def test_dividend_000333(self):
pipe = make_dividend_pipeline()
df = run_pipeline(pipe, '2014-1-1','2017-1-10')
result = select_by_sid(df, 333).resample('AS').first()
expected = pd.DataFrame({'股利':[2.0, 1.0, 1.2, 1.0]},
index = pd.date_range('2014', '2017',tz='utc',freq='AS'))
assert_frame_equal(result, expected)
def test_dividend_002006(self):
# 重点测试，无现金分红年度，其值为0.0
pipe = make_dividend_pipeline()
df = run_pipeline(pipe, '2012-1-1','2017-1-10')
result = select_by_sid(df, 2006).resample('AS').first()
expected = pd.DataFrame({'股利':[0.10, 0.00, 0.00, 0.00,0.00,0.02]},
index = pd.date_range('2012', '2017',tz='utc',freq='AS'))
assert_frame_equal(result, expected)
def test_finance_report_000333(self):
# 重点测试日期偏移。默认报告期后45天
pipe = make_finance_report_pipeline()
df = run_pipeline(pipe, '2017-2-14','2017-2-16')
result = select_by_sid(df, 333)
a001 = [1944482.0, 1793125.0, 1793125.0]
b001 = [11707795.0, 15984170.0, 15984170.0]
expected = pd.DataFrame({'货币资金(万元)':a001, '营业总收入(万元)':b001},
index = pd.date_range('2017-02-14', '2017-02-16',tz='utc'))
assert_frame_equal(result, expected)
def test_margin_000333(self):
# 当天查询上一交易日数据
pipe = make_margin_pipeline()
df = run_pipeline(pipe, '2017-11-29','2017-11-30')
result = select_by_sid(df, 333)
# 测试数据类型float64
long_amount = [2976903630.0, 3021909993.0]
total_amount = [2994651757.0, 3038726032.0]
expected = pd.DataFrame({'融资余额(万元)':long_amount,
'融资融券余额(万元)':total_amount},
index = pd.date_range('2017-11-29', '2017-11-30',tz='utc'))
assert_frame_equal(result, expected)
def test_stock_concept(self):
pipe = make_concept_pipeline()
start = '2017-11-27'
end = '2017-11-30'
df = run_pipeline(pipe, start,end)
# 浪潮信息
result_1 = select_by_sid(df, 977)
expected_1 = pd.DataFrame({'大数据':[True] * 4,
'特种药':[False] * 4,
'融资融券':[True] * 4,},
index = pd.date_range(start, end, tz='utc'))
assert_frame_equal(result_1, expected_1)
# 丰原药业
result_2 = select_by_sid(df, 153)
expected_2 = pd.DataFrame({'大数据':[False] * 4,
'特种药':[True] * 4,
'融资融券':[False] * 4,},
index = pd.date_range(start, end, tz='utc'))
assert_frame_equal(result_2, expected_2)
def test_HistoricalZscore(self):
# 因读取原始数据时，并没有考虑调整项。故不得选择送转股期间测试。
# 如果原始数据做同样的调整处理，才会通过测试。
# 关注点：按列计算z值， 单列输入
start = '2017-11-20'
end = '2017-12-1'
stock_codes = ['000001', '000002']
df = run_pipeline(make_pipeline(), start, end)
result = df.loc[end].values
expected = np.array([[-1.59106704, -1.23624812],
[-0.51367549, 1.18221339]])
assert_array_almost_equal(result, expected)
with self.assertRaises(ValueError):
HistoricalZScore(window_length=2, inputs=[USEquityPricing.close])
with self.assertRaises(ValueError):
HistoricalZScore(
window_length=4,
inputs=[USEquityPricing.close, USEquityPricing.volume])
def compute_factors(one_day):
"""
根据每天运行的pipeline结果，近似计算Fama-French 因子
"""
factors = run_pipeline(make_pipeline(), one_day, one_day)
# 获取后续使用的值
returns = factors['returns']
mkt_cap = factors.sort_values(['market_cap'], ascending=True)
be_me = factors.sort_values(['be_me'], ascending=True)
prior_returns = factors['prior_returns']
# 根据市值划分总体，构造6个资产组合
half = int(len(mkt_cap) * 0.5)
small_caps = mkt_cap[:half]
big_caps = mkt_cap[half:]
thirty = int(len(be_me) * 0.3)
seventy = int(len(be_me) * 0.7)
growth = be_me[:thirty]
neutral = be_me[thirty:seventy]
value = be_me[seventy:]
small_value = small_caps.index.intersection(value.index)
small_neutral = small_caps.index.intersection(neutral.index)
small_growth = small_caps.index.intersection(growth.index)
big_value = big_caps.index.intersection(value.index)
big_neutral = big_caps.index.intersection(neutral.index)
big_growth = big_caps.index.intersection(growth.index)
# 假设等权分配资金，取其投资组合回报率平均值
sv = returns[small_value].mean()
sn = returns[small_neutral].mean()
sg = returns[small_growth].mean()
bv = returns[big_value].mean()
bn = returns[big_neutral].mean()
bg = returns[big_growth].mean()
# 按降序排列
# 小市值组合收益率（序列没有列可以指定)
s_p = prior_returns.loc[small_caps.index].sort_values(ascending=False)
# 大市值组合收益率
b_p = prior_returns.loc[big_caps.index].sort_values(ascending=False)
shp = s_p.iloc[:int(len(s_p) * 0.3)].mean() # 小市值组合收益率前70%的均值
bhp = b_p.iloc[:int(len(b_p) * 0.3)].mean() # 大市值组合收益率前70%的均值
slp = s_p.iloc[int(len(s_p) * 0.7):].mean() # 小市值组合收益率后30%的均值
blp = b_p.iloc[int(len(b_p) * 0.7):].mean() # 大市值组合收益率后30%的均值
# 计算 SMB
smb = (sv + sn + sg) / 3 - (bv + bn + bg) / 3
# 计算 HML
hml = (sv + bv) / 2 - (sg + bg) / 2
# 计算 MOM
mom = (shp + bhp) / 2 - (slp + blp) / 2
return smb, hml, mom