeseries.TimeSeries(),
GraphItemMixin.TIME_SERIES_NEGATIVE: timeseries.TimeSeries()}
# Somehow get the events with aggregation in it
# aggregation == PredefinedGraph.AGGREGATION_AVG, AGGREGATION_SUM
# aggregation_period == PredefinedGraph.PERIOD_YEAR, PERIOD_MONTH,
# PERIOD_QUARTER, PERIOD_DAY
# between dt_start and dt_end
events = Event.agg_from_raw(
single_series, dt_start=dt_start, dt_end=dt_end,
schema_prefix=source.database_schema_name,
agg_function=aggregation,
agg_period=aggregation_period).using(source.database_name)
# Put the events in Timeseries objects in obj
for event in events:
obj[GraphItemMixin.TIME_SERIES_ALL][event.timestamp] = (
event.value, event.flag, event.comment)
if event.value >= 0:
obj[GraphItemMixin.TIME_SERIES_POSITIVE][
event.timestamp] = (
event.value, event.flag, event.comment)
else:
obj[GraphItemMixin.TIME_SERIES_NEGATIVE][
event.timestamp] = (
event.value, event.flag, event.comment)
# Now put the timeseries in the result
for k in obj.keys():
obj[k].location_id = single_series.location
obj[k].parameter_id = single_series.parameter
obj[k].time_step = single_series.timestep
obj[k].units = single_series.unit
result[single_series.location,
single_series.parameter, k] = obj[k]
return result
