跳转至

process

frequency_controller

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
class frequency_controller(object):
    def __init__(self, freq: str):
        self.homeplace = HomePlace()
        self.freq = freq

        if freq == "M":
            self.counts_one_year = 12
            self.time_shift = pd.DateOffset(months=1)
            self.states_files = (
                self.homeplace.daily_data_file + "states_monthly.parquet"
            )
            self.sts_files = self.homeplace.daily_data_file + "sts_monthly.parquet"
            self.comment_name = "月"
            self.days_in = 20
        elif freq == "W":
            self.counts_one_year = 52
            self.time_shift = pd.DateOffset(weeks=1)
            self.states_files = self.homeplace.daily_data_file + "states_weekly.parquet"
            self.sts_files = self.homeplace.daily_data_file + "sts_weekly.parquet"
            self.comment_name = "周"
            self.days_in = 5
        else:
            raise ValueError("'暂时不支持此频率哈🤒，目前仅支持月频'M'，和周频'W'")

    def next_end(self, x):
        """找到下个周期的最后一天"""
        if self.freq == "M":
            return x + pd.DateOffset(months=1) + pd.tseries.offsets.MonthEnd()
        elif self.freq == "W":
            return x + pd.DateOffset(weeks=1)
        else:
            raise ValueError("'暂时不支持此频率哈🤒，目前仅支持月频'M'，和周频'W'")

next_end(x)

找到下个周期的最后一天

Source code in pure_ocean_breeze/labor/process.py
Python
1426
1427
1428
1429
1430
1431
1432
1433
def next_end(self, x):
    """找到下个周期的最后一天"""
    if self.freq == "M":
        return x + pd.DateOffset(months=1) + pd.tseries.offsets.MonthEnd()
    elif self.freq == "W":
        return x + pd.DateOffset(weeks=1)
    else:
        raise ValueError("'暂时不支持此频率哈🤒，目前仅支持月频'M'，和周频'W'")

pure_coldwinter

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
class pure_coldwinter(object):
    # DONE: 可以自由添加其他要剔除的因子，或者替换某些要剔除的因子

    @classmethod
    @lru_cache(maxsize=None)
    def __init__(
        cls,
        momentum: bool = 1,
        earningsyield: bool = 1,
        growth: bool = 1,
        liquidity: bool = 1,
        size: bool = 1,
        leverage: bool = 1,
        beta: bool = 1,
        nonlinearsize: bool = 1,
        residualvolatility: bool = 1,
        booktoprice: bool = 1,
    ) -> None:
        """读入10种常用风格因子，并可以额外加入其他因子

        Parameters
        ----------
        facs_dict : Dict, optional
            额外加入的因子，名字为key，因子矩阵为value，形如`{'反转': ret20, '换手': tr20}`, by default None
        momentum : bool, optional
            是否删去动量因子, by default 1
        earningsyield : bool, optional
            是否删去盈利因子, by default 1
        growth : bool, optional
            是否删去成长因子, by default 1
        liquidity : bool, optional
            是否删去流动性因子, by default 1
        size : bool, optional
            是否删去规模因子, by default 1
        leverage : bool, optional
            是否删去杠杆因子, by default 1
        beta : bool, optional
            是否删去贝塔因子, by default 1
        nonlinearsize : bool, optional
            是否删去非线性市值因子, by default 1
        residualvolatility : bool, optional
            是否删去残差波动率因子, by default 1
        booktoprice : bool, optional
            是否删去账面市值比因子, by default 1
        """
        cls.homeplace = HomePlace()
        # barra因子数据
        styles = os.listdir(cls.homeplace.barra_data_file)
        styles = [i for i in styles if (i.endswith(".parquet")) and (i[0] != ".")]
        barras = {}
        for s in styles:
            k = s.split(".")[0]
            v = pd.read_parquet(cls.homeplace.barra_data_file + s).resample("M").last()
            barras[k] = v
        rename_dict = {
            "size": "市值",
            "nonlinearsize": "非线性市值",
            "booktoprice": "估值",
            "earningsyield": "盈利",
            "growth": "成长",
            "leverage": "杠杆",
            "liquidity": "流动性",
            "momentum": "动量",
            "residualvolatility": "波动率",
            "beta": "贝塔",
        }
        if momentum == 0:
            barras = {k: v for k, v in barras.items() if k != "momentum"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "momentum"}
        if earningsyield == 0:
            barras = {k: v for k, v in barras.items() if k != "earningsyield"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "earningsyield"}
        if growth == 0:
            barras = {k: v for k, v in barras.items() if k != "growth"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "growth"}
        if liquidity == 0:
            barras = {k: v for k, v in barras.items() if k != "liquidity"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "liquidity"}
        if size == 0:
            barras = {k: v for k, v in barras.items() if k != "size"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "size"}
        if leverage == 0:
            barras = {k: v for k, v in barras.items() if k != "leverage"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "leverage"}
        if beta == 0:
            barras = {k: v for k, v in barras.items() if k != "beta"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "beta"}
        if nonlinearsize == 0:
            barras = {k: v for k, v in barras.items() if k != "nonlinearsize"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "nonlinearsize"}
        if residualvolatility == 0:
            barras = {k: v for k, v in barras.items() if k != "residualvolatility"}
            rename_dict = {
                k: v for k, v in rename_dict.items() if k != "residualvolatility"
            }
        if booktoprice == 0:
            barras = {k: v for k, v in barras.items() if k != "booktoprice"}
            rename_dict = {k: v for k, v in rename_dict.items() if k != "booktoprice"}
        facs_dict = {
            "反转_20天收益率均值": boom_one(read_daily(ret=1)),
            "波动_20天收益率标准差": read_daily(ret=1)
            .rolling(20, min_periods=10)
            .std()
            .resample("M")
            .last(),
            "换手_20天换手率均值": boom_one(read_daily(tr=1)),
        }
        barras.update(facs_dict)
        rename_dict.update({k: k for k in facs_dict.keys()})
        cls.barras = barras
        cls.rename_dict = rename_dict
        sort_names = list(rename_dict.values())
        cls.sort_names = sort_names
        cls.barras_together = merge_many(
            list(barras.values()), list(barras.keys()), how="inner"
        )

    def __call__(self):
        """返回纯净因子值"""
        return self.snow_fac

    def set_factors_df_wide(self, df: pd.DataFrame, other_factors: dict = None):
        """传入因子数据，时间为索引，代码为列名"""
        df = df.resample("M").last()
        self.__corr = [
            df.corrwith(i, axis=1).mean() for i in list(self.barras.values())
        ]
        self.__corr = (
            pd.Series(
                self.__corr, index=[self.rename_dict[i] for i in self.barras.keys()]
            )
            .to_frame("相关系数")
            .T
        )
        self.__corr = self.__corr[self.sort_names]
        df = df.stack().reset_index()
        df.columns = ["date", "code", "fac"]
        self.factors = df
        self.corr_pri = pd.merge(df, self.barras_together, on=["date", "code"]).dropna()
        if other_factors is not None:
            other_factors = merge_many(
                list(other_factors.values()), list(other_factors.keys()), how="inner"
            )
            self.corr_pri = pd.merge(self.corr_pri, other_factors, on=["date", "code"])

    @property
    def corr(self) -> pd.DataFrame:
        """因子和10个常用风格因子的相关系数

        Returns
        -------
        pd.DataFrame
            因子和10个常用风格因子的相关系数
        """
        return self.__corr.copy()

    def ols_in_group(self, df):
        """对每个时间段进行回归，并计算残差"""
        xs = list(df.columns)
        xs = [i for i in xs if i != "fac"]
        xs_join = "+".join(xs)
        ols_formula = "fac~" + xs_join
        ols_result = smf.ols(ols_formula, data=df).fit()
        ols_ws = {i: ols_result.params[i] for i in xs}
        ols_b = ols_result.params["Intercept"]
        to_minus = [ols_ws[i] * df[i] for i in xs]
        to_minus = reduce(lambda x, y: x + y, to_minus)
        df = df.assign(snow_fac=df.fac - to_minus - ols_b)
        df = df[["snow_fac"]]
        df = df.rename(columns={"snow_fac": "fac"})
        return df

    def get_snow_fac(self):
        """获得纯净因子"""
        self.snow_fac = (
            self.corr_pri.set_index(["date", "code"])
            .groupby(["date"])
            .apply(self.ols_in_group)
        )
        self.snow_fac = self.snow_fac.unstack()
        self.snow_fac.columns = list(map(lambda x: x[1], list(self.snow_fac.columns)))

corr: pd.DataFrame property

因子和10个常用风格因子的相关系数

Returns

pd.DataFrame 因子和10个常用风格因子的相关系数

__call__()

返回纯净因子值

Source code in pure_ocean_breeze/labor/process.py
Python
3867
3868
3869
def __call__(self):
    """返回纯净因子值"""
    return self.snow_fac

__init__(momentum=1, earningsyield=1, growth=1, liquidity=1, size=1, leverage=1, beta=1, nonlinearsize=1, residualvolatility=1, booktoprice=1) cached classmethod

读入10种常用风格因子，并可以额外加入其他因子

Parameters

facs_dict : Dict, optional 额外加入的因子，名字为key，因子矩阵为value，形如{'反转': ret20, '换手': tr20}, by default None momentum : bool, optional 是否删去动量因子, by default 1 earningsyield : bool, optional 是否删去盈利因子, by default 1 growth : bool, optional 是否删去成长因子, by default 1 liquidity : bool, optional 是否删去流动性因子, by default 1 size : bool, optional 是否删去规模因子, by default 1 leverage : bool, optional 是否删去杠杆因子, by default 1 beta : bool, optional 是否删去贝塔因子, by default 1 nonlinearsize : bool, optional 是否删去非线性市值因子, by default 1 residualvolatility : bool, optional 是否删去残差波动率因子, by default 1 booktoprice : bool, optional 是否删去账面市值比因子, by default 1

Source code in pure_ocean_breeze/labor/process.py
Python
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
@classmethod
@lru_cache(maxsize=None)
def __init__(
    cls,
    momentum: bool = 1,
    earningsyield: bool = 1,
    growth: bool = 1,
    liquidity: bool = 1,
    size: bool = 1,
    leverage: bool = 1,
    beta: bool = 1,
    nonlinearsize: bool = 1,
    residualvolatility: bool = 1,
    booktoprice: bool = 1,
) -> None:
    """读入10种常用风格因子，并可以额外加入其他因子

    Parameters
    ----------
    facs_dict : Dict, optional
        额外加入的因子，名字为key，因子矩阵为value，形如`{'反转': ret20, '换手': tr20}`, by default None
    momentum : bool, optional
        是否删去动量因子, by default 1
    earningsyield : bool, optional
        是否删去盈利因子, by default 1
    growth : bool, optional
        是否删去成长因子, by default 1
    liquidity : bool, optional
        是否删去流动性因子, by default 1
    size : bool, optional
        是否删去规模因子, by default 1
    leverage : bool, optional
        是否删去杠杆因子, by default 1
    beta : bool, optional
        是否删去贝塔因子, by default 1
    nonlinearsize : bool, optional
        是否删去非线性市值因子, by default 1
    residualvolatility : bool, optional
        是否删去残差波动率因子, by default 1
    booktoprice : bool, optional
        是否删去账面市值比因子, by default 1
    """
    cls.homeplace = HomePlace()
    # barra因子数据
    styles = os.listdir(cls.homeplace.barra_data_file)
    styles = [i for i in styles if (i.endswith(".parquet")) and (i[0] != ".")]
    barras = {}
    for s in styles:
        k = s.split(".")[0]
        v = pd.read_parquet(cls.homeplace.barra_data_file + s).resample("M").last()
        barras[k] = v
    rename_dict = {
        "size": "市值",
        "nonlinearsize": "非线性市值",
        "booktoprice": "估值",
        "earningsyield": "盈利",
        "growth": "成长",
        "leverage": "杠杆",
        "liquidity": "流动性",
        "momentum": "动量",
        "residualvolatility": "波动率",
        "beta": "贝塔",
    }
    if momentum == 0:
        barras = {k: v for k, v in barras.items() if k != "momentum"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "momentum"}
    if earningsyield == 0:
        barras = {k: v for k, v in barras.items() if k != "earningsyield"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "earningsyield"}
    if growth == 0:
        barras = {k: v for k, v in barras.items() if k != "growth"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "growth"}
    if liquidity == 0:
        barras = {k: v for k, v in barras.items() if k != "liquidity"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "liquidity"}
    if size == 0:
        barras = {k: v for k, v in barras.items() if k != "size"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "size"}
    if leverage == 0:
        barras = {k: v for k, v in barras.items() if k != "leverage"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "leverage"}
    if beta == 0:
        barras = {k: v for k, v in barras.items() if k != "beta"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "beta"}
    if nonlinearsize == 0:
        barras = {k: v for k, v in barras.items() if k != "nonlinearsize"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "nonlinearsize"}
    if residualvolatility == 0:
        barras = {k: v for k, v in barras.items() if k != "residualvolatility"}
        rename_dict = {
            k: v for k, v in rename_dict.items() if k != "residualvolatility"
        }
    if booktoprice == 0:
        barras = {k: v for k, v in barras.items() if k != "booktoprice"}
        rename_dict = {k: v for k, v in rename_dict.items() if k != "booktoprice"}
    facs_dict = {
        "反转_20天收益率均值": boom_one(read_daily(ret=1)),
        "波动_20天收益率标准差": read_daily(ret=1)
        .rolling(20, min_periods=10)
        .std()
        .resample("M")
        .last(),
        "换手_20天换手率均值": boom_one(read_daily(tr=1)),
    }
    barras.update(facs_dict)
    rename_dict.update({k: k for k in facs_dict.keys()})
    cls.barras = barras
    cls.rename_dict = rename_dict
    sort_names = list(rename_dict.values())
    cls.sort_names = sort_names
    cls.barras_together = merge_many(
        list(barras.values()), list(barras.keys()), how="inner"
    )

get_snow_fac()

获得纯净因子

Source code in pure_ocean_breeze/labor/process.py
Python
3922
3923
3924
3925
3926
3927
3928
3929
3930
def get_snow_fac(self):
    """获得纯净因子"""
    self.snow_fac = (
        self.corr_pri.set_index(["date", "code"])
        .groupby(["date"])
        .apply(self.ols_in_group)
    )
    self.snow_fac = self.snow_fac.unstack()
    self.snow_fac.columns = list(map(lambda x: x[1], list(self.snow_fac.columns)))

ols_in_group(df)

对每个时间段进行回归，并计算残差

Source code in pure_ocean_breeze/labor/process.py
Python
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
def ols_in_group(self, df):
    """对每个时间段进行回归，并计算残差"""
    xs = list(df.columns)
    xs = [i for i in xs if i != "fac"]
    xs_join = "+".join(xs)
    ols_formula = "fac~" + xs_join
    ols_result = smf.ols(ols_formula, data=df).fit()
    ols_ws = {i: ols_result.params[i] for i in xs}
    ols_b = ols_result.params["Intercept"]
    to_minus = [ols_ws[i] * df[i] for i in xs]
    to_minus = reduce(lambda x, y: x + y, to_minus)
    df = df.assign(snow_fac=df.fac - to_minus - ols_b)
    df = df[["snow_fac"]]
    df = df.rename(columns={"snow_fac": "fac"})
    return df

set_factors_df_wide(df, other_factors=None)

传入因子数据，时间为索引，代码为列名

Source code in pure_ocean_breeze/labor/process.py
Python
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
def set_factors_df_wide(self, df: pd.DataFrame, other_factors: dict = None):
    """传入因子数据，时间为索引，代码为列名"""
    df = df.resample("M").last()
    self.__corr = [
        df.corrwith(i, axis=1).mean() for i in list(self.barras.values())
    ]
    self.__corr = (
        pd.Series(
            self.__corr, index=[self.rename_dict[i] for i in self.barras.keys()]
        )
        .to_frame("相关系数")
        .T
    )
    self.__corr = self.__corr[self.sort_names]
    df = df.stack().reset_index()
    df.columns = ["date", "code", "fac"]
    self.factors = df
    self.corr_pri = pd.merge(df, self.barras_together, on=["date", "code"]).dropna()
    if other_factors is not None:
        other_factors = merge_many(
            list(other_factors.values()), list(other_factors.keys()), how="inner"
        )
        self.corr_pri = pd.merge(self.corr_pri, other_factors, on=["date", "code"])

pure_fall

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
class pure_fall(object):
    # DONE：修改为因子文件名可以带“日频_“，也可以不带“日频_“
    def __init__(self, daily_path: str) -> None:
        """一个使用mysql中的分钟数据，来更新因子值的框架

