def load_factor(self, fact_name: str, **kwargs): """ 优先直接获取数据--否则数据库调取--最后实时计算 :param fact_name: :param kwargs: :return: """ if kwargs.get('factor_value', None) is None: # self.db.query_factor_data("EP_ttm", "Fin") if kwargs['cal']: try: fact_raw_data = self.Factor.factor[ fact_name + '_data_raw'](**kwargs['factor_params']) # TODO self.data_input["factor_raw_data"] = fact_raw_data except Exception as e: print(e) print( f"{dt.datetime.now().strftime('%X')}: Unable to load raw data that to calculate factor!" ) return else: factor_class = self.Factor.factor[fact_name]( data=self.data_input["factor_raw_data"].copy( deep=True), **kwargs['factor_params']) else: factor_data_ = self.db.query_factor_data( factor_name=fact_name, db_name=kwargs['db_name']) print( f"{dt.datetime.now().strftime('%X')}: Get factor data from MySQL!" ) factor_data_.set_index( [KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) factor_class = FactorInfo() factor_class.data = factor_data_[fact_name] factor_class.factor_name = fact_name else: print( f"{dt.datetime.now().strftime('%X')}: Get factor data from input!" ) kwargs['factor_value'].set_index( [KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) factor_class = FactorInfo() factor_class.data = kwargs['factor_value'][fact_name] factor_class.factor_name = fact_name self.fact_name = factor_class.factor_name self.factor_dict[self.fact_name] = factor_class
def Value002(cls, data: pd.DataFrame, net_profit_cut: str = FISN.Net_Pro_Cut.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False): """ 扣非市盈率倒数(EP_cut_TTM): 市盈率(扣除非经常性损益)倒数 :param data: :param net_profit_cut: :param total_mv: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[net_profit_cut] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def HighFreq035(cls, data: pd.DataFrame, n: int = 21): """ 博弈因子(Stren):存在涨停主卖为零情况,会导致分母为0,根据数值特征范围将分母为零的计算设置为2 """ factor_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[ 'buy_amount'] = data['BuyAll_AM_120min'] + data['BuyAll_PM_120min'] data['sale_amount'] = data['SaleAll_AM_120min'] + data[ 'SaleAll_PM_120min'] # 升序 w = cls.Half_time(n) data[['buy_amount_w', 'sale_amount_w']] = data[['buy_amount', 'sale_amount']].groupby( KN.STOCK_ID.value, group_keys=False).rolling( n, min_periods=n).apply(lambda x: (x * w).sum()) data[factor_name] = data['buy_amount_w'] / data['sale_amount_w'] # 无穷大值设置为2 data[factor_name][np.isinf(data[factor_name])] = 2 F = FactorInfo() F.data = data[factor_name] F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Distribution031(cls, data: pd.DataFrame, n: int = 20, **kwargs): """ N日上午收益与下午收益差和(AVP) """ factor_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) # Calculate AM and PM returns on stocks data['am_ret_stock'] = data['2hPrice'] / data['open'] - 1 data['pm_ret_stock'] = data['4hPrice'] / data['2hPrice'] - 1 data['diff'] = data['am_ret_stock'] - data['pm_ret_stock'] # filter momentum data[factor_name] = data['diff'].groupby( KN.STOCK_ID.value, group_keys=False).rolling(n).sum() F = FactorInfo() F.data = data[factor_name] F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Value004(cls, data: pd.DataFrame, operator_income: str = FISN.Op_Income.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False): """ 市销率倒数(TTM)(SP_TTM):市销率倒数 :param data: :param operator_income: :param total_mv: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[operator_income] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Value008(cls, data: pd.DataFrame, Surplus_Reserves: str = FBSN.Surplus_Reserves.value, Undistributed_Profit: str = FBSN.Undistributed_Profit.