def __init__(self,
provider: str,
level: int,
cap: CompactFactor = None,
db_interface: DBInterface = None):
self.db_interface = db_interface if db_interface else get_db_interface(
)
self.calendar = TradingCalendar(self.db_interface)
self.date = self.calendar.yesterday()
self.data_reader = AShareDataReader(self.db_interface)
self.industry = self.data_reader.industry(provider=provider,
level=level)
self.cap = cap if cap else self.data_reader.stock_free_floating_market_cap
def major_index_valuation(db_interface: DBInterface = None):
if db_interface is None:
db_interface = get_db_interface()
data = db_interface.read_table('指数日行情',
['市盈率TTM', '市净率']).dropna(how='all')
tmp = data.groupby('ID').rank()
latest = data.groupby('ID').tail(1)
percentile = tmp.groupby('ID').tail(1) / tmp.groupby('ID').max()
percentile.columns = [f'{it}分位' for it in percentile.columns]
ret = pd.concat([latest, percentile], axis=1)
ret = ret.loc[:, sorted(ret.columns)].reset_index()
index_name_dict = dict(
zip(constants.STOCK_INDEXES.values(), constants.STOCK_INDEXES.keys()))
ret['ID'] = ret['ID'].map(index_name_dict)
return ret.set_index(['DateTime', 'ID'])
def __init__(self,
date: dt.datetime = None,
db_interface: DBInterface = None):
self.db_interface = db_interface if db_interface else get_db_interface(
)
self.calendar = TradingCalendar(self.db_interface)
if date is None:
date = dt.datetime.combine(dt.date.today(), dt.time())
self.date = date
records = utils.load_excel('自编指数配置.xlsx')
self.tickers = [it['ticker'] for it in records]
self.tbd_indexes = list(
set(self.tickers) - set(self.must_keep_indexes))
start_date = self.calendar.offset(date, -22)
index_factor = ContinuousFactor('自合成指数',
'收益率',
db_interface=self.db_interface)
self.cache = index_factor.get_data(start_date=start_date,
end_date=date).unstack()
self.industry_cache = []
def plot_factor_return(factor_name: str,
weight: bool = True,
industry_neutral: bool = True,
bins: int = 5,
start_date: DateUtils.DateType = None,
end_date: DateUtils.DateType = None,
db_interface: DBInterface = None) -> plt.Figure:
if db_interface is None:
db_interface = get_db_interface()
ids = utils.generate_factor_bin_names(factor_name,
weight=weight,
industry_neutral=industry_neutral,
bins=bins)
data = db_interface.read_table('因子分组收益率',
ids=ids,
start_date=start_date,
end_date=end_date)
df = (data.unstack() + 1).cumprod()
bin_names_info = [utils.decompose_bin_names(it) for it in df.columns]
diff_series = df[ids[0]] - df[ids[-1]]
df.columns = [it['group'] for it in bin_names_info]
diff_series.name = f'{utils.decompose_bin_names(ids[0])["group"]}-{utils.decompose_bin_names(ids[-1])["group"]}'
fig, axes = plt.subplots(2, 1, figsize=(15, 8), sharex='col')
df.plot(ax=axes[0])
industry_neutral_str = '行业中性' if industry_neutral else '非行业中性'
weight_str = '市值加权' if weight else '等权'
axes[0].set_title(
f'{factor_name} 分组收益率({industry_neutral_str}, {weight_str})')
plot_dt = df.index.get_level_values('DateTime')
axes[0].set_xlim(left=plot_dt[0], right=plot_dt[-1])
axes[0].grid(True)
diff_series.plot(ax=axes[1])
axes[1].grid(True)
axes[1].legend()
return fig
def setUp(self) -> None:
set_global_config('config.json')
self.db_interface = get_db_interface()