        Parameters
        ----------
        daily_path : str
            日频因子值存储文件的名字，请以'.parquet'结尾
        """
        self.homeplace = HomePlace()
        # 将分钟数据拼成一张日频因子表
        self.daily_factors = None
        self.daily_factors_path = self.homeplace.factor_data_file + "日频_" + daily_path

    def __call__(self, monthly=False):
        """为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表"""
        if monthly:
            return self.monthly_factors.copy()
        else:
            try:
                return self.daily_factors.copy()
            except Exception:
                return self.monthly_factors.copy()

    def wide_to_long(self, df, i):
        """将宽数据转化为长数据，用于因子表转化和拼接"""
        df = df.stack().reset_index()
        df.columns = ["date", "code", i]
        df = df.set_index(["date", "code"])
        return df

    def de_in_group(self, df, help_names):
        """对每个时间，分别做回归，剔除相关因子"""
        ols_order = "fac~" + "+".join(help_names)
        ols_result = smf.ols(ols_order, data=df).fit()
        params = {i: ols_result.params[i] for i in help_names}
        predict = [params[i] * df[i] for i in help_names]
        predict = reduce(lambda x, y: x + y, predict)
        df.fac = df.fac - predict - ols_result.params["Intercept"]
        df = df[["fac"]]
        return df

    def standardlize_in_cross_section(self, df):
        """
        在横截面上做标准化
        输入的df应为，列名是股票代码，索引是时间
        """
        df = df.T
        df = (df - df.mean()) / df.std()
        df = df.T
        return df

__call__(monthly=False)

为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表

Source code in pure_ocean_breeze/labor/process.py
Python
2978
2979
2980
2981
2982
2983
2984
2985
2986
def __call__(self, monthly=False):
    """为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表"""
    if monthly:
        return self.monthly_factors.copy()
    else:
        try:
            return self.daily_factors.copy()
        except Exception:
            return self.monthly_factors.copy()

__init__(daily_path)

一个使用mysql中的分钟数据，来更新因子值的框架

Parameters

daily_path : str 日频因子值存储文件的名字，请以'.parquet'结尾

Source code in pure_ocean_breeze/labor/process.py
Python
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
def __init__(self, daily_path: str) -> None:
    """一个使用mysql中的分钟数据，来更新因子值的框架

    Parameters
    ----------
    daily_path : str
        日频因子值存储文件的名字，请以'.parquet'结尾
    """
    self.homeplace = HomePlace()
    # 将分钟数据拼成一张日频因子表
    self.daily_factors = None
    self.daily_factors_path = self.homeplace.factor_data_file + "日频_" + daily_path

de_in_group(df, help_names)

对每个时间，分别做回归，剔除相关因子

Source code in pure_ocean_breeze/labor/process.py
Python
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
def de_in_group(self, df, help_names):
    """对每个时间，分别做回归，剔除相关因子"""
    ols_order = "fac~" + "+".join(help_names)
    ols_result = smf.ols(ols_order, data=df).fit()
    params = {i: ols_result.params[i] for i in help_names}
    predict = [params[i] * df[i] for i in help_names]
    predict = reduce(lambda x, y: x + y, predict)
    df.fac = df.fac - predict - ols_result.params["Intercept"]
    df = df[["fac"]]
    return df

standardlize_in_cross_section(df)

在横截面上做标准化 输入的df应为，列名是股票代码，索引是时间

Source code in pure_ocean_breeze/labor/process.py
Python
3006
3007
3008
3009
3010
3011
3012
3013
3014
def standardlize_in_cross_section(self, df):
    """
    在横截面上做标准化
    输入的df应为，列名是股票代码，索引是时间
    """
    df = df.T
    df = (df - df.mean()) / df.std()
    df = df.T
    return df

wide_to_long(df, i)

将宽数据转化为长数据，用于因子表转化和拼接

Source code in pure_ocean_breeze/labor/process.py
Python
2988
2989
2990
2991
2992
2993
def wide_to_long(self, df, i):
    """将宽数据转化为长数据，用于因子表转化和拼接"""
    df = df.stack().reset_index()
    df.columns = ["date", "code", i]
    df = df.set_index(["date", "code"])
    return df

pure_fall_frequent

Bases: object

对单只股票单日进行操作

Source code in pure_ocean_breeze/labor/process.py
Python
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
class pure_fall_frequent(object):
    """对单只股票单日进行操作"""

    def __init__(
        self,
        factor_file: str,
        project: str = None,
        startdate: int = None,
        enddate: int = None,
        questdb_host: str = "127.0.0.1",
        kind: str = "stock",
        clickhouse: bool = 0,
        questdb: bool = 0,
        questdb_web_port: str = "9001",
        ignore_history_in_questdb: bool = 0,
        groupby_target: list = ["date", "code"],
    ) -> None:
        """基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

        Parameters
        ----------
        factor_file : str
            用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
        project : str, optional
            该因子所属项目，即子文件夹名称, by default None
        startdate : int, optional
            起始时间，形如20121231，为开区间, by default None
        enddate : int, optional
            截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
        questdb_host: str, optional
            questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
        kind : str, optional
            类型为股票还是指数，指数为'index', by default "stock"
        clickhouse : bool, optional
            使用clickhouse作为数据源，如果postgresql与本参数都为0，将依然从clickhouse中读取, by default 0
        questdb : bool, optional
            使用questdb作为数据源, by default 0
        questdb_web_port : str, optional
            questdb的web_port, by default '9001'
        ignore_history_in_questdb : bool, optional
            打断后重新从头计算，清除在questdb中的记录
        groupby_target: list, optional
            groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算，
            则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']
        """
        homeplace = HomePlace()
        self.kind = kind
        self.groupby_target = groupby_target
        if clickhouse == 0 and questdb == 0:
            clickhouse = 1
        self.clickhouse = clickhouse
        self.questdb = questdb
        self.questdb_web_port = questdb_web_port
        if clickhouse == 1:
            # 连接clickhouse
            self.chc = ClickHouseClient("minute_data")
        elif questdb == 1:
            self.chc = Questdb(host=questdb_host, web_port=questdb_web_port)
        # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
        pinyin = Pinyin()
        self.factor_file_pinyin = pinyin.get_pinyin(
            factor_file.replace(".parquet", ""), ""
        )
        self.factor_steps = Questdb(host=questdb_host, web_port=questdb_web_port)
        if project is not None:
            if not os.path.exists(homeplace.factor_data_file + project):
                os.makedirs(homeplace.factor_data_file + project)
            else:
                logger.info(f"当前正在{project}项目中……")
        else:
            logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
        # 完整的因子文件路径
        if project is not None:
            factor_file = homeplace.factor_data_file + project + "/" + factor_file
        else:
            factor_file = homeplace.factor_data_file + factor_file
        self.factor_file = factor_file
        # 读入之前的因子
        if os.path.exists(factor_file):
            factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
            )
            factor_old = self.factor_steps.get_data_with_tuple(
                f"select * from '{self.factor_file_pinyin}'"
            ).drop_duplicates(subset=["date", "code"])
            factor_old = factor_old.pivot(index="date", columns="code", values="fac")
            factor_old = factor_old.sort_index()
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif ignore_history_in_questdb and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
            )
            factor_old = self.factor_steps.do_order(
                f"drop table '{self.factor_file_pinyin}'"
            )
            self.factor_old = None
            self.dates_old = []
            logger.info("删除完毕，正在重新计算")
        else:
            self.factor_old = None
            self.dates_old = []
            logger.info("这个因子以前没有，正在重新计算")
        # 读取当前所有的日子
        dates_all = self.chc.show_all_dates(f"minute_data_{kind}")
        dates_all = [int(i) for i in dates_all]
        if startdate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i >= startdate]
        if enddate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i <= enddate]
        self.dates_all = dates_all
        # 需要新补充的日子
        self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
        if len(self.dates_new) == 0:
            ...
        elif len(self.dates_new) == 1:
            self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
            print(f"只缺一天{self.dates_new[0]}")
        else:
            dates = [pd.Timestamp(str(i)) for i in self.dates_new]
            intervals = [[]] * len(dates)
            interbee = 0
            intervals[0] = intervals[0] + [dates[0]]
            for i in range(len(dates) - 1):
                val1 = dates[i]
                val2 = dates[i + 1]
                if val2 - val1 < pd.Timedelta(days=30):
                    ...
                else:
                    interbee = interbee + 1
                intervals[interbee] = intervals[interbee] + [val2]
            intervals = [i for i in intervals if len(i) > 0]
            print(f"共{len(intervals)}个时间区间，分别是")
            for date in intervals:
                print(f"从{date[0]}到{date[-1]}")
            self.dates_new_intervals = intervals
        self.factor_new = []

    def __call__(self) -> pd.DataFrame:
        """获得经运算产生的因子

        Returns
        -------
        `pd.DataFrame`
            经运算产生的因子值
        """
        return self.factor.copy()

    def forward_dates(self, dates, many_days):
        dates_index = [self.dates_all.index(i) for i in dates]

        def value(x, a):
            if x >= 0:
                return a[x]
            else:
                return None

        return [value(i - many_days, self.dates_all) for i in dates_index]

    def select_one_calculate(
        self,
        date: pd.Timestamp,
        func: Callable,
        fields: str = "*",
        show_time: bool = 0,
    ) -> None:
        the_func = partial(func)
        if not isinstance(date, int):
            date = int(datetime.datetime.strftime(date, "%Y%m%d"))
        # 开始计算因子值
        if self.clickhouse == 1:
            sql_order = f"select {fields} from minute_data.minute_data_{self.kind} where date={date * 100} order by code,date,num"
        else:
            sql_order = (
                f"select {fields} from minute_data_{self.kind} where date='{date}'"
            )
        if show_time:
            df = self.chc.get_data_show_time(sql_order)
        else:
            df = self.chc.get_data(sql_order)
        if self.clickhouse == 1:
            df = ((df.set_index("code")) / 100).reset_index()
        else:
            df.num = df.num.astype(int)
            df.date = df.date.astype(int)
            df = df.sort_values(["date", "num"])
        df = df.groupby(self.groupby_target).apply(the_func)
        if self.groupby_target == ["date", "code"]:
            df = df.to_frame("fac").reset_index()
            df.columns = ["date", "code", "fac"]
        else:
            df = df.reset_index()
        if (df is not None) and (df.shape[0] > 0):
            df = df.pivot(columns="code", index="date", values="fac")
            df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
            to_save = df.stack().reset_index()
            to_save.columns = ["date", "code", "fac"]
            self.factor_steps.write_via_df(
                to_save, self.factor_file_pinyin, tuple_col="fac"
            )
            return df

    def select_many_calculate(
        self,
        dates: List[pd.Timestamp],
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
        use_mpire: bool = 0,
    ) -> None:
        the_func = partial(func)
        factor_new = []
        dates = [int(datetime.datetime.strftime(i, "%Y%m%d")) for i in dates]
        if many_days == 1:
            # 将需要更新的日子分块，每200天一组，一起运算
            dates_new_len = len(dates)
            cut_points = list(range(0, dates_new_len, chunksize)) + [dates_new_len - 1]
            if cut_points[-1] == cut_points[-2]:
                cut_points = cut_points[:-1]
            cuts = tuple(zip(cut_points[:-many_days], cut_points[many_days:]))
            df_first = self.select_one_calculate(
                date=dates[0],
                func=func,
                fields=fields,
                show_time=show_time,
            )
            factor_new.append(df_first)

            def cal_one(date1, date2):
                if self.clickhouse == 1:
                    sql_order = f"select {fields} from minute_data.minute_data_{self.kind} where date>{dates[date1] * 100} and date<={dates[date2] * 100} order by code,date,num"
                else:
                    sql_order = f"select {fields} from minute_data_{self.kind} where cast(date as int)>{dates[date1]} and cast(date as int)<={dates[date2]} order by code,date,num"
                if show_time:
                    df = self.chc.get_data_show_time(sql_order)
                else:
                    df = self.chc.get_data(sql_order)
                if self.clickhouse == 1:
                    df = ((df.set_index("code")) / 100).reset_index()
                else:
                    df.num = df.num.astype(int)
                    df.date = df.date.astype(int)
                    df = df.sort_values(["date", "num"])
                df = df.groupby(self.groupby_target).apply(the_func)
                if self.groupby_target == ["date", "code"]:
                    df = df.to_frame("fac").reset_index()
                    df.columns = ["date", "code", "fac"]
                else:
                    df = df.reset_index()
                df = df.pivot(columns="code", index="date", values="fac")
                df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
                to_save = df.stack().reset_index()
                to_save.columns = ["date", "code", "fac"]
                self.factor_steps.write_via_df(
                    to_save, self.factor_file_pinyin, tuple_col="fac"
                )
                return df

            if n_jobs > 1:
                if use_mpire:
                    with WorkerPool(n_jobs=n_jobs) as pool:
                        factor_new_more = pool.map(
                            cal_one,
                            cut_points[:-many_days],
                            cut_points[many_days:],
                            progress_bar=True,
                        )
                else:
                    with concurrent.futures.ThreadPoolExecutor(
                        max_workers=n_jobs
                    ) as executor:
                        factor_new_more = list(
                            tqdm.auto.tqdm(
                                executor.map(
                                    cal_one,
                                    cut_points[:-many_days],
                                    cut_points[many_days:],
                                ),
                                total=len(cut_points[many_days:]),
                            )
                        )
                factor_new = factor_new + factor_new_more
            else:
                # 开始计算因子值
                for date1, date2 in tqdm.auto.tqdm(cuts, desc="不知乘月几人归，落月摇情满江树。"):
                    df = cal_one(date1, date2)
                    factor_new.append(df)
        else:

            def cal_two(date1, date2):
                if date1 is not None:
                    if self.clickhouse == 1:
                        sql_order = f"select {fields} from minute_data.minute_data_{self.kind} where date>{date1*100} and date<={date2*100} order by code,date,num"
                    else:
                        sql_order = f"select {fields} from minute_data_{self.kind} where cast(date as int)>{date1} and cast(date as int)<={date2} order by code,date,num"
                    if show_time:
                        df = self.chc.get_data_show_time(sql_order)
                    else:
                        df = self.chc.get_data(sql_order)
                    if self.clickhouse == 1:
                        df = ((df.set_index("code")) / 100).reset_index()
                    else:
                        df.num = df.num.astype(int)
                        df.date = df.date.astype(int)
                        df = df.sort_values(["date", "num"])
                    if self.groupby_target == [
                        "date",
                        "code",
                    ] or self.groupby_target == ["code"]:
                        df = df.groupby(["code"]).apply(the_func).reset_index()
                    else:
                        df = the_func(df)
                    df = df.assign(date=date2)
                    df.columns = ["code", "fac", "date"]
                    df = df.pivot(columns="code", index="date", values="fac")
                    df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
                    to_save = df.stack().reset_index()
                    to_save.columns = ["date", "code", "fac"]
                    self.factor_steps.write_via_df(
                        to_save, self.factor_file_pinyin, tuple_col="fac"
                    )
                    return df

            pairs = self.forward_dates(dates, many_days=many_days)
            cuts2 = tuple(zip(pairs, dates))
            if n_jobs > 1:
                if use_mpire:
                    with WorkerPool(n_jobs=n_jobs) as pool:
                        factor_new_more = pool.map(
                            cal_two, pairs, dates, progress_bar=True
                        )
                else:
                    with concurrent.futures.ThreadPoolExecutor(
                        max_workers=n_jobs
                    ) as executor:
                        factor_new_more = list(
                            tqdm.auto.tqdm(
                                executor.map(cal_two, pairs, dates),
                                total=len(pairs),
                            )
                        )
                factor_new = factor_new + factor_new_more
            else:
                # 开始计算因子值
                for date1, date2 in tqdm.auto.tqdm(cuts2, desc="知不可乎骤得，托遗响于悲风。"):
                    df = cal_two(date1, date2)
                    factor_new.append(df)

        if len(factor_new) > 0:
            factor_new = pd.concat(factor_new)
            return factor_new
        else:
            return None

    def select_any_calculate(
        self,
        dates: List[pd.Timestamp],
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
        use_mpire: bool = 0,
    ) -> None:
        if len(dates) == 1 and many_days == 1:
            res = self.select_one_calculate(
                dates[0],
                func=func,
                fields=fields,
                show_time=show_time,
            )
        else:
            res = self.select_many_calculate(
                dates=dates,
                func=func,
                fields=fields,
                chunksize=chunksize,
                show_time=show_time,
                many_days=many_days,
                n_jobs=n_jobs,
                use_mpire=use_mpire,
            )
        if res is not None:
            self.factor_new.append(res)
        return res