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False): """ 股息率倒数(DP_TTM):股息率(近12个月现金红利和) 股息 = 期末留存收益 - 期初留存收益 留存收益 = 盈余公积 + 未分配利润 """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data["RE"] = data[Surplus_Reserves] + data[Undistributed_Profit] data[func_name] = data["RE"] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Solvency010(cls, data: pd.DataFrame, currency: str = FBSN.Currency.value, tradable_asset: str = FBSN.Tradable_Asset.value, op_net_cash_flow: str = FCFSN.Op_Net_CF.value, short_borrow: str = FBSN.ST_Borrow.value, short_bond_payable: str = FBSN.ST_Bond_Payable.value, short_iliq_liability_1y: str = FBSN.ST_IL_LB_1Y.value, switch: bool = False) -> FactorInfo: """ 短期偿债能力指标1(ShortDebt1_CFPA_qoq):(现金及现金等价物 + TTM经营性现金流)/短期有息负债 现金及现金等价物 = 货币资金 + 交易性金融资产 经营性现金流 = 经营性现金流量净额 短期有息负债 = 短期借款 + 短期应付债券 + 一年内到期的非流动负债 :param data: :param currency: :param tradable_asset: :param op_net_cash_flow: :param short_borrow: :param short_bond_payable: :param short_iliq_liability_1y: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([SN.REPORT_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) # 短期偿债能力指标 ShortDebt1_CFPA = data[[currency, tradable_asset, op_net_cash_flow]].sum(skipna=True, axis=1) / \ data[[short_borrow, short_bond_payable, short_iliq_liability_1y]].sum(skipna=True, axis=1) # switch inf to Nan ShortDebt1_CFPA[np.isinf(ShortDebt1_CFPA)] = np.nan data[func_name] = ShortDebt1_CFPA.groupby( KN.STOCK_ID.value).apply(lambda x: x.diff(1) / abs(x.shift(1))) if switch: data_fact = cls()._switch_freq(data_=data, name=func_name) else: data_fact = None data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Value009(cls, data: pd.DataFrame, operator_income: str = FISN.Op_Income.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False): """ 企业价值倍数倒数(最新财报,扣除现金)(EV2EBITDA_LR):企业价值(扣除现金)/息税折旧摊销前利润 企业价值 = 总市值 + 负债总计 - 无息负债 - 货币资金 :param data: :param operator_income: :param total_mv: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[operator_income] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Momentum009(cls, data: pd.DataFrame, close_price: str = PVN.CLOSE.value, bm_price: str = 'index_close', n: int = 20): """ 市场alpha因子 """ factor_name = sys._getframe().f_code.co_name + f'_{n}' data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) # ret data = data.groupby(KN.STOCK_ID.value).pct_change().dropna() data_new = data.groupby( KN.STOCK_ID.value, group_keys=False).apply(lambda x: cls._reg_rolling( x, bm_price, close_price, False, True, n)) data_new.name = factor_name F = FactorInfo() F.data = data_new F.factor_type = 'MTM' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Momentum025(cls, data: pd.DataFrame, high_price: str = PVN.HIGH.value, close_price: str = PVN.CLOSE.value, n: int = 1) -> FactorInfo: """ 动量CTH收益率绝对值均值(MTM_CTH_abs):N日收盘价与最高价收益率绝对值均值 :return: """ factor_name = sys._getframe().f_code.co_name + f'_{n}' data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data['return'] = data.groupby( KN.STOCK_ID.value, group_keys=False).apply( lambda x: abs(x[close_price] / x[high_price] - 1)) data[factor_name] = data['return'].groupby( KN.STOCK_ID.value, group_keys=False).rolling(n, min_periods=1).mean() F = FactorInfo() F.data = data[factor_name] F.factor_type = 'MTM' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Momentum013(cls, data: pd.DataFrame, close_price: str = PVN.CLOSE.value, n: int = 1) -> FactorInfo: """ 动量CTC收益率(MTM_CTC):N日收盘价计算的收益率均值 :return: """ factor_name = sys._getframe().f_code.co_name + f'_{n}' data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data['return'] = data[close_price].groupby( KN.STOCK_ID.value).pct_change() data[factor_name] = data['return'].groupby( KN.STOCK_ID.value, group_keys=False).rolling(n, min_periods=1).mean() F = FactorInfo() F.data = data[factor_name] F.factor_type = 