def test_condition04(self): # corresponds to condition01 with inflation
# test data
thresholds = {}
limits = {}
timeline = [
ud.datestr2ts(strdate)
for strdate in ['2014-06-01', '2015-10-01', '2017-06-01']
]
realSeries = {
'stock1': [(0, 10.0), (1, 11.0), (2, 12.0)],
'stock2': [(0, 10.0), (1, 11.0), (2, 12.0)],
}
param_adjinflat = True
present = datetime.fromtimestamp(timeline[2])
thresholds['stock1'] = ud.findThresholds(realSeries['stock1'],
timeline, param_adjinflat)
thresholds['stock2'] = ud.findThresholds(realSeries['stock2'],
timeline, param_adjinflat)
limits['stock1'] = ud.findLimits(realSeries['stock1'], timeline,
param_adjinflat)
limits['stock2'] = ud.findLimits(realSeries['stock2'], timeline,
param_adjinflat)
param_models = [('MA', None)]
param_prices = ['close']
weights = {(ticker, modelType, priceType): 1.0
for ticker in ['stock2', 'stock1']
for (modelType, _) in param_models
for priceType in param_prices}
ws = sum(weights.values())
nt = len(realSeries.keys())
weights = {key: weights[key] / ws * nt for key in weights}
# expected response
ref_de2twMap = {
0: ('threshold', ('stock1', None, None)),
1: ('threshold', ('stock1', param_adjinflat, present)),
2: ('threshold', ('stock2', None, None)),
3: ('threshold', ('stock2', param_adjinflat, present)),
4: ('weights', ('stock1', 'MA', 'close')),
5: ('weights', ('stock2', 'MA', 'close'))
}
ref_bounds = [(0, 1), (0, 1), (0, 1), (0, 1), (0, 1), (0, 1)]
ref_x0 = np.array([0, 1, 0, 1, 1, 1])
# actual response
bounds, de2twMap, x0 = ud.tw2debv(thresholds, weights, limits,
param_models, param_prices)
# compares expected and actual responses
success = ref_bounds == bounds and ref_de2twMap == de2twMap and np.allclose(
ref_x0, x0)
self.assertTrue(success)
def test_condition06(self): # corresponds to condition02 with inflation
# test data
thresholds = {}
limits = {}
timeline = [
ud.datestr2ts(strdate)
for strdate in ['2014-06-01', '2015-10-01', '2017-06-01']
]
realSeries = {
'stock1': [(0, 10.0), (1, 11.0), (2, 12.0)],
'stock2': [(0, 10.0), (1, 11.0), (2, 12.0)],
}
param_adjinflat = True
present = datetime.fromtimestamp(timeline[2])
thresholds['stock1'] = ud.findThresholds(realSeries['stock1'],
timeline, param_adjinflat)
thresholds['stock2'] = ud.findThresholds(realSeries['stock2'],
timeline, param_adjinflat)
limits['stock1'] = ud.findLimits(realSeries['stock1'], timeline,
param_adjinflat)
limits['stock2'] = ud.findLimits(realSeries['stock2'], timeline,
param_adjinflat)
param_models = [('MA', None), ('ARIMA', None)]
param_prices = ['close', 'open']
weights = {(ticker, modelType, priceType): 1.0
for ticker in ['stock2', 'stock1']
for (modelType, _) in param_models
for priceType in param_prices}
ws = sum(weights.values())
nt = len(realSeries.keys())
weights = {key: weights[key] / ws * nt for key in weights}
bounds, de2twMap, x0 = ud.tw2debv(thresholds, weights, limits,
param_models, param_prices)
# expected response
ref_de2twMap = (thresholds, weights)
# actual response
(_thresholds, _weights) = ud.dev2tw(x0, de2twMap, limits)
# compares expected and actual responses
success = thresholds == _thresholds and weights == _weights
self.assertTrue(success)