    @staticmethod
    def for_cross_via_str(func):
        """返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值
        例如
        ```python
        return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
        ```
        上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列"""

        def full_run(df, *args, **kwargs):
            codes = sorted(list(set(df.code)))
            res = func(df, *args, **kwargs)
            if isinstance(res[0], list):
                kind = 1
                res = [",".join(i) for i in res]
            else:
                kind = 0
            df = pd.DataFrame({"code": codes, "fac": res})
            if kind:
                df.fac = df.fac.apply(lambda x: [float(i) for i in x.split(",")])
            return df

        return full_run

    @staticmethod
    def for_cross_via_zip(func):
        """返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值
        例如
        ```python
        return (
                    pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                    pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                )
        ```
        上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值
        """

        def full_run(df, *args, **kwargs):
            res = func(df, *args, **kwargs)
            if isinstance(res, pd.Series):
                res = res.reset_index()
                res.columns = ["code", "fac"]
                return res
            elif isinstance(res, pd.DataFrame):
                res.columns = [f"fac{i}" for i in range(len(res.columns))]
                res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
                res = res[["fac"]].reset_index()
                res.columns = ["code", "fac"]
                return res
            elif res is None:
                ...
            else:
                res = pd.concat(res, axis=1)
                res.columns = [f"fac{i}" for i in range(len(res.columns))]
                res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
                res = res[["fac"]].reset_index()
                res.columns = ["code", "fac"]
                return res

        return full_run

    def get_daily_factors_one(
        self,
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
        use_mpire: bool = 0,
    ):
        if len(self.dates_new) > 0:
            for interval in self.dates_new_intervals:
                df = self.select_any_calculate(
                    dates=interval,
                    func=func,
                    fields=fields,
                    chunksize=chunksize,
                    show_time=show_time,
                    many_days=many_days,
                    n_jobs=n_jobs,
                    use_mpire=use_mpire,
                )
            if len(self.factor_new) > 0:
                self.factor_new = pd.concat(self.factor_new)
                # 拼接新的和旧的
                self.factor = pd.concat([self.factor_old, self.factor_new]).sort_index()
                self.factor = drop_duplicates_index(self.factor.dropna(how="all"))
                new_end_date = datetime.datetime.strftime(
                    self.factor.index.max(), "%Y%m%d"
                )
                # 存入本地
                self.factor.to_parquet(self.factor_file)
                logger.info(f"截止到{new_end_date}的因子值计算完了")
                # 删除存储在questdb的中途备份数据
                try:
                    self.factor_steps.do_order(
                        f"drop table '{self.factor_file_pinyin}'"
                    )
                    logger.info("备份在questdb的表格已删除")
                except Exception:
                    logger.warning("删除questdb中表格时，存在某个未知错误，请当心")
            else:
                logger.warning("由于某种原因，更新的因子值计算失败，建议检查🤒")
                # 拼接新的和旧的
                self.factor = pd.concat([self.factor_old]).sort_index()
                self.factor = drop_duplicates_index(self.factor.dropna(how="all"))

        else:
            self.factor = drop_duplicates_index(self.factor_old)
            # 存入本地
            self.factor.to_parquet(self.factor_file)
            new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
            logger.info(f"当前截止到{new_end_date}的因子值已经是最新的了")

    @kk.desktop_sender(title="嘿，分钟数据处理完啦～🎈")
    def get_daily_factors_two(
        self,
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
    ):
        self.get_daily_factors_one(
            func=func,
            fields=fields,
            chunksize=chunksize,
            show_time=show_time,
            many_days=many_days,
            n_jobs=n_jobs,
        )

    def get_daily_factors(
        self,
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
    ) -> None:
        """每次抽取chunksize天的截面上全部股票的分钟数据
        对每天的股票的数据计算因子值

        Parameters
        ----------
        func : Callable
            用于计算因子值的函数
        fields : str, optional
            股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open'
            提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*"
        chunksize : int, optional
            每次读取的截面上的天数, by default 10
        show_time : bool, optional
            展示每次读取数据所需要的时间, by default 0
        many_days : int, optional
            计算某天的因子值时，需要使用之前多少天的数据
        n_jobs : int, optional
            并行数量, by default 1
        """
        try:
            self.get_daily_factors_two(
                func=func,
                fields=fields,
                chunksize=chunksize,
                show_time=show_time,
                many_days=many_days,
                n_jobs=n_jobs,
            )
        except Exception:
            self.get_daily_factors_one(
                func=func,
                fields=fields,
                chunksize=chunksize,
                show_time=show_time,
                many_days=many_days,
                n_jobs=n_jobs,
            )

    def drop_table(self):
        """直接删除存储在questdb中的暂存数据"""
        try:
            self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
            logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
        except Exception:
            logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

__call__()

获得经运算产生的因子

Returns

pd.DataFrame 经运算产生的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
3308
3309
3310
3311
3312
3313
3314
3315
3316
def __call__(self) -> pd.DataFrame:
    """获得经运算产生的因子

    Returns
    -------
    `pd.DataFrame`
        经运算产生的因子值
    """
    return self.factor.copy()

__init__(factor_file, project=None, startdate=None, enddate=None, questdb_host='127.0.0.1', kind='stock', clickhouse=0, questdb=0, questdb_web_port='9001', ignore_history_in_questdb=0, groupby_target=['date', 'code'])

基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

Parameters

factor_file : str 用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾 project : str, optional 该因子所属项目，即子文件夹名称, by default None startdate : int, optional 起始时间，形如20121231，为开区间, by default None enddate : int, optional 截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None questdb_host: str, optional questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1' kind : str, optional 类型为股票还是指数，指数为'index', by default "stock" clickhouse : bool, optional 使用clickhouse作为数据源，如果postgresql与本参数都为0，将依然从clickhouse中读取, by default 0 questdb : bool, optional 使用questdb作为数据源, by default 0 questdb_web_port : str, optional questdb的web_port, by default '9001' ignore_history_in_questdb : bool, optional 打断后重新从头计算，清除在questdb中的记录 groupby_target: list, optional groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算， 则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']

Source code in pure_ocean_breeze/labor/process.py
Python
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
def __init__(
    self,
    factor_file: str,
    project: str = None,
    startdate: int = None,
    enddate: int = None,
    questdb_host: str = "127.0.0.1",
    kind: str = "stock",
    clickhouse: bool = 0,
    questdb: bool = 0,
    questdb_web_port: str = "9001",
    ignore_history_in_questdb: bool = 0,
    groupby_target: list = ["date", "code"],
) -> None:
    """基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

    Parameters
    ----------
    factor_file : str
        用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
    project : str, optional
        该因子所属项目，即子文件夹名称, by default None
    startdate : int, optional
        起始时间，形如20121231，为开区间, by default None
    enddate : int, optional
        截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
    questdb_host: str, optional
        questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
    kind : str, optional
        类型为股票还是指数，指数为'index', by default "stock"
    clickhouse : bool, optional
        使用clickhouse作为数据源，如果postgresql与本参数都为0，将依然从clickhouse中读取, by default 0
    questdb : bool, optional
        使用questdb作为数据源, by default 0
    questdb_web_port : str, optional
        questdb的web_port, by default '9001'
    ignore_history_in_questdb : bool, optional
        打断后重新从头计算，清除在questdb中的记录
    groupby_target: list, optional
        groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算，
        则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']
    """
    homeplace = HomePlace()
    self.kind = kind
    self.groupby_target = groupby_target
    if clickhouse == 0 and questdb == 0:
        clickhouse = 1
    self.clickhouse = clickhouse
    self.questdb = questdb
    self.questdb_web_port = questdb_web_port
    if clickhouse == 1:
        # 连接clickhouse
        self.chc = ClickHouseClient("minute_data")
    elif questdb == 1:
        self.chc = Questdb(host=questdb_host, web_port=questdb_web_port)
    # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
    pinyin = Pinyin()
    self.factor_file_pinyin = pinyin.get_pinyin(
        factor_file.replace(".parquet", ""), ""
    )
    self.factor_steps = Questdb(host=questdb_host, web_port=questdb_web_port)
    if project is not None:
        if not os.path.exists(homeplace.factor_data_file + project):
            os.makedirs(homeplace.factor_data_file + project)
        else:
            logger.info(f"当前正在{project}项目中……")
    else:
        logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
    # 完整的因子文件路径
    if project is not None:
        factor_file = homeplace.factor_data_file + project + "/" + factor_file
    else:
        factor_file = homeplace.factor_data_file + factor_file
    self.factor_file = factor_file
    # 读入之前的因子
    if os.path.exists(factor_file):
        factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
        )
        factor_old = self.factor_steps.get_data_with_tuple(
            f"select * from '{self.factor_file_pinyin}'"
        ).drop_duplicates(subset=["date", "code"])
        factor_old = factor_old.pivot(index="date", columns="code", values="fac")
        factor_old = factor_old.sort_index()
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif ignore_history_in_questdb and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
        )
        factor_old = self.factor_steps.do_order(
            f"drop table '{self.factor_file_pinyin}'"
        )
        self.factor_old = None
        self.dates_old = []
        logger.info("删除完毕，正在重新计算")
    else:
        self.factor_old = None
        self.dates_old = []
        logger.info("这个因子以前没有，正在重新计算")
    # 读取当前所有的日子
    dates_all = self.chc.show_all_dates(f"minute_data_{kind}")
    dates_all = [int(i) for i in dates_all]
    if startdate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i >= startdate]
    if enddate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i <= enddate]
    self.dates_all = dates_all
    # 需要新补充的日子
    self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
    if len(self.dates_new) == 0:
        ...
    elif len(self.dates_new) == 1:
        self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
        print(f"只缺一天{self.dates_new[0]}")
    else:
        dates = [pd.Timestamp(str(i)) for i in self.dates_new]
        intervals = [[]] * len(dates)
        interbee = 0
        intervals[0] = intervals[0] + [dates[0]]
        for i in range(len(dates) - 1):
            val1 = dates[i]
            val2 = dates[i + 1]
            if val2 - val1 < pd.Timedelta(days=30):
                ...
            else:
                interbee = interbee + 1
            intervals[interbee] = intervals[interbee] + [val2]
        intervals = [i for i in intervals if len(i) > 0]
        print(f"共{len(intervals)}个时间区间，分别是")
        for date in intervals:
            print(f"从{date[0]}到{date[-1]}")
        self.dates_new_intervals = intervals
    self.factor_new = []

drop_table()

直接删除存储在questdb中的暂存数据

Source code in pure_ocean_breeze/labor/process.py
Python
3741
3742
3743
3744
3745
3746
3747
def drop_table(self):
    """直接删除存储在questdb中的暂存数据"""
    try:
        self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
        logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
    except Exception:
        logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

for_cross_via_str(func) staticmethod

返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值 例如

Python
1
return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列

Source code in pure_ocean_breeze/labor/process.py
Python
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
@staticmethod
def for_cross_via_str(func):
    """返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值
    例如
    ```python
    return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
    ```
    上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列"""

    def full_run(df, *args, **kwargs):
        codes = sorted(list(set(df.code)))
        res = func(df, *args, **kwargs)
        if isinstance(res[0], list):
            kind = 1
            res = [",".join(i) for i in res]
        else:
            kind = 0
        df = pd.DataFrame({"code": codes, "fac": res})
        if kind:
            df.fac = df.fac.apply(lambda x: [float(i) for i in x.split(",")])
        return df

    return full_run

for_cross_via_zip(func) staticmethod

返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值 例如

Python
1
2
3
4
return (
            pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
            pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
        )
上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值

Source code in pure_ocean_breeze/labor/process.py
Python
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
@staticmethod
def for_cross_via_zip(func):
    """返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值
    例如
    ```python
    return (
                pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
            )
    ```
    上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值
    """

    def full_run(df, *args, **kwargs):
        res = func(df, *args, **kwargs)
        if isinstance(res, pd.Series):
            res = res.reset_index()
            res.columns = ["code", "fac"]
            return res
        elif isinstance(res, pd.DataFrame):
            res.columns = [f"fac{i}" for i in range(len(res.columns))]
            res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
            res = res[["fac"]].reset_index()
            res.columns = ["code", "fac"]
            return res
        elif res is None:
            ...
        else:
            res = pd.concat(res, axis=1)
            res.columns = [f"fac{i}" for i in range(len(res.columns))]
            res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
            res = res[["fac"]].reset_index()
            res.columns = ["code", "fac"]
            return res

    return full_run

get_daily_factors(func, fields='*', chunksize=10, show_time=0, many_days=1, n_jobs=1)

每次抽取chunksize天的截面上全部股票的分钟数据 对每天的股票的数据计算因子值

Parameters

func : Callable 用于计算因子值的函数 fields : str, optional 股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open' 提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*" chunksize : int, optional 每次读取的截面上的天数, by default 10 show_time : bool, optional 展示每次读取数据所需要的时间, by default 0 many_days : int, optional 计算某天的因子值时，需要使用之前多少天的数据 n_jobs : int, optional 并行数量, by default 1

Source code in pure_ocean_breeze/labor/process.py
Python
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
def get_daily_factors(
    self,
    func: Callable,
    fields: str = "*",
    chunksize: int = 10,
    show_time: bool = 0,
    many_days: int = 1,
    n_jobs: int = 1,
) -> None:
    """每次抽取chunksize天的截面上全部股票的分钟数据
    对每天的股票的数据计算因子值

    Parameters
    ----------
    func : Callable
        用于计算因子值的函数
    fields : str, optional
        股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open'
        提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*"
    chunksize : int, optional
        每次读取的截面上的天数, by default 10
    show_time : bool, optional
        展示每次读取数据所需要的时间, by default 0
    many_days : int, optional
        计算某天的因子值时，需要使用之前多少天的数据
    n_jobs : int, optional
        并行数量, by default 1
    """
    try:
        self.get_daily_factors_two(
            func=func,
            fields=fields,
            chunksize=chunksize,
            show_time=show_time,
            many_days=many_days,
            n_jobs=n_jobs,
        )
    except Exception:
        self.get_daily_factors_one(
            func=func,
            fields=fields,
            chunksize=chunksize,
            show_time=show_time,
            many_days=many_days,
            n_jobs=n_jobs,
        )

pure_fall_nature

Source code in pure_ocean_breeze/labor/process.py
Python
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
6948
6949
class pure_fall_nature:
    def __init__(
        self,
        factor_file: str,
        project: str = None,
        startdate: int = None,
        enddate: int = None,
        questdb_host: str = "127.0.0.1",
        ignore_history_in_questdb: bool = 0,
        groupby_code: bool = 1,
    ) -> None:
        """基于股票逐笔数据，计算因子值，每天的因子值只用到当日的数据