'MTM' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Momentum017(cls, data: pd.DataFrame, open_price: str = PVN.OPEN.value, close_price: str = PVN.CLOSE.value, n: int = 1) -> FactorInfo: """ 动量OTC收益率均值(MTM_OTC):N日开盘价与收盘价收益率均值 :return: """ factor_name = sys._getframe().f_code.co_name + f'_{n}' data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data['return'] = data.groupby( KN.STOCK_ID.value, group_keys=False).apply(lambda x: (x[ open_price] / x[close_price].shift(1) - 1).shift(-1)) data[factor_name] = data['return'].groupby( KN.STOCK_ID.value, group_keys=False).rolling(n, min_periods=1).mean() F = FactorInfo() F.data = data[factor_name] F.factor_type = 'MTM' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Profit013(cls, data: pd.DataFrame, net_profit_in: str = FISN.Net_Pro_In.value, total_asset: str = FBSN.Total_Asset.value, switch: bool = False): """ 总资产净利率(TTM)(ROA_TTM) """ func_name = sys._getframe().f_code.co_name data.set_index([SN.REPORT_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[net_profit_in] / data[total_asset] data[func_name][np.isinf(data[func_name])] = np.nan if switch: data_fact = cls()._switch_freq(data_=data, name=func_name) else: data_fact = None data.reset_index(inplace=True) F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Momentum012(cls, data: pd.DataFrame, price: str = PVN.CLOSE.value, n: int = 20): """ 路径动量因子(MTM_PathLen) """ factor_name = sys._getframe().f_code.co_name + f'_{n}' data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data['stand'] = data[price].groupby( KN.STOCK_ID.value, group_keys=False).apply( lambda x: (x - x.rolling(n, min_periods=1).mean()) / x.rolling( n, min_periods=1).std(ddof=1)) data['stand'][np.isinf(data['stand'])] = np.nan data['diff'] = data['stand'].groupby(KN.STOCK_ID.value).diff(1).abs() data[factor_name] = data['diff'].groupby(KN.STOCK_ID.value, group_keys=False).rolling( n, min_periods=1).sum() F = FactorInfo() F.data = data[factor_name] F.factor_type = 'MTM' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Value014(cls, data: pd.DataFrame, free_cash_flow: str = FCFSN.Free_Cash_Flow.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False): """ 市现率倒数(自由现金流,TTM)(FCFP_TTM):市现率倒数(自由现金流) :param data: :param free_cash_flow: :param total_mv: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[free_cash_flow] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Profit026(cls, data: pd.DataFrame, net_profit_in: str = FISN.Net_Pro_In.value, operator_income: str = FISN.Op_Income.value, switch: bool = False): """ 当期净利润率(NP) """ func_name = sys._getframe().f_code.co_name data.set_index([SN.REPORT_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[net_profit_in] / data[operator_income] data[np.isinf(data[func_name])] = 0 if switch: data_fact = cls()._switch_freq(data_=data, name=func_name) else: data_fact = None data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Value011(cls, data: pd.DataFrame, net_asset_ex: str = FBSN.Net_Asset_Ex.value, total_mv: str = PVN.TOTAL_MV.value, switch: bool = False) -> FactorInfo: """ 市净率倒数(TTM)(BP_TTM):市净率的倒数 :param data: :param net_asset_ex: :param total_mv: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) data[func_name] = data[net_asset_ex] / data[total_mv] data_fact = data[func_name].copy(deep=True) data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def VolPrice017(cls, data: pd.DataFrame, n: int = 20, **kwargs): """PMA 特殊""" factor_name = sys._getframe().f_code.co_name data.set_index([KN.TRADE_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) # Calculate AM and PM returns on stocks data['am_ret_stock'] = data['2hPrice'] / data['open'] - 1 data['pm_ret_stock'] = data['4hPrice'] / data['2hPrice'] - 1 # filter momentum