def factor_definition(self):
s = time.time()
needData = self.needData # 计算所需数据
needData.loc[needData[t.TRDSTAT].isin([5, 6]),
t.PCTCHG] = np.nan # 将停牌对应 股票收益率 设为 NaN
factor = -Calculator.Sumif(x=needData[t.PCTCHG],
condition=needData[t.PCTCHG] > 0,
num=5) # 计算5日动量
factor = factor.to_frame()
factor.columns = [self.factorName]
print('factor {0} done with {1} seconds'.format(
self.factorName,
time.time() - s))
return factor
def test_return_shape(self):
"""
测试 返回值形状
:return:
"""
calculator = Calculator()
simDataX = simDataGen(colnames=('colX', ))
simDataY = simDataGen(colnames=('colY', ))
dataShape = simDataX.shape
# test functions
self.assertTrue(
np.all(dataShape == calculator.Delay(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Sumif(x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
condition=simDataX > 0,
retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Diff(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Max(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Min(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Mean(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Sum(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Var(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Std(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Wma(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Sma(
x=simDataX, n=3, m=2, by=ALIAS_FIELDS.STKCD,
retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Decaylinear(
x=simDataX, d=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Rank(
x=simDataX, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.TsToMax(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.TsToMin(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.FindRank(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD, retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.Corr(x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.RegAlpha(x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.RegBeta(x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
retDf=True).shape))
self.assertTrue(
np.all(dataShape == calculator.RegResi(x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
retDf=True).shape))
print('DataFrame return shape test passed')
self.assertTrue(
np.all(dataShape[0] == calculator.Delay(
x=simDataX, num=3, by=ALIAS_FIELDS.STKCD).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Sumif(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
condition=simDataX > 0,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Diff(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Max(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Min(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Mean(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Sum(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Var(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Std(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Wma(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Sma(
x=simDataX,
n=3,
m=2,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Decaylinear(
x=simDataX,
d=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Rank(
x=simDataX,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.TsToMax(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.TsToMin(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.FindRank(
x=simDataX,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.Corr(
x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.RegAlpha(
x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.RegBeta(
x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
self.assertTrue(
np.all(dataShape[0] == calculator.RegResi(
x=simDataX,
y=simDataY,
num=3,
by=ALIAS_FIELDS.STKCD,
).shape[0]))
print('Series return shape test passed')