        Parameters
        ----------
        factor_file : str
            用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
        project : str, optional
            该因子所属项目，即子文件夹名称, by default None
        startdate : int, optional
            起始时间，形如20121231，为开区间, by default None
        enddate : int, optional
            截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
        questdb_host: str, optional
            questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
        ignore_history_in_questdb : bool, optional
            打断后重新从头计算，清除在questdb中的记录
        groupby_target: list, optional
            groupby计算时，分组的依据, by default ['code']
        """
        homeplace = HomePlace()
        self.groupby_code = groupby_code
        # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
        pinyin = Pinyin()
        self.factor_file_pinyin = pinyin.get_pinyin(
            factor_file.replace(".parquet", ""), ""
        )
        self.factor_steps = Questdb(host=questdb_host)
        if project is not None:
            if not os.path.exists(homeplace.factor_data_file + project):
                os.makedirs(homeplace.factor_data_file + project)
            else:
                logger.info(f"当前正在{project}项目中……")
        else:
            logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
        # 完整的因子文件路径
        if project is not None:
            factor_file = homeplace.factor_data_file + project + "/" + factor_file
        else:
            factor_file = homeplace.factor_data_file + factor_file
        self.factor_file = factor_file
        # 读入之前的因子
        if os.path.exists(factor_file):
            factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
            )
            factor_old = self.factor_steps.get_data_with_tuple(
                f"select * from '{self.factor_file_pinyin}'"
            ).drop_duplicates(subset=["date", "code"])
            factor_old = factor_old.pivot(index="date", columns="code", values="fac")
            factor_old = factor_old.sort_index()
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif ignore_history_in_questdb and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
            )
            factor_old = self.factor_steps.do_order(
                f"drop table '{self.factor_file_pinyin}'"
            )
            self.factor_old = None
            self.dates_old = []
            logger.info("删除完毕，正在重新计算")
        else:
            self.factor_old = None
            self.dates_old = []
            logger.info("这个因子以前没有，正在重新计算")
        # 读取当前所有的日子
        dates_all = os.listdir(homeplace.tick_by_tick_data)
        dates_all = [i.split(".")[0] for i in dates_all if i.endswith(".parquet")]
        dates_all = [i.replace("-", "") for i in dates_all]
        dates_all = [int(i) for i in dates_all if "20" if i]
        if startdate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i >= startdate]
        if enddate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i <= enddate]
        self.dates_all = dates_all
        # 需要新补充的日子
        self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
        if len(self.dates_new) == 0:
            ...
        elif len(self.dates_new) == 1:
            self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
            print(f"只缺一天{self.dates_new[0]}")
        else:
            dates = [pd.Timestamp(str(i)) for i in self.dates_new]
            intervals = [[]] * len(dates)
            interbee = 0
            intervals[0] = intervals[0] + [dates[0]]
            for i in range(len(dates) - 1):
                val1 = dates[i]
                val2 = dates[i + 1]
                if val2 - val1 < pd.Timedelta(days=30):
                    ...
                else:
                    interbee = interbee + 1
                intervals[interbee] = intervals[interbee] + [val2]
            intervals = [i for i in intervals if len(i) > 0]
            print(f"共{len(intervals)}个时间区间，分别是")
            for date in intervals:
                print(f"从{date[0]}到{date[-1]}")
            self.dates_new_intervals = intervals
        self.factor_new = []
        self.age = read_daily(age=1)
        self.state = read_daily(state=1)
        self.closes_unadj = read_daily(close=1, unadjust=1).shift(1)

    def __call__(self) -> pd.DataFrame:
        """获得经运算产生的因子

        Returns
        -------
        `pd.DataFrame`
            经运算产生的因子值
        """
        return self.factor.copy()

    def select_one_calculate(
        self,
        date: pd.Timestamp,
        func: Callable,
        fields: str = "*",
        resample_frequency: str = None,
        opens_in: bool = 0,
        highs_in: bool = 0,
        lows_in: bool = 0,
        amounts_in: bool = 0,
        merge_them: bool = 0,
    ) -> None:
        the_func = partial(func)
        if not isinstance(date, int):
            date = int(datetime.datetime.strftime(date, "%Y%m%d"))
        parquet_name = (
            homeplace.tick_by_tick_data
            + str(date)[:4]
            + "-"
            + str(date)[4:6]
            + "-"
            + str(date)[6:]
            + ".parquet"
        )
        if resample_frequency is not None:
            fields = "date,code,price,amount"
        # 开始计算因子值
        cursor = duckdb.connect()
        df = (
            cursor.execute(f"select {fields} from '{parquet_name}';")
            .arrow()
            .to_pandas()
        )
        date = df.date.iloc[0]
        date0 = pd.Timestamp(year=date.year, month=date.month, day=date.day)
        age_here = self.age.loc[pd.Timestamp(pd.Timestamp(df.date.iloc[0]).date())]
        age_here = age_here.where(age_here > 180, np.nan).dropna()
        state_here = self.state.loc[pd.Timestamp(pd.Timestamp(df.date.iloc[0]).date())]
        state_here = state_here.where(state_here > 0, np.nan).dropna()
        df = df[df.code.isin(age_here.index)]
        df = df[df.code.isin(state_here.index)]

        if resample_frequency is not None:
            date = df.date.iloc[0]
            date0 = pd.Timestamp(year=date.year, month=date.month, day=date.day)
            head = self.closes_unadj.loc[date0].to_frame("head_temp").T
            df = df[df.code.isin(head.columns)]
            price = df.drop_duplicates(subset=["code", "date"], keep="last").pivot(
                index="date", columns="code", values="price"
            )
            closes = price.resample(resample_frequency).last()
            head = head[[i for i in head.columns if i in closes.columns]]
            price = pd.concat([head, closes])
            closes = closes.ffill().iloc[1:, :]
            self.closes = closes
            names = []

            if opens_in:
                price = df.drop_duplicates(subset=["code", "date"], keep="first").pivot(
                    index="date", columns="code", values="price"
                )
                opens = price.resample(resample_frequency).first()
                opens = np.isnan(opens).replace(True, 1).replace(
                    False, 0
                ) * closes.shift(1) + opens.fillna(0)
                self.opens = opens
                names.append("open")
            else:
                self.opens = None

            if highs_in:
                price = (
                    df.sort_values(["code", "date", "price"])
                    .drop_duplicates(subset=["code", "date"], keep="last")
                    .pivot(index="date", columns="code", values="price")
                )
                highs = price.resample(resample_frequency).max()
                highs = np.isnan(highs).replace(True, 1).replace(
                    False, 0
                ) * closes.shift(1) + highs.fillna(0)
                self.highs = highs
                names.append("high")
            else:
                self.highs = None

            if lows_in:
                price = (
                    df.sort_values(["code", "date", "price"])
                    .drop_duplicates(subset=["code", "date"], keep="first")
                    .pivot(index="date", columns="code", values="price")
                )
                lows = price.resample(resample_frequency).min()
                lows = np.isnan(lows).replace(True, 1).replace(False, 0) * closes.shift(
                    1
                ) + lows.fillna(0)
                self.lows = lows
                names.append("low")
            else:
                self.low = None

            names.append("close")
            if amounts_in:
                amounts = df.groupby(["code", "date"]).amount.sum().reset_index()
                amounts = amounts.pivot(index="date", columns="code", values="amount")
                amounts = amounts.resample(resample_frequency).sum().fillna(0)
                self.amounts = amounts
                names.append("amount")
            else:
                self.amounts = None

            if merge_them:
                self.data = merge_many(
                    [
                        i
                        for i in [
                            self.opens,
                            self.highs,
                            self.lows,
                            self.closes,
                            self.amounts,
                        ]
                        if i is not None
                    ],
                    names,
                )

        if self.groupby_code:
            df = df.groupby(["code"]).apply(the_func)
        else:
            df = the_func(df)
            if isinstance(df, pd.DataFrame):
                df.columns = [f"fac{i}" for i in range(len(df.columns))]
                df = df.assign(fac=list(zip(*[df[i] for i in list(df.columns)])))
                df = df[["fac"]]
            elif isinstance(df, list) or isinstance(df, tuple):
                df = pd.concat(list(df), axis=1)
                df.columns = [f"fac{i}" for i in range(len(df.columns))]
                df = df.assign(fac=list(zip(*[df[i] for i in list(df.columns)])))
                df = df[["fac"]]
        df = df.reset_index()
        df.columns = ["code", "fac"]
        df.insert(
            0, "date", pd.Timestamp(year=date.year, month=date.month, day=date.day)
        )
        if (df is not None) and (df.shape[0] > 0):
            df1 = df.pivot(columns="code", index="date", values="fac")
            self.factor_steps.write_via_df(df, self.factor_file_pinyin, tuple_col="fac")
            return df1

    def get_daily_factors(
        self,
        func: Callable,
        n_jobs: int = 1,
        fields: str = "*",
        resample_frequency: str = None,
        opens_in: bool = 0,
        highs_in: bool = 0,
        lows_in: bool = 0,
        amounts_in: bool = 0,
        merge_them: bool = 0,
        use_mpire: bool = 0,
    ) -> None:
        """每次抽取chunksize天的截面上全部股票的分钟数据
        对每天的股票的数据计算因子值

        Parameters
        ----------
        func : Callable
            用于计算因子值的函数
        n_jobs : int, optional
            并行数量, by default 1
        fields : str, optional
            要读取的字段，可选包含`date,code,price,amount,saleamount,buyamount,action,saleid,saleprice,buyid,buyprice`，其中date,code必须包含, by default `'*'`
        resample_frequency : str, optional
            将逐笔数据转化为秒级或分钟频数据，可以填写要转化的频率，如'3s'（3秒数据），'1m'（1分钟数据），
            指定此参数后，将自动生成一个self.closes的收盘价矩阵(index为时间,columns为股票代码,values为收盘价)，
            可在循环计算的函数中使用`self.closes`来调用计算好的值, by default None
        opens_in : bool, optional
            在resample_frequency不为None的情况下，可以使用此参数，提前计算好开盘价矩阵(index为时间,columns为股票代码,values为开盘价)，
            可在循环计算的函数中使用`self.opens`来调用计算好的值，by default 0
        highs_in : bool, optional
            在resample_frequency不为None的情况下，可以使用此参数，提前计算好最高价矩阵(index为时间,columns为股票代码,values为最高价)，
            可在循环计算的函数中使用`self.highs`来调用计算好的值，by default 0
        lows_in : bool, optional
            在resample_frequency不为None的情况下，可以使用此参数，提前计算好最低价矩阵(index为时间,columns为股票代码,values为最低价)，
            可在循环计算的函数中使用`self.lows`来调用计算好的值，by default 0
        amounts_in : bool, optional
            在resample_frequency不为None的情况下，可以使用此参数，提前计算好成交额矩阵(index为时间,columns为股票代码,values为成交量)，
            可在循环计算的函数中使用`self.amounts`来调用计算好的值，by default 0
        merge_them : bool, optional
            在resample_frequency不为None的情况下，可以使用此参数，将计算好的因子值合并到一起，生成类似于分钟数据的sql形式，by default 0
        use_mpire : bool, optional
            并行是否使用mpire，默认使用concurrent，by default 0
        """
        if len(self.dates_new) > 0:
            if n_jobs > 1:
                if use_mpire:
                    with WorkerPool(n_jobs=n_jobs) as pool:
                        self.factor_new = pool.map(
                            lambda x: self.select_one_calculate(
                                date=x,
                                func=func,
                                fields=fields,
                                resample_frequency=resample_frequency,
                                opens_in=opens_in,
                                highs_in=highs_in,
                                lows_in=lows_in,
                                amounts_in=amounts_in,
                                merge_them=merge_them,
                            ),
                            self.dates_new,
                            progress_bar=True,
                        )
                else:
                    with concurrent.futures.ThreadPoolExecutor(
                        max_workers=n_jobs
                    ) as executor:
                        self.factor_new = list(
                            tqdm.auto.tqdm(
                                executor.map(
                                    lambda x: self.select_one_calculate(
                                        date=x,
                                        func=func,
                                        fields=fields,
                                        resample_frequency=resample_frequency,
                                        opens_in=opens_in,
                                        highs_in=highs_in,
                                        lows_in=lows_in,
                                        amounts_in=amounts_in,
                                        merge_them=merge_them,
                                    ),
                                    self.dates_new,
                                ),
                                total=len(self.dates_new),
                            )
                        )
            else:
                for date in tqdm.auto.tqdm(self.dates_new, "您现在处于单核运算状态，建议仅在调试时使用单核"):
                    df = self.select_one_calculate(
                        date=date,
                        func=func,
                        resample_frequency=resample_frequency,
                        opens_in=opens_in,
                        highs_in=highs_in,
                        lows_in=lows_in,
                        amounts_in=amounts_in,
                        merge_them=merge_them,
                    )
                    self.factor_new.append(df)
            # 拼接新的和旧的
            if self.factor_old is not None:
                self.factor = pd.concat(
                    [self.factor_old] + self.factor_new
                ).sort_index()
            else:
                self.factor = pd.concat(self.factor_new).sort_index()
            self.factor = drop_duplicates_index(self.factor.dropna(how="all"))
            new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
            # 存入本地
            self.factor.to_parquet(self.factor_file)
            logger.info(f"截止到{new_end_date}的因子值计算完了")
            # 删除存储在questdb的中途备份数据
            try:
                self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
                logger.info("备份在questdb的表格已删除")
            except Exception:
                logger.warning("删除questdb中表格时，存在某个未知错误，请当心")

        else:
            self.factor = drop_duplicates_index(self.factor_old)
            # 存入本地
            self.factor.to_parquet(self.factor_file)
            new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
            logger.info(f"当前截止到{new_end_date}的因子值已经是最新的了")

    def drop_table(self):
        """直接删除存储在questdb中的暂存数据"""
        try:
            self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
            logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
        except Exception:
            logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

__call__()

获得经运算产生的因子

Returns

pd.DataFrame 经运算产生的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
6658
6659
6660
6661
6662
6663
6664
6665
6666
def __call__(self) -> pd.DataFrame:
    """获得经运算产生的因子

    Returns
    -------
    `pd.DataFrame`
        经运算产生的因子值
    """
    return self.factor.copy()

__init__(factor_file, project=None, startdate=None, enddate=None, questdb_host='127.0.0.1', ignore_history_in_questdb=0, groupby_code=1)

基于股票逐笔数据，计算因子值，每天的因子值只用到当日的数据

Parameters

factor_file : str 用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾 project : str, optional 该因子所属项目，即子文件夹名称, by default None startdate : int, optional 起始时间，形如20121231，为开区间, by default None enddate : int, optional 截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None questdb_host: str, optional questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1' ignore_history_in_questdb : bool, optional 打断后重新从头计算，清除在questdb中的记录 groupby_target: list, optional groupby计算时，分组的依据, by default ['code']

Source code in pure_ocean_breeze/labor/process.py
Python
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
def __init__(
    self,
    factor_file: str,
    project: str = None,
    startdate: int = None,
    enddate: int = None,
    questdb_host: str = "127.0.0.1",
    ignore_history_in_questdb: bool = 0,
    groupby_code: bool = 1,
) -> None:
    """基于股票逐笔数据，计算因子值，每天的因子值只用到当日的数据

    Parameters
    ----------
    factor_file : str
        用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
    project : str, optional
        该因子所属项目，即子文件夹名称, by default None
    startdate : int, optional
        起始时间，形如20121231，为开区间, by default None
    enddate : int, optional
        截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
    questdb_host: str, optional
        questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
    ignore_history_in_questdb : bool, optional
        打断后重新从头计算，清除在questdb中的记录
    groupby_target: list, optional
        groupby计算时，分组的依据, by default ['code']
    """
    homeplace = HomePlace()
    self.groupby_code = groupby_code
    # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
    pinyin = Pinyin()
    self.factor_file_pinyin = pinyin.get_pinyin(
        factor_file.replace(".parquet", ""), ""
    )
    self.factor_steps = Questdb(host=questdb_host)
    if project is not None:
        if not os.path.exists(homeplace.factor_data_file + project):
            os.makedirs(homeplace.factor_data_file + project)
        else:
            logger.info(f"当前正在{project}项目中……")
    else:
        logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
    # 完整的因子文件路径
    if project is not None:
        factor_file = homeplace.factor_data_file + project + "/" + factor_file
    else:
        factor_file = homeplace.factor_data_file + factor_file
    self.factor_file = factor_file
    # 读入之前的因子
    if os.path.exists(factor_file):
        factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
        )
        factor_old = self.factor_steps.get_data_with_tuple(
            f"select * from '{self.factor_file_pinyin}'"
        ).drop_duplicates(subset=["date", "code"])
        factor_old = factor_old.pivot(index="date", columns="code", values="fac")
        factor_old = factor_old.sort_index()
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif ignore_history_in_questdb and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
        )
        factor_old = self.factor_steps.do_order(
            f"drop table '{self.factor_file_pinyin}'"
        )
        self.factor_old = None
        self.dates_old = []
        logger.info("删除完毕，正在重新计算")
    else:
        self.factor_old = None
        self.dates_old = []
        logger.info("这个因子以前没有，正在重新计算")
    # 读取当前所有的日子
    dates_all = os.listdir(homeplace.tick_by_tick_data)
    dates_all = [i.split(".")[0] for i in dates_all if i.endswith(".parquet")]
    dates_all = [i.replace("-", "") for i in dates_all]
    dates_all = [int(i) for i in dates_all if "20" if i]
    if startdate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i >= startdate]
    if enddate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i <= enddate]
    self.dates_all = dates_all
    # 需要新补充的日子
    self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
    if len(self.dates_new) == 0:
        ...
    elif len(self.dates_new) == 1:
        self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
        print(f"只缺一天{self.dates_new[0]}")
    else:
        dates = [pd.Timestamp(str(i)) for i in self.dates_new]
        intervals = [[]] * len(dates)
        interbee = 0
        intervals[0] = intervals[0] + [dates[0]]
        for i in range(len(dates) - 1):
            val1 = dates[i]
            val2 = dates[i + 1]
            if val2 - val1 < pd.Timedelta(days=30):
                ...
            else:
                interbee = interbee + 1
            intervals[interbee] = intervals[interbee] + [val2]
        intervals = [i for i in intervals if len(i) > 0]
        print(f"共{len(intervals)}个时间区间，分别是")
        for date in intervals:
            print(f"从{date[0]}到{date[-1]}")
        self.dates_new_intervals = intervals
    self.factor_new = []
    self.age = read_daily(age=1)
    self.state = read_daily(state=1)
    self.closes_unadj = read_daily(close=1, unadjust=1).shift(1)

drop_table()

