def execute(self, strategy):
forecasts = {}
for forecaster in self.forecasters:
forecasts[forecaster.name] = forecaster(strategy).data
forecasts = Panel(forecasts)
mean_fcst = self.normalise(forecasts.mean(axis = 'items'))
return Signal(mean_fcst, [-20, 20], forecasts)
def execute(self, strategy):
prices = strategy.indicator_prices
slow = self.slow(prices)
fast = self.fast(prices)
vol = self.vol(prices)
forecast = self.normalise((fast - slow) / vol)
return Signal(forecast, [-20, 20], Panel({"Fast" : fast, "Slow" : slow, "Volatility" : vol}))
def execute(self, strategy):
measures = {}
for pars in self.par_pairs:
name = "ewmac_{}_{}".format(max(pars), min(pars))
ewmac = EWMAC(EMA(max(pars)), EMA(min(pars)), self.vol)
measures[name] = ewmac(strategy).data
measures = Panel(measures)
return Signal(measures.mean(axis = 'items'), [-20, 20], measures)
def execute(self, strategy):
prices = strategy.indicator_prices
fast_ema = self.fast(prices)
slow_ema = self.slow(prices)
ind_data = strategy.get_empty_dataframe(fill_data = 'Down')
ind_data[fast_ema > slow_ema] = 'Up'
return Signal(ind_data, ['Up', 'Down'], {'Fast':fast_ema, 'Slow':slow_ema})
def execute(self, strategy):
prices = strategy.indicator_prices
fast_ema = self.fast(prices)
mid_ema = self.mid(prices)
slow_ema = self.slow(prices)
levels = (fast_ema > mid_ema) & (mid_ema > slow_ema)
ind_data = strategy.get_empty_dataframe(fill_data = 'Down')
ind_data[levels] = 'Up'
return Signal(ind_data, ['Up', 'Down'], {'Fast':fast_ema, 'Mid':mid_ema, 'Slow':slow_ema})
def execute(self, strategy):
prices = strategy.indicator_prices
breakout = self.breakout(prices)
high = breakout["high"]
low = breakout["low"]
ind_data = strategy.get_empty_dataframe()
# Note: dataframe.where returns the existing values where the
# mask is True, and replaces them with other where False.
# So we need to invert the mask.
# Counterintuitive I think...
ind_data = ind_data.where(~(prices.data == high), 'Up')
ind_data = ind_data.where(~(prices.data == low), 'Down')
ind_data = ind_data.ffill()
# We need to remove any remaining Nans (at the start of the df),
# otherwise later string comparisons will throw an error because
# it thinks it is a mixed datatype.
ind_data = ind_data.fillna('-')
return Signal(ind_data, ['Up', 'Down'], breakout)
def execute(self, strategy):
prices = strategy.indicator_prices
indicator = self.measure(prices)
vol = self.vol(prices)
forecast = self.normalise((self.base - indicator) / vol)
return Signal(forecast, [-20, 20], Panel({"Base" : self.base, "Ind" : indicator, "Volatility" : vol}))
def execute(self, strategy):
prices = strategy.indicator_prices
fast = self.fast(prices)
slow = self.slow(prices)
return Signal(fast - slow, ['forecast'], {'fast':fast, 'slow':slow})