data[factor_name] = data.groupby( KN.TRADE_DATE.value, group_keys=False).apply( lambda x: cls._reg(x, 'am_ret_stock', 'pm_ret_stock')) # data['mean'] = data['res'].groupby(KN.STOCK_ID.value, # group_keys=False).rolling(n, min_periods=1).apply(np.nanmean) # data['std'] = data['res'].groupby(KN.STOCK_ID.value, # group_keys=False).rolling(n, min_periods=1).apply(np.nanstd) # data[factor_name] = data['mean'] / data['std'] # data[factor_name][np.isinf(data[factor_name])] = 0 data[factor_name] = data['pm_ret_stock'] F = FactorInfo() F.data = data[factor_name] F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = factor_name return F
def Solvency012(cls, data: pd.DataFrame, total_asset: str = FBSN.Total_Asset.value, currency: str = FBSN.Currency.value, tradable_asset: str = FBSN.Tradable_Asset.value, short_borrow: str = FBSN.ST_Borrow.value, short_bond_payable: str = FBSN.ST_Bond_Payable.value, short_iliq_liability_1y: str = FBSN.ST_IL_LB_1Y.value, switch: bool = False): """ 短期偿债能力指标3(ShortDebt3_CFPA_qoq):(现金及现金等价物 - 短期有息负债)/ 总资产 现金及现金等价物 = 货币资金 + 交易性金融资产 短期有息负债 = 短期借款 + 短期应付债券 + 一年内到期的非流动负债 :param data: :param total_asset: :param currency: :param tradable_asset: :param short_borrow: :param short_bond_payable: :param short_iliq_liability_1y: :param switch: :return: """ func_name = sys._getframe().f_code.co_name data.set_index([SN.REPORT_DATE.value, KN.STOCK_ID.value], inplace=True) data.sort_index(inplace=True) x1 = data[[currency, tradable_asset]].sum(skipna=True, axis=1) x2 = data[[short_borrow, short_bond_payable, short_iliq_liability_1y]].sum(skipna=True, axis=1) y = data[total_asset] # 短期偿债能力指标 ShortDebt2_CFPA = (x1 - x2) / y data[func_name] = ShortDebt2_CFPA.groupby( KN.STOCK_ID.value).apply(lambda x: x.diff(1) / abs(x.shift(1))) # switch inf to Nan data[func_name][np.isinf(data[func_name])] = np.nan if switch: data_fact = cls()._switch_freq(data_=data, name=func_name) else: data_fact = None data = data.reset_index() F = FactorInfo() F.data_raw = data[[ SN.ANN_DATE.value, KN.STOCK_ID.value, SN.REPORT_DATE.value, func_name ]] F.data = data_fact F.factor_type = data['type'][0] F.factor_category = cls().__class__.__name__ F.factor_name = func_name return F
def Distribution027(cls, data: pd.DataFrame, **kwargs): """N日分钟振幅波动稳定性(HFD_ret_std)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def Distribution008(cls, data: pd.DataFrame, **kwargs): """高频量价相关性(HFD_Corr_Vol_P)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def VolPrice009(cls, data: pd.DataFrame, **kwargs): """改进反转(Rev_improve)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def VolPrice008(cls, data: pd.DataFrame, **kwargs): """大单驱动涨幅(MOM_bigOrder)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def Distribution019(cls, data: pd.DataFrame, **kwargs): """成交量差分绝对值均值(Vol_diff_abs_mean)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def VolPrice011(cls, data: pd.DataFrame, **kwargs): """聪明钱因子(SmartQ)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def Distribution007(cls, data: pd.DataFrame, **kwargs): """高频下行波动占比(HFD_std_down_occupy)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def Distribution009(cls, data: pd.DataFrame, **kwargs): """高频收益偏度(HFD_ret_skew)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def VolPrice013(cls, data: pd.DataFrame, **kwargs): """轨迹非流动因子(Illiq_Track)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def VolPrice012(cls, data: pd.DataFrame, **kwargs): """高频反转因子(HFD_Rev)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F
def Distribution024(cls, data: pd.DataFrame, **kwargs): """均匀分布主动占比因子(Event_Amt_R)""" F = FactorInfo() F.data = data F.factor_type = 'HFD' F.factor_category = cls().__class__.__name__ F.factor_name = data.name return F