直接删除存储在questdb中的暂存数据

Source code in pure_ocean_breeze/labor/process.py
Python
6943
6944
6945
6946
6947
6948
6949
def drop_table(self):
    """直接删除存储在questdb中的暂存数据"""
    try:
        self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
        logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
    except Exception:
        logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

get_daily_factors(func, n_jobs=1, fields='*', resample_frequency=None, opens_in=0, highs_in=0, lows_in=0, amounts_in=0, merge_them=0, use_mpire=0)

每次抽取chunksize天的截面上全部股票的分钟数据 对每天的股票的数据计算因子值

Parameters

func : Callable 用于计算因子值的函数 n_jobs : int, optional 并行数量, by default 1 fields : str, optional 要读取的字段，可选包含date,code,price,amount,saleamount,buyamount,action,saleid,saleprice,buyid,buyprice，其中date,code必须包含, by default '*' resample_frequency : str, optional 将逐笔数据转化为秒级或分钟频数据，可以填写要转化的频率，如'3s'（3秒数据），'1m'（1分钟数据）， 指定此参数后，将自动生成一个self.closes的收盘价矩阵(index为时间,columns为股票代码,values为收盘价)， 可在循环计算的函数中使用self.closes来调用计算好的值, by default None opens_in : bool, optional 在resample_frequency不为None的情况下，可以使用此参数，提前计算好开盘价矩阵(index为时间,columns为股票代码,values为开盘价)， 可在循环计算的函数中使用self.opens来调用计算好的值，by default 0 highs_in : bool, optional 在resample_frequency不为None的情况下，可以使用此参数，提前计算好最高价矩阵(index为时间,columns为股票代码,values为最高价)， 可在循环计算的函数中使用self.highs来调用计算好的值，by default 0 lows_in : bool, optional 在resample_frequency不为None的情况下，可以使用此参数，提前计算好最低价矩阵(index为时间,columns为股票代码,values为最低价)， 可在循环计算的函数中使用self.lows来调用计算好的值，by default 0 amounts_in : bool, optional 在resample_frequency不为None的情况下，可以使用此参数，提前计算好成交额矩阵(index为时间,columns为股票代码,values为成交量)， 可在循环计算的函数中使用self.amounts来调用计算好的值，by default 0 merge_them : bool, optional 在resample_frequency不为None的情况下，可以使用此参数，将计算好的因子值合并到一起，生成类似于分钟数据的sql形式，by default 0 use_mpire : bool, optional 并行是否使用mpire，默认使用concurrent，by default 0

Source code in pure_ocean_breeze/labor/process.py
Python
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
def get_daily_factors(
    self,
    func: Callable,
    n_jobs: int = 1,
    fields: str = "*",
    resample_frequency: str = None,
    opens_in: bool = 0,
    highs_in: bool = 0,
    lows_in: bool = 0,
    amounts_in: bool = 0,
    merge_them: bool = 0,
    use_mpire: bool = 0,
) -> None:
    """每次抽取chunksize天的截面上全部股票的分钟数据
    对每天的股票的数据计算因子值

    Parameters
    ----------
    func : Callable
        用于计算因子值的函数
    n_jobs : int, optional
        并行数量, by default 1
    fields : str, optional
        要读取的字段，可选包含`date,code,price,amount,saleamount,buyamount,action,saleid,saleprice,buyid,buyprice`，其中date,code必须包含, by default `'*'`
    resample_frequency : str, optional
        将逐笔数据转化为秒级或分钟频数据，可以填写要转化的频率，如'3s'（3秒数据），'1m'（1分钟数据），
        指定此参数后，将自动生成一个self.closes的收盘价矩阵(index为时间,columns为股票代码,values为收盘价)，
        可在循环计算的函数中使用`self.closes`来调用计算好的值, by default None
    opens_in : bool, optional
        在resample_frequency不为None的情况下，可以使用此参数，提前计算好开盘价矩阵(index为时间,columns为股票代码,values为开盘价)，
        可在循环计算的函数中使用`self.opens`来调用计算好的值，by default 0
    highs_in : bool, optional
        在resample_frequency不为None的情况下，可以使用此参数，提前计算好最高价矩阵(index为时间,columns为股票代码,values为最高价)，
        可在循环计算的函数中使用`self.highs`来调用计算好的值，by default 0
    lows_in : bool, optional
        在resample_frequency不为None的情况下，可以使用此参数，提前计算好最低价矩阵(index为时间,columns为股票代码,values为最低价)，
        可在循环计算的函数中使用`self.lows`来调用计算好的值，by default 0
    amounts_in : bool, optional
        在resample_frequency不为None的情况下，可以使用此参数，提前计算好成交额矩阵(index为时间,columns为股票代码,values为成交量)，
        可在循环计算的函数中使用`self.amounts`来调用计算好的值，by default 0
    merge_them : bool, optional
        在resample_frequency不为None的情况下，可以使用此参数，将计算好的因子值合并到一起，生成类似于分钟数据的sql形式，by default 0
    use_mpire : bool, optional
        并行是否使用mpire，默认使用concurrent，by default 0
    """
    if len(self.dates_new) > 0:
        if n_jobs > 1:
            if use_mpire:
                with WorkerPool(n_jobs=n_jobs) as pool:
                    self.factor_new = pool.map(
                        lambda x: self.select_one_calculate(
                            date=x,
                            func=func,
                            fields=fields,
                            resample_frequency=resample_frequency,
                            opens_in=opens_in,
                            highs_in=highs_in,
                            lows_in=lows_in,
                            amounts_in=amounts_in,
                            merge_them=merge_them,
                        ),
                        self.dates_new,
                        progress_bar=True,
                    )
            else:
                with concurrent.futures.ThreadPoolExecutor(
                    max_workers=n_jobs
                ) as executor:
                    self.factor_new = list(
                        tqdm.auto.tqdm(
                            executor.map(
                                lambda x: self.select_one_calculate(
                                    date=x,
                                    func=func,
                                    fields=fields,
                                    resample_frequency=resample_frequency,
                                    opens_in=opens_in,
                                    highs_in=highs_in,
                                    lows_in=lows_in,
                                    amounts_in=amounts_in,
                                    merge_them=merge_them,
                                ),
                                self.dates_new,
                            ),
                            total=len(self.dates_new),
                        )
                    )
        else:
            for date in tqdm.auto.tqdm(self.dates_new, "您现在处于单核运算状态，建议仅在调试时使用单核"):
                df = self.select_one_calculate(
                    date=date,
                    func=func,
                    resample_frequency=resample_frequency,
                    opens_in=opens_in,
                    highs_in=highs_in,
                    lows_in=lows_in,
                    amounts_in=amounts_in,
                    merge_them=merge_them,
                )
                self.factor_new.append(df)
        # 拼接新的和旧的
        if self.factor_old is not None:
            self.factor = pd.concat(
                [self.factor_old] + self.factor_new
            ).sort_index()
        else:
            self.factor = pd.concat(self.factor_new).sort_index()
        self.factor = drop_duplicates_index(self.factor.dropna(how="all"))
        new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
        # 存入本地
        self.factor.to_parquet(self.factor_file)
        logger.info(f"截止到{new_end_date}的因子值计算完了")
        # 删除存储在questdb的中途备份数据
        try:
            self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
            logger.info("备份在questdb的表格已删除")
        except Exception:
            logger.warning("删除questdb中表格时，存在某个未知错误，请当心")

    else:
        self.factor = drop_duplicates_index(self.factor_old)
        # 存入本地
        self.factor.to_parquet(self.factor_file)
        new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
        logger.info(f"当前截止到{new_end_date}的因子值已经是最新的了")

pure_fall_second

Bases: object

对单只股票单日进行操作

Source code in pure_ocean_breeze/labor/process.py
Python
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
class pure_fall_second(object):
    """对单只股票单日进行操作"""

    def __init__(
        self,
        factor_file: str,
        project: str = None,
        startdate: int = None,
        enddate: int = None,
        questdb_host: str = "127.0.0.1",
        ignore_history_in_questdb: bool = 0,
        groupby_target: list = ["date", "code"],
    ) -> None:
        """基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

        Parameters
        ----------
        factor_file : str
            用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
        project : str, optional
            该因子所属项目，即子文件夹名称, by default None
        startdate : int, optional
            起始时间，形如20121231，为开区间, by default None
        enddate : int, optional
            截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
        questdb_host: str, optional
            questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
        ignore_history_in_questdb : bool, optional
            打断后重新从头计算，清除在questdb中的记录
        groupby_target: list, optional
            groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算，
            则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']
        """
        homeplace = HomePlace()
        self.groupby_target = groupby_target
        self.chc = ClickHouseClient("second_data")
        # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
        pinyin = Pinyin()
        self.factor_file_pinyin = pinyin.get_pinyin(
            factor_file.replace(".parquet", ""), ""
        )
        self.factor_steps = Questdb(host=questdb_host)
        if project is not None:
            if not os.path.exists(homeplace.factor_data_file + project):
                os.makedirs(homeplace.factor_data_file + project)
            else:
                logger.info(f"当前正在{project}项目中……")
        else:
            logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
        # 完整的因子文件路径
        if project is not None:
            factor_file = homeplace.factor_data_file + project + "/" + factor_file
        else:
            factor_file = homeplace.factor_data_file + factor_file
        self.factor_file = factor_file
        # 读入之前的因子
        if os.path.exists(factor_file):
            factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
            )
            factor_old = self.factor_steps.get_data_with_tuple(
                f"select * from '{self.factor_file_pinyin}'"
            ).drop_duplicates(subset=["date", "code"])
            factor_old = factor_old.pivot(index="date", columns="code", values="fac")
            factor_old = factor_old.sort_index()
            self.factor_old = factor_old
            # 已经算好的日子
            dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
            self.dates_old = dates_old
        elif ignore_history_in_questdb and self.factor_file_pinyin in list(
            self.factor_steps.get_data("show tables").table
        ):
            logger.info(
                f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
            )
            factor_old = self.factor_steps.do_order(
                f"drop table '{self.factor_file_pinyin}'"
            )
            self.factor_old = None
            self.dates_old = []
            logger.info("删除完毕，正在重新计算")
        else:
            self.factor_old = None
            self.dates_old = []
            logger.info("这个因子以前没有，正在重新计算")
        # 读取当前所有的日子
        dates_all = self.chc.show_all_dates(f"second_data_stock_10s")
        dates_all = [int(i) for i in dates_all]
        if startdate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i >= startdate]
        if enddate is None:
            ...
        else:
            dates_all = [i for i in dates_all if i <= enddate]
        self.dates_all = dates_all
        # 需要新补充的日子
        self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
        if len(self.dates_new) == 0:
            ...
        elif len(self.dates_new) == 1:
            self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
            print(f"只缺一天{self.dates_new[0]}")
        else:
            dates = [pd.Timestamp(str(i)) for i in self.dates_new]
            intervals = [[]] * len(dates)
            interbee = 0
            intervals[0] = intervals[0] + [dates[0]]
            for i in range(len(dates) - 1):
                val1 = dates[i]
                val2 = dates[i + 1]
                if val2 - val1 < pd.Timedelta(days=30):
                    ...
                else:
                    interbee = interbee + 1
                intervals[interbee] = intervals[interbee] + [val2]
            intervals = [i for i in intervals if len(i) > 0]
            print(f"共{len(intervals)}个时间区间，分别是")
            for date in intervals:
                print(f"从{date[0]}到{date[-1]}")
            self.dates_new_intervals = intervals
        self.factor_new = []

    def __call__(self) -> pd.DataFrame:
        """获得经运算产生的因子

        Returns
        -------
        `pd.DataFrame`
            经运算产生的因子值
        """
        return self.factor.copy()

    def forward_dates(self, dates, many_days):
        dates_index = [self.dates_all.index(i) for i in dates]

        def value(x, a):
            if x >= 0:
                return a[x]
            else:
                return None

        return [value(i - many_days, self.dates_all) for i in dates_index]

    def select_one_calculate(
        self,
        date: pd.Timestamp,
        func: Callable,
        fields: str = "*",
    ) -> None:
        the_func = partial(func)
        if not isinstance(date, int):
            date = int(datetime.datetime.strftime(date, "%Y%m%d"))
        # 开始计算因子值

        sql_order = f"select {fields} from second_data.second_data_stock_10s where toYYYYMMDD(date)=date order by code,date"
        df = self.chc.get_data(sql_order)
        df = ((df.set_index(["code", "date"])) / 100).reset_index()
        df = df.groupby(self.groupby_target).apply(the_func)
        if self.groupby_target == ["date", "code"]:
            df = df.to_frame("fac").reset_index()
            df.columns = ["date", "code", "fac"]
        else:
            df = df.reset_index()
        if (df is not None) and (df.shape[0] > 0):
            df = df.pivot(columns="code", index="date", values="fac")
            df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
            to_save = df.stack().reset_index()
            to_save.columns = ["date", "code", "fac"]
            self.factor_steps.write_via_df(
                to_save, self.factor_file_pinyin, tuple_col="fac"
            )
            return df

    def select_many_calculate(
        self,
        dates: List[pd.Timestamp],
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        many_days: int = 1,
        n_jobs: int = 1,
    ) -> None:
        the_func = partial(func)
        factor_new = []
        dates = [int(datetime.datetime.strftime(i, "%Y%m%d")) for i in dates]
        if many_days == 1:
            # 将需要更新的日子分块，每200天一组，一起运算
            dates_new_len = len(dates)
            cut_points = list(range(0, dates_new_len, chunksize)) + [dates_new_len - 1]
            if cut_points[-1] == cut_points[-2]:
                cut_points = cut_points[:-1]
            cuts = tuple(zip(cut_points[:-many_days], cut_points[many_days:]))
            df_first = self.select_one_calculate(
                date=dates[0],
                func=func,
                fields=fields,
            )
            factor_new.append(df_first)

            def cal_one(date1, date2):
                if self.clickhouse == 1:
                    sql_order = f"select {fields} from minute_data.minute_data_{self.kind} where date>{dates[date1] * 100} and date<={dates[date2] * 100} order by code,date,num"
                else:
                    sql_order = f"select {fields} from minute_data_{self.kind} where cast(date as int)>{dates[date1]} and cast(date as int)<={dates[date2]} order by code,date,num"

                df = self.chc.get_data(sql_order)
                if self.clickhouse == 1:
                    df = ((df.set_index("code")) / 100).reset_index()
                else:
                    df.num = df.num.astype(int)
                    df.date = df.date.astype(int)
                    df = df.sort_values(["date", "num"])
                df = df.groupby(self.groupby_target).apply(the_func)
                if self.groupby_target == ["date", "code"]:
                    df = df.to_frame("fac").reset_index()
                    df.columns = ["date", "code", "fac"]
                else:
                    df = df.reset_index()
                df = df.pivot(columns="code", index="date", values="fac")
                df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
                to_save = df.stack().reset_index()
                to_save.columns = ["date", "code", "fac"]
                self.factor_steps.write_via_df(
                    to_save, self.factor_file_pinyin, tuple_col="fac"
                )
                return df

            if n_jobs > 1:
                with concurrent.futures.ThreadPoolExecutor(
                    max_workers=n_jobs
                ) as executor:
                    factor_new_more = list(
                        tqdm.auto.tqdm(executor.map(cal_one, cuts), total=len(cuts))
                    )
                factor_new = factor_new + factor_new_more
            else:
                # 开始计算因子值
                for date1, date2 in tqdm.auto.tqdm(cuts, desc="不知乘月几人归，落月摇情满江树。"):
                    df = cal_one(date1, date2)
                    factor_new.append(df)
        else:

