def get_demand_raw(self, _from, to):
"""Use for demand chart."""
validate_date(_from)
validate_date(to)
from_adj = _from
to_adj = to
dayly = pd.date_range(from_adj, to_adj, freq="D", tz=pytz.utc)
values = [self.get_day_evaporation_on(date) for date in dayly]
ts = pd.Series(values, dayly, name="evaporation")
# use fillna with ffill instead of ts.interpolate() to keep the data
# as accurate as possible
ts.fillna(method="ffill", inplace=True)
# divide by amount of quarter hours in a day
return ts[from_adj:to_adj]
def get_rain(self, which, _from, to, *args, **kwargs):
validate_date(_from)
validate_date(to)
rain_filter_id = self.basin.rain_filter_id
rain_location_id = self.basin.rain_location_id
rain = self._get_timeseries_as_pd_series(
rain_filter_id,
rain_location_id,
RAIN_PARAMETER_IDS[which],
_from, to, 'rain_' + which
)
# convert to quarterly figures
rain = rain.resample('15min', fill_method='ffill')
# 4 quarters in an hour
if which != 'kwadrant':
rain /= 4
return rain
def get_fill(self, _from, to, *args, **kwargs):
validate_date(_from)
validate_date(to)
# basin parameters
fill_filter_id = self.basin.filter_id
fill_location_id = self.basin.location_id
fill_parameter_id = self.basin.parameter_id
# max_storage can come from request, therefore from self.constants
max_storage = self.constants.max_storage
# values are given in percentage as to max_storage
ts = self._get_timeseries_as_pd_series(
fill_filter_id,
fill_location_id,
fill_parameter_id,
_from, to, 'fill'
)
ts = self.fill_pct_to_m3(ts, max_storage)
return ts
def get_demand(self, _from, to):
"""Use for prediction chart. Convert demand from m2 to m3."""
validate_date(_from)
validate_date(to)
# ensure we deal with broader range to avoid resample edge problems
from_adj = _from
to_adj = to
dayly = pd.date_range(from_adj, to_adj, freq="D", tz=pytz.utc)
values = [self.get_day_evaporation_on(date) for date in dayly]
ts = pd.Series(values, dayly, name="evaporation")
ts = ts.resample("15min")
# use fillna with ffill instead of ts.interpolate() to keep the data
# as accurate as possible
ts.fillna(method="ffill", inplace=True)
# divide by amount of quarter hours in a day
ts /= 24 * 4
ts *= float(0.001)
ts *= self.crop_surface
ts *= self.basin.recirculation
result = ts[from_adj:to_adj]
return result
def predict_fill(self, _from, to, outflow_open=None,
outflow_closed=None, outflow_capacity=0):
# do some input validation here, to ensure we are dealing with sane
# numbers
validate_date(_from)
validate_date(to)
rain_mean = self.fews_data.get_rain('mean', _from, to)
# create a no rain series
periods = (to - _from).days * 4 * 24 + 1
values = np.zeros(periods)
dates = pd.date_range(_from, periods=periods, freq='15min',
tz=pytz.utc)
rain_zero = pd.Series(values, dates, name='uitstroom')
# gebruik de datum van de laatst beschikbaar regenvoorspelling als
# from en to waarden
_from_rain = rain_mean.index[0]
to_rain = rain_mean.index[-1]
# retrieve fill: just take any data we have,
# so we can compare measurements with predictions
current_fill = self.fews_data.get_current_fill(to_rain)
current_fill_m3 = current_fill['current_fill_m3']
# bereken watervraag over deze periode
demand_m3_rain = self.demand_table.get_demand(_from_rain, to_rain)
demand_m3_zero_rain = self.demand_table.get_demand(_from, to)
# leidt aantal periodes af uit een vd 'input' tijdreeksen
periods_rain = len(rain_mean)
# bereken uitstroom
max_uitstroom_m3 = self.calc_max_uitstroom(_from_rain, periods_rain)
if len(max_uitstroom_m3) != len(rain_mean):
raise Exception('%s != %s' % (len(max_uitstroom_m3),
len(rain_mean)))
# return ook tussentijdse waarden, vnml. voor debugging
result = {
'scenarios': {},
'history': fill_m3_to_pct(
current_fill['fill_history_m3'],
self.constants.max_storage),
'current_fill': fill_m3_to_pct(
current_fill_m3, self.constants.max_storage),
'intermediate': {
'rain_mean': rain_mean,
'max_uitstroom': max_uitstroom_m3,
'demand': demand_m3_rain,
}
}
# bereken de drie scenarios
calc_scenarios = {
'no_rain': (rain_zero, demand_m3_zero_rain),
'mean': (rain_mean, demand_m3_rain),
}
for scenario, (rain, demand_m3_) in calc_scenarios.items():
result['scenarios'][scenario] = self.predict_scenario(
_from, current_fill_m3, demand_m3_, rain, outflow_open,
outflow_closed, outflow_capacity)
return result