            def cal_two(date1, date2):
                if date1 is not None:
                    if self.clickhouse == 1:
                        sql_order = f"select {fields} from minute_data.minute_data_{self.kind} where date>{date1*100} and date<={date2*100} order by code,date,num"
                    else:
                        sql_order = f"select {fields} from minute_data_{self.kind} where cast(date as int)>{date1} and cast(date as int)<={date2} order by code,date,num"

                    df = self.chc.get_data(sql_order)
                    if self.clickhouse == 1:
                        df = ((df.set_index("code")) / 100).reset_index()
                    else:
                        df.num = df.num.astype(int)
                        df.date = df.date.astype(int)
                        df = df.sort_values(["date", "num"])
                    if self.groupby_target == [
                        "date",
                        "code",
                    ] or self.groupby_target == ["code"]:
                        df = df.groupby(["code"]).apply(the_func).reset_index()
                    else:
                        df = the_func(df)
                    df = df.assign(date=date2)
                    df.columns = ["code", "fac", "date"]
                    df = df.pivot(columns="code", index="date", values="fac")
                    df.index = pd.to_datetime(df.index.astype(str), format="%Y%m%d")
                    to_save = df.stack().reset_index()
                    to_save.columns = ["date", "code", "fac"]
                    self.factor_steps.write_via_df(
                        to_save, self.factor_file_pinyin, tuple_col="fac"
                    )
                    return df

            pairs = self.forward_dates(dates, many_days=many_days)
            cuts2 = tuple(zip(pairs, dates))
            if n_jobs > 1:
                with concurrent.futures.ThreadPoolExecutor(
                    max_workers=n_jobs
                ) as executor:
                    factor_new_more = list(
                        tqdm.auto.tqdm(executor.map(cal_two, cuts2), total=len(cuts2))
                    )
                factor_new = factor_new + factor_new_more
            else:
                # 开始计算因子值
                for date1, date2 in tqdm.auto.tqdm(cuts2, desc="知不可乎骤得，托遗响于悲风。"):
                    df = cal_two(date1, date2)
                    factor_new.append(df)

        if len(factor_new) > 0:
            factor_new = pd.concat(factor_new)
            return factor_new
        else:
            return None

    def select_any_calculate(
        self,
        dates: List[pd.Timestamp],
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
    ) -> None:
        if len(dates) == 1 and many_days == 1:
            res = self.select_one_calculate(
                dates[0],
                func=func,
                fields=fields,
                show_time=show_time,
            )
        else:
            res = self.select_many_calculate(
                dates=dates,
                func=func,
                fields=fields,
                chunksize=chunksize,
                show_time=show_time,
                many_days=many_days,
                n_jobs=n_jobs,
            )
        if res is not None:
            self.factor_new.append(res)
        return res

    @staticmethod
    def for_cross_via_str(func):
        """返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值
        例如
        ```python
        return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
        ```
        上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列"""

        def full_run(df, *args, **kwargs):
            codes = sorted(list(set(df.code)))
            res = func(df, *args, **kwargs)
            if isinstance(res[0], list):
                kind = 1
                res = [",".join(i) for i in res]
            else:
                kind = 0
            df = pd.DataFrame({"code": codes, "fac": res})
            if kind:
                df.fac = df.fac.apply(lambda x: [float(i) for i in x.split(",")])
            return df

        return full_run

    @staticmethod
    def for_cross_via_zip(func):
        """返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值
        例如
        ```python
        return (
                    pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                    pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                )
        ```
        上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值
        """

        def full_run(df, *args, **kwargs):
            res = func(df, *args, **kwargs)
            if isinstance(res, pd.Series):
                res = res.reset_index()
                res.columns = ["code", "fac"]
                return res
            elif isinstance(res, pd.DataFrame):
                res.columns = [f"fac{i}" for i in range(len(res.columns))]
                res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
                res = res[["fac"]].reset_index()
                res.columns = ["code", "fac"]
                return res
            elif res is None:
                ...
            else:
                res = pd.concat(res, axis=1)
                res.columns = [f"fac{i}" for i in range(len(res.columns))]
                res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
                res = res[["fac"]].reset_index()
                res.columns = ["code", "fac"]
                return res

        return full_run

    @kk.desktop_sender(title="嘿，分钟数据处理完啦～🎈")
    def get_daily_factors(
        self,
        func: Callable,
        fields: str = "*",
        chunksize: int = 10,
        show_time: bool = 0,
        many_days: int = 1,
        n_jobs: int = 1,
    ) -> None:
        """每次抽取chunksize天的截面上全部股票的分钟数据
        对每天的股票的数据计算因子值

        Parameters
        ----------
        func : Callable
            用于计算因子值的函数
        fields : str, optional
            股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open'
            提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*"
        chunksize : int, optional
            每次读取的截面上的天数, by default 10
        show_time : bool, optional
            展示每次读取数据所需要的时间, by default 0
        many_days : int, optional
            计算某天的因子值时，需要使用之前多少天的数据
        n_jobs : int, optional
            并行数量，不建议设置为大于2的数，此外当此参数大于1时，请使用questdb数据库来读取分钟数据, by default 1
        """
        if len(self.dates_new) > 0:
            for interval in self.dates_new_intervals:
                df = self.select_any_calculate(
                    dates=interval,
                    func=func,
                    fields=fields,
                    chunksize=chunksize,
                    show_time=show_time,
                    many_days=many_days,
                    n_jobs=n_jobs,
                )
            self.factor_new = pd.concat(self.factor_new)
            # 拼接新的和旧的
            self.factor = pd.concat([self.factor_old, self.factor_new]).sort_index()
            self.factor = drop_duplicates_index(self.factor.dropna(how="all"))
            new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
            # 存入本地
            self.factor.to_parquet(self.factor_file)
            logger.info(f"截止到{new_end_date}的因子值计算完了")
            # 删除存储在questdb的中途备份数据
            try:
                self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
                logger.info("备份在questdb的表格已删除")
            except Exception:
                logger.warning("删除questdb中表格时，存在某个未知错误，请当心")

        else:
            self.factor = drop_duplicates_index(self.factor_old)
            # 存入本地
            self.factor.to_parquet(self.factor_file)
            new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
            logger.info(f"当前截止到{new_end_date}的因子值已经是最新的了")

    def drop_table(self):
        """直接删除存储在questdb中的暂存数据"""
        try:
            self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
            logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
        except Exception:
            logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

__call__()

获得经运算产生的因子

Returns

pd.DataFrame 经运算产生的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
6188
6189
6190
6191
6192
6193
6194
6195
6196
def __call__(self) -> pd.DataFrame:
    """获得经运算产生的因子

    Returns
    -------
    `pd.DataFrame`
        经运算产生的因子值
    """
    return self.factor.copy()

__init__(factor_file, project=None, startdate=None, enddate=None, questdb_host='127.0.0.1', ignore_history_in_questdb=0, groupby_target=['date', 'code'])

基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

Parameters

factor_file : str 用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾 project : str, optional 该因子所属项目，即子文件夹名称, by default None startdate : int, optional 起始时间，形如20121231，为开区间, by default None enddate : int, optional 截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None questdb_host: str, optional questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1' ignore_history_in_questdb : bool, optional 打断后重新从头计算，清除在questdb中的记录 groupby_target: list, optional groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算， 则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']

Source code in pure_ocean_breeze/labor/process.py
Python
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
def __init__(
    self,
    factor_file: str,
    project: str = None,
    startdate: int = None,
    enddate: int = None,
    questdb_host: str = "127.0.0.1",
    ignore_history_in_questdb: bool = 0,
    groupby_target: list = ["date", "code"],
) -> None:
    """基于clickhouse的分钟数据，计算因子值，每天的因子值只用到当日的数据

    Parameters
    ----------
    factor_file : str
        用于保存因子值的文件名，需为parquet文件，以'.parquet'结尾
    project : str, optional
        该因子所属项目，即子文件夹名称, by default None
    startdate : int, optional
        起始时间，形如20121231，为开区间, by default None
    enddate : int, optional
        截止时间，形如20220814，为闭区间，为空则计算到最近数据, by default None
    questdb_host: str, optional
        questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
    ignore_history_in_questdb : bool, optional
        打断后重新从头计算，清除在questdb中的记录
    groupby_target: list, optional
        groupby计算时，分组的依据，使用此参数时，自定义函数的部分，如果指定按照['date']分组groupby计算，
        则返回时，应当返回一个两列的dataframe，第一列为股票代码，第二列为为因子值, by default ['date','code']
    """
    homeplace = HomePlace()
    self.groupby_target = groupby_target
    self.chc = ClickHouseClient("second_data")
    # 将计算到一半的因子，存入questdb中，避免中途被打断后重新计算，表名即为因子文件名的汉语拼音
    pinyin = Pinyin()
    self.factor_file_pinyin = pinyin.get_pinyin(
        factor_file.replace(".parquet", ""), ""
    )
    self.factor_steps = Questdb(host=questdb_host)
    if project is not None:
        if not os.path.exists(homeplace.factor_data_file + project):
            os.makedirs(homeplace.factor_data_file + project)
        else:
            logger.info(f"当前正在{project}项目中……")
    else:
        logger.warning("当前因子不属于任何项目，这将造成因子数据文件夹的混乱，不便于管理，建议指定一个项目名称")
    # 完整的因子文件路径
    if project is not None:
        factor_file = homeplace.factor_data_file + project + "/" + factor_file
    else:
        factor_file = homeplace.factor_data_file + factor_file
    self.factor_file = factor_file
    # 读入之前的因子
    if os.path.exists(factor_file):
        factor_old = drop_duplicates_index(pd.read_parquet(self.factor_file))
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif (not ignore_history_in_questdb) and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，现在将读取上次的数据，继续计算"
        )
        factor_old = self.factor_steps.get_data_with_tuple(
            f"select * from '{self.factor_file_pinyin}'"
        ).drop_duplicates(subset=["date", "code"])
        factor_old = factor_old.pivot(index="date", columns="code", values="fac")
        factor_old = factor_old.sort_index()
        self.factor_old = factor_old
        # 已经算好的日子
        dates_old = sorted(list(factor_old.index.strftime("%Y%m%d").astype(int)))
        self.dates_old = dates_old
    elif ignore_history_in_questdb and self.factor_file_pinyin in list(
        self.factor_steps.get_data("show tables").table
    ):
        logger.info(
            f"上次计算途中被打断，已经将数据备份在questdb数据库的表{self.factor_file_pinyin}中，但您选择重新计算，所以正在删除原来的数据，从头计算"
        )
        factor_old = self.factor_steps.do_order(
            f"drop table '{self.factor_file_pinyin}'"
        )
        self.factor_old = None
        self.dates_old = []
        logger.info("删除完毕，正在重新计算")
    else:
        self.factor_old = None
        self.dates_old = []
        logger.info("这个因子以前没有，正在重新计算")
    # 读取当前所有的日子
    dates_all = self.chc.show_all_dates(f"second_data_stock_10s")
    dates_all = [int(i) for i in dates_all]
    if startdate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i >= startdate]
    if enddate is None:
        ...
    else:
        dates_all = [i for i in dates_all if i <= enddate]
    self.dates_all = dates_all
    # 需要新补充的日子
    self.dates_new = sorted([i for i in dates_all if i not in self.dates_old])
    if len(self.dates_new) == 0:
        ...
    elif len(self.dates_new) == 1:
        self.dates_new_intervals = [[pd.Timestamp(str(self.dates_new[0]))]]
        print(f"只缺一天{self.dates_new[0]}")
    else:
        dates = [pd.Timestamp(str(i)) for i in self.dates_new]
        intervals = [[]] * len(dates)
        interbee = 0
        intervals[0] = intervals[0] + [dates[0]]
        for i in range(len(dates) - 1):
            val1 = dates[i]
            val2 = dates[i + 1]
            if val2 - val1 < pd.Timedelta(days=30):
                ...
            else:
                interbee = interbee + 1
            intervals[interbee] = intervals[interbee] + [val2]
        intervals = [i for i in intervals if len(i) > 0]
        print(f"共{len(intervals)}个时间区间，分别是")
        for date in intervals:
            print(f"从{date[0]}到{date[-1]}")
        self.dates_new_intervals = intervals
    self.factor_new = []

drop_table()

直接删除存储在questdb中的暂存数据

Source code in pure_ocean_breeze/labor/process.py
Python
6516
6517
6518
6519
6520
6521
6522
def drop_table(self):
    """直接删除存储在questdb中的暂存数据"""
    try:
        self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
        logger.success(f"暂存在questdb中的数据表格'{self.factor_file_pinyin}'已经删除")
    except Exception:
        logger.warning(f"您要删除的表格'{self.factor_file_pinyin}'已经不存在了，请检查")

for_cross_via_str(func) staticmethod

返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值 例如

Python
1
return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列

Source code in pure_ocean_breeze/labor/process.py
Python
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
@staticmethod
def for_cross_via_str(func):
    """返回值为两层的list，每一个里层的小list为单个股票在这一天的返回值
    例如
    ```python
    return [[0.11,0.24,0.55],[2.59,1.99,0.43],[1.32,8.88,7.77]……]
    ```
    上例中，每个股票一天返回三个因子值，里层的list按照股票代码顺序排列"""

    def full_run(df, *args, **kwargs):
        codes = sorted(list(set(df.code)))
        res = func(df, *args, **kwargs)
        if isinstance(res[0], list):
            kind = 1
            res = [",".join(i) for i in res]
        else:
            kind = 0
        df = pd.DataFrame({"code": codes, "fac": res})
        if kind:
            df.fac = df.fac.apply(lambda x: [float(i) for i in x.split(",")])
        return df

    return full_run

for_cross_via_zip(func) staticmethod

返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值 例如

Python
1
2
3
4
return (
            pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
            pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
        )
上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值

Source code in pure_ocean_breeze/labor/process.py
Python
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
@staticmethod
def for_cross_via_zip(func):
    """返回值为多个pd.Series，每个pd.Series的index为股票代码，values为单个因子值
    例如
    ```python
    return (
                pd.Series([1.54,8.77,9.99……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
                pd.Series([3.54,6.98,9.01……],index=['000001.SZ','000002.SZ','000004.SZ'……]),
            )
    ```
    上例中，每个股票一天返回两个因子值，每个pd.Series对应一个因子值
    """

    def full_run(df, *args, **kwargs):
        res = func(df, *args, **kwargs)
        if isinstance(res, pd.Series):
            res = res.reset_index()
            res.columns = ["code", "fac"]
            return res
        elif isinstance(res, pd.DataFrame):
            res.columns = [f"fac{i}" for i in range(len(res.columns))]
            res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
            res = res[["fac"]].reset_index()
            res.columns = ["code", "fac"]
            return res
        elif res is None:
            ...
        else:
            res = pd.concat(res, axis=1)
            res.columns = [f"fac{i}" for i in range(len(res.columns))]
            res = res.assign(fac=list(zip(*[res[i] for i in list(res.columns)])))
            res = res[["fac"]].reset_index()
            res.columns = ["code", "fac"]
            return res

    return full_run

get_daily_factors(func, fields='*', chunksize=10, show_time=0, many_days=1, n_jobs=1)

每次抽取chunksize天的截面上全部股票的分钟数据 对每天的股票的数据计算因子值

Parameters

func : Callable 用于计算因子值的函数 fields : str, optional 股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open' 提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*" chunksize : int, optional 每次读取的截面上的天数, by default 10 show_time : bool, optional 展示每次读取数据所需要的时间, by default 0 many_days : int, optional 计算某天的因子值时，需要使用之前多少天的数据 n_jobs : int, optional 并行数量，不建议设置为大于2的数，此外当此参数大于1时，请使用questdb数据库来读取分钟数据, by default 1

Source code in pure_ocean_breeze/labor/process.py
Python
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
@kk.desktop_sender(title="嘿，分钟数据处理完啦～🎈")
def get_daily_factors(
    self,
    func: Callable,
    fields: str = "*",
    chunksize: int = 10,
    show_time: bool = 0,
    many_days: int = 1,
    n_jobs: int = 1,
) -> None:
    """每次抽取chunksize天的截面上全部股票的分钟数据
    对每天的股票的数据计算因子值

    Parameters
    ----------
    func : Callable
        用于计算因子值的函数
    fields : str, optional
        股票数据涉及到哪些字段，排除不必要的字段，可以节约读取数据的时间，形如'date,code,num,close,amount,open'
        提取出的数据，自动按照code,date,num排序，因此code,date,num是必不可少的字段, by default "*"
    chunksize : int, optional
        每次读取的截面上的天数, by default 10
    show_time : bool, optional
        展示每次读取数据所需要的时间, by default 0
    many_days : int, optional
        计算某天的因子值时，需要使用之前多少天的数据
    n_jobs : int, optional
        并行数量，不建议设置为大于2的数，此外当此参数大于1时，请使用questdb数据库来读取分钟数据, by default 1
    """
    if len(self.dates_new) > 0:
        for interval in self.dates_new_intervals:
            df = self.select_any_calculate(
                dates=interval,
                func=func,
                fields=fields,
                chunksize=chunksize,
                show_time=show_time,
                many_days=many_days,
                n_jobs=n_jobs,
            )
        self.factor_new = pd.concat(self.factor_new)
        # 拼接新的和旧的
        self.factor = pd.concat([self.factor_old, self.factor_new]).sort_index()
        self.factor = drop_duplicates_index(self.factor.dropna(how="all"))
        new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
        # 存入本地
        self.factor.to_parquet(self.factor_file)
        logger.info(f"截止到{new_end_date}的因子值计算完了")
        # 删除存储在questdb的中途备份数据
        try:
            self.factor_steps.do_order(f"drop table '{self.factor_file_pinyin}'")
            logger.info("备份在questdb的表格已删除")
        except Exception:
            logger.warning("删除questdb中表格时，存在某个未知错误，请当心")

    else:
        self.factor = drop_duplicates_index(self.factor_old)
        # 存入本地
        self.factor.to_parquet(self.factor_file)
        new_end_date = datetime.datetime.strftime(self.factor.index.max(), "%Y%m%d")
        logger.info(f"当前截止到{new_end_date}的因子值已经是最新的了")

pure_fallmount

Bases: pure_fall

继承自父类，专为做因子截面标准化之后相加和因子剔除其他辅助因子的作用

Source code in pure_ocean_breeze/labor/process.py
Python
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
class pure_fallmount(pure_fall):
    """继承自父类，专为做因子截面标准化之后相加和因子剔除其他辅助因子的作用"""

    def __init__(self, monthly_factors):
        """输入月度因子值，以设定新的对象"""
        super(pure_fall, self).__init__()
        self.monthly_factors = monthly_factors

    def __call__(self, monthly=False):
        """为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表"""
        if monthly:
            return self.monthly_factors.copy()
        else:
            try:
                return self.daily_factors.copy()
            except Exception:
                return self.monthly_factors.copy()

    def __add__(self, selfas):
        """返回一个对象，而非一个表格，如需表格请调用对象"""
        fac1 = self.standardlize_in_cross_section(self.monthly_factors)
        fac2s = []
        if not isinstance(selfas, Iterable):
            if not STATES["NO_LOG"]:
                logger.warning(f"{selfas} is changed into Iterable")
            selfas = (selfas,)
        for selfa in selfas:
            fac2 = self.standardlize_in_cross_section(selfa.monthly_factors)
            fac2s.append(fac2)
        for i in fac2s:
            fac1 = fac1 + i
        new_pure = pure_fallmount(fac1)
        return new_pure

    def __mul__(self, selfas):
        """将几个因子横截面标准化之后，使其都为正数，然后因子值相乘"""
        fac1 = self.standardlize_in_cross_section(self.monthly_factors)
        fac1 = fac1 - fac1.min()
        fac2s = []
        if not isinstance(selfas, Iterable):
            if not STATES["NO_LOG"]:
                logger.warning(f"{selfas} is changed into Iterable")
            selfas = (selfas,)
        for selfa in selfas:
            fac2 = self.standardlize_in_cross_section(selfa.monthly_factors)
            fac2 = fac2 - fac2.min()
            fac2s.append(fac2)
        for i in fac2s:
            fac1 = fac1 * i
        new_pure = pure_fall()
        new_pure.monthly_factors = fac1
        return new_pure

    def __sub__(self, selfa):
        """返回对象，如需表格，请调用对象"""
        tqdm.auto.tqdm.pandas()
        if not isinstance(selfa, Iterable):
            if not STATES["NO_LOG"]:
                logger.warning(f"{selfa} is changed into Iterable")
            selfa = (selfa,)
        fac_main = self.wide_to_long(self.monthly_factors, "fac")
        fac_helps = [i.monthly_factors for i in selfa]
        help_names = ["help" + str(i) for i in range(1, (len(fac_helps) + 1))]
        fac_helps = list(map(self.wide_to_long, fac_helps, help_names))
        fac_helps = pd.concat(fac_helps, axis=1)
        facs = pd.concat([fac_main, fac_helps], axis=1).dropna()
        facs = facs.groupby("date").progress_apply(
            lambda x: self.de_in_group(x, help_names)
        )
        facs = facs.unstack()
        facs.columns = list(map(lambda x: x[1], list(facs.columns)))
        new_pure = pure_fallmount(facs)
        return new_pure

    def __gt__(self, selfa):
        """用于输出25分组表格，使用时，以x>y的形式使用，其中x,y均为pure_fall对象
        计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe
        x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子"""
        x = self.monthly_factors.copy()
        y = selfa.monthly_factors.copy()
        x = x.stack().reset_index()
        y = y.stack().reset_index()
        x.columns = ["date", "code", "fac"]
        y.columns = ["date", "code", "fac"]
        shen = pure_moon()
        x = x.groupby("date").apply(lambda df: shen.get_groups(df, 5))
        x = (
            x.reset_index(drop=True)
            .drop(columns=["fac"])
            .rename(columns={"group": "groupx"})
        )
        xy = pd.merge(x, y, on=["date", "code"])
        xy = xy.groupby(["date", "groupx"]).apply(lambda df: shen.get_groups(df, 5))
        xy = (
            xy.reset_index(drop=True)
            .drop(columns=["fac"])
            .rename(columns={"group": "groupy"})
        )
        xy = xy.assign(fac=xy.groupx * 5 + xy.groupy)
        xy = xy[["date", "code", "fac"]]
        xy = xy.set_index(["date", "code"]).unstack()
        xy.columns = [i[1] for i in list(xy.columns)]
        new_pure = pure_fallmount(xy)
        return new_pure

    def __rshift__(self, selfa):
        """用于输出100分组表格，使用时，以x>>y的形式使用，其中x,y均为pure_fall对象
        计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe
        x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子"""
        x = self.monthly_factors.copy()
        y = selfa.monthly_factors.copy()
        x = x.stack().reset_index()
        y = y.stack().reset_index()
        x.columns = ["date", "code", "fac"]
        y.columns = ["date", "code", "fac"]
        shen = pure_moon()
        x = x.groupby("date").apply(lambda df: shen.get_groups(df, 10))
        x = (
            x.reset_index(drop=True)
            .drop(columns=["fac"])
            .rename(columns={"group": "groupx"})
        )
        xy = pd.merge(x, y, on=["date", "code"])
        xy = xy.groupby(["date", "groupx"]).apply(lambda df: shen.get_groups(df, 10))
        xy = (
            xy.reset_index(drop=True)
            .drop(columns=["fac"])
            .rename(columns={"group": "groupy"})
        )
        xy = xy.assign(fac=xy.groupx * 10 + xy.groupy)
        xy = xy[["date", "code", "fac"]]
        xy = xy.set_index(["date", "code"]).unstack()
        xy.columns = [i[1] for i in list(xy.columns)]
        new_pure = pure_fallmount(xy)
        return new_pure

__add__(selfas)

返回一个对象，而非一个表格，如需表格请调用对象

Source code in pure_ocean_breeze/labor/process.py
Python
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
def __add__(self, selfas):
    """返回一个对象，而非一个表格，如需表格请调用对象"""
    fac1 = self.standardlize_in_cross_section(self.monthly_factors)
    fac2s = []
    if not isinstance(selfas, Iterable):
        if not STATES["NO_LOG"]:
            logger.warning(f"{selfas} is changed into Iterable")
        selfas = (selfas,)
    for selfa in selfas:
        fac2 = self.standardlize_in_cross_section(selfa.monthly_factors)
        fac2s.append(fac2)
    for i in fac2s:
        fac1 = fac1 + i
    new_pure = pure_fallmount(fac1)
    return new_pure

__call__(monthly=False)

为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表

Source code in pure_ocean_breeze/labor/process.py
Python
3025
3026
3027
3028
3029
3030
3031
3032
3033
def __call__(self, monthly=False):
    """为了防止属性名太多，忘记了要调用哪个才是结果，因此可以直接输出月度数据表"""
    if monthly:
        return self.monthly_factors.copy()
    else:
        try:
            return self.daily_factors.copy()
        except Exception:
            return self.monthly_factors.copy()

__gt__(selfa)

用于输出25分组表格，使用时，以x>y的形式使用，其中x,y均为pure_fall对象 计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子

Source code in pure_ocean_breeze/labor/process.py
Python
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
def __gt__(self, selfa):
    """用于输出25分组表格，使用时，以x>y的形式使用，其中x,y均为pure_fall对象
    计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe
    x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子"""
    x = self.monthly_factors.copy()
    y = selfa.monthly_factors.copy()
    x = x.stack().reset_index()
    y = y.stack().reset_index()
    x.columns = ["date", "code", "fac"]
    y.columns = ["date", "code", "fac"]
    shen = pure_moon()
    x = x.groupby("date").apply(lambda df: shen.get_groups(df, 5))
    x = (
        x.reset_index(drop=True)
        .drop(columns=["fac"])
        .rename(columns={"group": "groupx"})
    )
    xy = pd.merge(x, y, on=["date", "code"])
    xy = xy.groupby(["date", "groupx"]).apply(lambda df: shen.get_groups(df, 5))
    xy = (
        xy.reset_index(drop=True)
        .drop(columns=["fac"])
        .rename(columns={"group": "groupy"})
    )
    xy = xy.assign(fac=xy.groupx * 5 + xy.groupy)
    xy = xy[["date", "code", "fac"]]
    xy = xy.set_index(["date", "code"]).unstack()
    xy.columns = [i[1] for i in list(xy.columns)]
    new_pure = pure_fallmount(xy)
    return new_pure

__init__(monthly_factors)

输入月度因子值，以设定新的对象

Source code in pure_ocean_breeze/labor/process.py
Python
3020
3021
3022
3023
def __init__(self, monthly_factors):
    """输入月度因子值，以设定新的对象"""
    super(pure_fall, self).__init__()
    self.monthly_factors = monthly_factors

__mul__(selfas)

将几个因子横截面标准化之后，使其都为正数，然后因子值相乘

Source code in pure_ocean_breeze/labor/process.py
Python
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
def __mul__(self, selfas):
    """将几个因子横截面标准化之后，使其都为正数，然后因子值相乘"""
    fac1 = self.standardlize_in_cross_section(self.monthly_factors)
    fac1 = fac1 - fac1.min()
    fac2s = []
    if not isinstance(selfas, Iterable):
        if not STATES["NO_LOG"]:
            logger.warning(f"{selfas} is changed into Iterable")
        selfas = (selfas,)
    for selfa in selfas:
        fac2 = self.standardlize_in_cross_section(selfa.monthly_factors)
        fac2 = fac2 - fac2.min()
        fac2s.append(fac2)
    for i in fac2s:
        fac1 = fac1 * i
    new_pure = pure_fall()
    new_pure.monthly_factors = fac1
    return new_pure

__rshift__(selfa)

用于输出100分组表格，使用时，以x>>y的形式使用，其中x,y均为pure_fall对象 计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子

Source code in pure_ocean_breeze/labor/process.py
Python
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
def __rshift__(self, selfa):
    """用于输出100分组表格，使用时，以x>>y的形式使用，其中x,y均为pure_fall对象
    计算时使用的是他们的月度因子表，即self.monthly_factors属性，为宽数据形式的dataframe
    x应为首先用来的分组的主因子，y为在x分组后的组内继续分组的次因子"""
    x = self.monthly_factors.copy()
    y = selfa.monthly_factors.copy()
    x = x.stack().reset_index()
    y = y.stack().reset_index()
    x.columns = ["date", "code", "fac"]
    y.columns = ["date", "code", "fac"]
    shen = pure_moon()
    x = x.groupby("date").apply(lambda df: shen.get_groups(df, 10))
    x = (
        x.reset_index(drop=True)
        .drop(columns=["fac"])
        .rename(columns={"group": "groupx"})
    )
    xy = pd.merge(x, y, on=["date", "code"])
    xy = xy.groupby(["date", "groupx"]).apply(lambda df: shen.get_groups(df, 10))
    xy = (
        xy.reset_index(drop=True)
        .drop(columns=["fac"])
        .rename(columns={"group": "groupy"})
    )
    xy = xy.assign(fac=xy.groupx * 10 + xy.groupy)
    xy = xy[["date", "code", "fac"]]
    xy = xy.set_index(["date", "code"]).unstack()
    xy.columns = [i[1] for i in list(xy.columns)]
    new_pure = pure_fallmount(xy)
    return new_pure

__sub__(selfa)

返回对象，如需表格，请调用对象

Source code in pure_ocean_breeze/labor/process.py
Python
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
def __sub__(self, selfa):
    """返回对象，如需表格，请调用对象"""
    tqdm.auto.tqdm.pandas()
    if not isinstance(selfa, Iterable):
        if not STATES["NO_LOG"]:
            logger.warning(f"{selfa} is changed into Iterable")
        selfa = (selfa,)
    fac_main = self.wide_to_long(self.monthly_factors, "fac")
    fac_helps = [i.monthly_factors for i in selfa]
    help_names = ["help" + str(i) for i in range(1, (len(fac_helps) + 1))]
    fac_helps = list(map(self.wide_to_long, fac_helps, help_names))
    fac_helps = pd.concat(fac_helps, axis=1)
    facs = pd.concat([fac_main, fac_helps], axis=1).dropna()
    facs = facs.groupby("date").progress_apply(
        lambda x: self.de_in_group(x, help_names)
    )
    facs = facs.unstack()
    facs.columns = list(map(lambda x: x[1], list(facs.columns)))
    new_pure = pure_fallmount(facs)
    return new_pure

pure_fama

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
class pure_fama(object):
    # @lru_cache(maxsize=None)
    def __init__(
        self,
        factors: List[pd.DataFrame],
        minus_group: Union[list, float] = 3,
        backsee: int = 20,
        rets: pd.DataFrame = None,
        value_weighted: bool = 1,
        add_market: bool = 1,
        add_market_series: pd.Series = None,
        factors_names: list = None,
        betas_rets: bool = 0,
        total_cap: bool = 0,
    ) -> None:
        """使用fama三因子的方法，将个股的收益率，拆分出各个因子带来的收益率以及特质的收益率
        分别计算每一期，各个因子收益率的值，超额收益率，因子的暴露，以及特质收益率

        Parameters
        ----------
        factors : List[pd.DataFrame]
            用于解释收益的各个因子值，每一个都是index为时间，columns为股票代码，values为因子值的dataframe
        minus_group : Union[list, float], optional
            每一个因子将截面上的股票分为几组, by default 3
        backsee : int, optional
            做时序回归时，回看的天数, by default 20
        rets : pd.DataFrame, optional
            每只个股的收益率，index为时间，columns为股票代码，values为收益率，默认使用当日日间收益率, by default None
        value_weighted : bool, optional
            是否使用流通市值加权, by default 1
        add_market : bool, optional
            是否加入市场收益率因子，默认使用中证全指的每日日间收益率, by default 1
        add_market_series : bool, optional
            加入的市场收益率的数据，如果没指定，则使用中证全指的日间收益率, by default None
        factors_names : list, optional
            各个因子的名字，默认为fac0(市场收益率因子，如果没有，则从fac1开始),fac1,fac2,fac3, by default None
        betas_rets : bool, optional
            是否计算每只个股的由于暴露在每个因子上所带来的收益率, by default 0
        total_cap : bool, optional
            加权时使用总市值, by default 0
        """
        start = max(
            [int(datetime.datetime.strftime(i.index.min(), "%Y%m%d")) for i in factors]
        )
        self.backsee = backsee
        self.factors = factors
        self.factors_names = factors_names
        if isinstance(minus_group, int):
            minus_group = [minus_group] * len(factors)
        self.minus_group = minus_group
        if rets is None:
            closes = read_daily(close=1, start=start)
            rets = closes / closes.shift(1) - 1
        self.rets = rets
        self.factors_group = [
            to_group(i, group=j) for i, j in zip(self.factors, self.minus_group)
        ]
        self.factors_group_long = [(i == 0) + 0 for i in self.factors_group]
        self.factors_group_short = [
            (i == (j - 1)) + 0 for i, j in zip(self.factors_group, self.minus_group)
        ]
        self.value_weighted = value_weighted
        if value_weighted:
            if total_cap:
                self.cap = read_daily(total_cap=1, start=start)
            self.cap = read_daily(flow_cap=1, start=start)
            self.factors_group_long = [self.cap * i for i in self.factors_group_long]
            self.factors_group_short = [self.cap * i for i in self.factors_group_short]
            self.factors_group_long = [
                (i.T / i.T.sum()).T for i in self.factors_group_long
            ]
            self.factors_group_short = [
                (i.T / i.T.sum()).T for i in self.factors_group_short
            ]
            self.factors_rets_long = [
                (self.rets * i).sum(axis=1).to_frame(f"fac{num+1}")
                for num, i in enumerate(self.factors_group_long)
            ]
            self.factors_rets_short = [
                (self.rets * i).sum(axis=1).to_frame(f"fac{num+1}")
                for num, i in enumerate(self.factors_group_short)
            ]
        else:
            self.factors_rets_long = [
                (self.rets * i).mean(axis=1).to_frame(f"fac{num+1}")
                for num, i in enumerate(self.factors_group_long)
            ]
            self.factors_rets_short = [
                (self.rets * i).mean(axis=1).to_frame(f"fac{num+1}")
                for num, i in enumerate(self.factors_group_short)
            ]
        self.rets_long = pd.concat(self.factors_rets_long, axis=1)
        self.rets_short = pd.concat(self.factors_rets_short, axis=1)
        self.__factors_rets = self.rets_long - self.rets_short
        if add_market_series is not None:
            add_market = 1
        self.add_market = add_market
        if add_market:
            if add_market_series is None:
                closes = read_market(close=1, every_stock=0, start=start).to_frame(
                    "fac0"
                )
            else:
                closes = add_market_series.to_frame("fac0")
            rets = closes / closes.shift(1) - 1
            self.__factors_rets = pd.concat([rets, self.__factors_rets], axis=1)
            if factors_names is not None:
                factors_names = ["市场"] + factors_names
        self.__data = self.make_df(self.rets, self.__factors_rets)
        tqdm.auto.tqdm.pandas()
        self.__coefficients = (
            self.__data.groupby("code").progress_apply(self.ols_in).reset_index()
        )
        self.__coefficients = self.__coefficients.rename(
            columns={
                i: "co" + i for i in list(self.__coefficients.columns) if "fac" in i
            }
        )
        self.__data = pd.merge(
            self.__data.reset_index(), self.__coefficients, on=["date", "code"]
        )
        betas = [
            self.__data[i] * self.__data["co" + i]
            for i in list(self.__data.columns)
            if i.startswith("fac")
        ]
        betas = sum(betas)
        self.__data = self.__data.assign(
            idiosyncratic=self.__data.ret - self.__data.intercept - betas
        )
        self.__idiosyncratic = self.__data.pivot(
            index="date", columns="code", values="idiosyncratic"
        )
        self.__alphas = self.__data.pivot(
            index="date", columns="code", values="intercept"
        )
        if factors_names is None:
            self.__betas = {
                i: self.__data.pivot(index="date", columns="code", values=i)
                for i in list(self.__data.columns)
                if i.startswith("fac")
            }
        else:
            facs = [i for i in list(self.__data.columns) if i.startswith("fac")]
            self.__betas = {
                factors_names[num]: self.__data.pivot(
                    index="date", columns="code", values=i
                )
                for num, i in enumerate(facs)
            }
        if betas_rets:
            if add_market:
                if add_market_series is None:
                    factors = [read_market(close=1, start=start)] + factors
                else:
                    factors = [
                        pd.DataFrame(
                            {k: add_market_series for k in list(factors[0].columns)},
                            index=factors[0].index,
                        )
                    ] + factors
            self.__betas_rets = {
                d1[0]: d1[1] * d2 for d1, d2 in zip(self.__betas, factors)
            }
        else:
            self.__betas_rets = "您如果想计算各个股票在各个因子的收益率，请先指定betas_rets参数为True"

    @property
    def idiosyncratic(self):
        return self.__idiosyncratic

    @property
    def data(self):
        return self.__data

    @property
    def alphas(self):
        return self.__alphas

    @property
    def betas(self):
        return self.__betas

    @property
    def betas_rets(self):
        return self.__betas_rets

    @property
    def factors_rets(self):
        return self.__factors_rets

    @property
    def coefficients(self):
        return self.__coefficients

    def __call__(self):
        return self.idiosyncratic

    def make_df(self, rets, facs):
        rets = rets.stack().reset_index()
        rets.columns = ["date", "code", "ret"]
        facs = facs.reset_index()
        facs.columns = ["date"] + list(facs.columns)[1:]
        df = pd.merge(rets, facs, on=["date"])
        df = df.set_index("date")
        return df

    def ols_in(self, df):
        try:
            if self.add_market:
                x = df[["fac0"] + [f"fac{i+1}" for i in range(len(self.factors))]]
            else:
                x = df[[f"fac{i+1}" for i in range(len(self.factors))]]
            ols = po.PandasRollingOLS(
                y=df[["ret"]],
                x=x,
                window=self.backsee,
            )
            betas = ols.beta
            alpha = ols.alpha
            return pd.concat([alpha, betas], axis=1)
        except Exception:
            # 有些数据总共不足，那就跳过
            ...

__init__(factors, minus_group=3, backsee=20, rets=None, value_weighted=1, add_market=1, add_market_series=None, factors_names=None, betas_rets=0, total_cap=0)

使用fama三因子的方法，将个股的收益率，拆分出各个因子带来的收益率以及特质的收益率 分别计算每一期，各个因子收益率的值，超额收益率，因子的暴露，以及特质收益率

Parameters

factors : List[pd.DataFrame] 用于解释收益的各个因子值，每一个都是index为时间，columns为股票代码，values为因子值的dataframe minus_group : Union[list, float], optional 每一个因子将截面上的股票分为几组, by default 3 backsee : int, optional 做时序回归时，回看的天数, by default 20 rets : pd.DataFrame, optional 每只个股的收益率，index为时间，columns为股票代码，values为收益率，默认使用当日日间收益率, by default None value_weighted : bool, optional 是否使用流通市值加权, by default 1 add_market : bool, optional 是否加入市场收益率因子，默认使用中证全指的每日日间收益率, by default 1 add_market_series : bool, optional 加入的市场收益率的数据，如果没指定，则使用中证全指的日间收益率, by default None factors_names : list, optional 各个因子的名字，默认为fac0(市场收益率因子，如果没有，则从fac1开始),fac1,fac2,fac3, by default None betas_rets : bool, optional 是否计算每只个股的由于暴露在每个因子上所带来的收益率, by default 0 total_cap : bool, optional 加权时使用总市值, by default 0

Source code in pure_ocean_breeze/labor/process.py
Python
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
def __init__(
    self,
    factors: List[pd.DataFrame],
    minus_group: Union[list, float] = 3,
    backsee: int = 20,
    rets: pd.DataFrame = None,
    value_weighted: bool = 1,
    add_market: bool = 1,
    add_market_series: pd.Series = None,
    factors_names: list = None,
    betas_rets: bool = 0,
    total_cap: bool = 0,
) -> None:
    """使用fama三因子的方法，将个股的收益率，拆分出各个因子带来的收益率以及特质的收益率
    分别计算每一期，各个因子收益率的值，超额收益率，因子的暴露，以及特质收益率

    Parameters
    ----------
    factors : List[pd.DataFrame]
        用于解释收益的各个因子值，每一个都是index为时间，columns为股票代码，values为因子值的dataframe
    minus_group : Union[list, float], optional
        每一个因子将截面上的股票分为几组, by default 3
    backsee : int, optional
        做时序回归时，回看的天数, by default 20
    rets : pd.DataFrame, optional
        每只个股的收益率，index为时间，columns为股票代码，values为收益率，默认使用当日日间收益率, by default None
    value_weighted : bool, optional
        是否使用流通市值加权, by default 1
    add_market : bool, optional
        是否加入市场收益率因子，默认使用中证全指的每日日间收益率, by default 1
    add_market_series : bool, optional
        加入的市场收益率的数据，如果没指定，则使用中证全指的日间收益率, by default None
    factors_names : list, optional
        各个因子的名字，默认为fac0(市场收益率因子，如果没有，则从fac1开始),fac1,fac2,fac3, by default None
    betas_rets : bool, optional
        是否计算每只个股的由于暴露在每个因子上所带来的收益率, by default 0
    total_cap : bool, optional
        加权时使用总市值, by default 0
    """
    start = max(
        [int(datetime.datetime.strftime(i.index.min(), "%Y%m%d")) for i in factors]
    )
    self.backsee = backsee
    self.factors = factors
    self.factors_names = factors_names
    if isinstance(minus_group, int):
        minus_group = [minus_group] * len(factors)
    self.minus_group = minus_group
    if rets is None:
        closes = read_daily(close=1, start=start)
        rets = closes / closes.shift(1) - 1
    self.rets = rets
    self.factors_group = [
        to_group(i, group=j) for i, j in zip(self.factors, self.minus_group)
    ]
    self.factors_group_long = [(i == 0) + 0 for i in self.factors_group]
    self.factors_group_short = [
        (i == (j - 1)) + 0 for i, j in zip(self.factors_group, self.minus_group)
    ]
    self.value_weighted = value_weighted
    if value_weighted:
        if total_cap:
            self.cap = read_daily(total_cap=1, start=start)
        self.cap = read_daily(flow_cap=1, start=start)
        self.factors_group_long = [self.cap * i for i in self.factors_group_long]
        self.factors_group_short = [self.cap * i for i in self.factors_group_short]
        self.factors_group_long = [
            (i.T / i.T.sum()).T for i in self.factors_group_long
        ]
        self.factors_group_short = [
            (i.T / i.T.sum()).T for i in self.factors_group_short
        ]
        self.factors_rets_long = [
            (self.rets * i).sum(axis=1).to_frame(f"fac{num+1}")
            for num, i in enumerate(self.factors_group_long)
        ]
        self.factors_rets_short = [
            (self.rets * i).sum(axis=1).to_frame(f"fac{num+1}")
            for num, i in enumerate(self.factors_group_short)
        ]
    else:
        self.factors_rets_long = [
            (self.rets * i).mean(axis=1).to_frame(f"fac{num+1}")
            for num, i in enumerate(self.factors_group_long)
        ]
        self.factors_rets_short = [
            (self.rets * i).mean(axis=1).to_frame(f"fac{num+1}")
            for num, i in enumerate(self.factors_group_short)
        ]
    self.rets_long = pd.concat(self.factors_rets_long, axis=1)
    self.rets_short = pd.concat(self.factors_rets_short, axis=1)
    self.__factors_rets = self.rets_long - self.rets_short
    if add_market_series is not None:
        add_market = 1
    self.add_market = add_market
    if add_market:
        if add_market_series is None:
            closes = read_market(close=1, every_stock=0, start=start).to_frame(
                "fac0"
            )
        else:
            closes = add_market_series.to_frame("fac0")
        rets = closes / closes.shift(1) - 1
        self.__factors_rets = pd.concat([rets, self.__factors_rets], axis=1)
        if factors_names is not None:
            factors_names = ["市场"] + factors_names
    self.__data = self.make_df(self.rets, self.__factors_rets)
    tqdm.auto.tqdm.pandas()
    self.__coefficients = (
        self.__data.groupby("code").progress_apply(self.ols_in).reset_index()
    )
    self.__coefficients = self.__coefficients.rename(
        columns={
            i: "co" + i for i in list(self.__coefficients.columns) if "fac" in i
        }
    )
    self.__data = pd.merge(
        self.__data.reset_index(), self.__coefficients, on=["date", "code"]
    )
    betas = [
        self.__data[i] * self.__data["co" + i]
        for i in list(self.__data.columns)
        if i.startswith("fac")
    ]
    betas = sum(betas)
    self.__data = self.__data.assign(
        idiosyncratic=self.__data.ret - self.__data.intercept - betas
    )
    self.__idiosyncratic = self.__data.pivot(
        index="date", columns="code", values="idiosyncratic"
    )
    self.__alphas = self.__data.pivot(
        index="date", columns="code", values="intercept"
    )
    if factors_names is None:
        self.__betas = {
            i: self.__data.pivot(index="date", columns="code", values=i)
            for i in list(self.__data.columns)
            if i.startswith("fac")
        }
    else:
        facs = [i for i in list(self.__data.columns) if i.startswith("fac")]
        self.__betas = {
            factors_names[num]: self.__data.pivot(
                index="date", columns="code", values=i
            )
            for num, i in enumerate(facs)
        }
    if betas_rets:
        if add_market:
            if add_market_series is None:
                factors = [read_market(close=1, start=start)] + factors
            else:
                factors = [
                    pd.DataFrame(
                        {k: add_market_series for k in list(factors[0].columns)},
                        index=factors[0].index,
                    )
                ] + factors
        self.__betas_rets = {
            d1[0]: d1[1] * d2 for d1, d2 in zip(self.__betas, factors)
        }
    else:
        self.__betas_rets = "您如果想计算各个股票在各个因子的收益率，请先指定betas_rets参数为True"

pure_helper

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
class pure_helper(object):
    def __init__(
        self,
        df_main: pd.DataFrame,
        df_helper: pd.DataFrame,
        func: Callable = None,
        group: int = 10,
    ) -> None:
        """使用因子b的值大小，对因子a进行分组，并可以在组内进行某种操作

        Parameters
        ----------
        df_main : pd.DataFrame
            要被分组并进行操作的因子
        df_helper : pd.DataFrame
            用来做分组的依据
        func : Callable, optional
            分组后，组内要进行的操作, by default None
        group : int, optional
            要分的组数, by default 10
        """
        self.df_main = df_main
        self.df_helper = df_helper
        self.func = func
        self.group = group
        if self.func is None:
            self.__data = self.sort_a_with_b()
        else:
            self.__data = self.sort_a_with_b_func()

    @property
    def data(self):
        return self.__data

    def __call__(self) -> pd.DataFrame:
        return self.data

    def sort_a_with_b(self):
        dfb = to_group(self.df_helper, group=self.group)
        dfb = dfb.stack().reset_index()
        dfb.columns = ["date", "code", "group"]
        dfa = self.df_main.stack().reset_index()
        dfa.columns = ["date", "code", "target"]
        df = pd.merge(dfa, dfb, on=["date", "code"])
        return df

    def sort_a_with_b_func(self):
        the_func = partial(self.func)
        df = self.sort_a_with_b().drop(columns=["code"])
        df = (
            df.groupby(["date", "group"])
            .apply(the_func)
            .drop(columns=["group"])
            .reset_index()
        )
        df = df.pivot(index="date", columns="group", values="target")
        df.columns = [f"group{str(int(i+1))}" for i in list(df.columns)]
        return df

__init__(df_main, df_helper, func=None, group=10)

使用因子b的值大小，对因子a进行分组，并可以在组内进行某种操作

Parameters

df_main : pd.DataFrame 要被分组并进行操作的因子 df_helper : pd.DataFrame 用来做分组的依据 func : Callable, optional 分组后，组内要进行的操作, by default None group : int, optional 要分的组数, by default 10

Source code in pure_ocean_breeze/labor/process.py
Python
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
def __init__(
    self,
    df_main: pd.DataFrame,
    df_helper: pd.DataFrame,
    func: Callable = None,
    group: int = 10,
) -> None:
    """使用因子b的值大小，对因子a进行分组，并可以在组内进行某种操作

    Parameters
    ----------
    df_main : pd.DataFrame
        要被分组并进行操作的因子
    df_helper : pd.DataFrame
        用来做分组的依据
    func : Callable, optional
        分组后，组内要进行的操作, by default None
    group : int, optional
        要分的组数, by default 10
    """
    self.df_main = df_main
    self.df_helper = df_helper
    self.func = func
    self.group = group
    if self.func is None:
        self.__data = self.sort_a_with_b()
    else:
        self.__data = self.sort_a_with_b_func()

pure_linprog

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
class pure_linprog(object):
    def __init__(
        self,
        facs: pd.DataFrame,
        total_caps: pd.DataFrame = None,
        indu_dummys: pd.DataFrame = None,
        index_weights_hs300: pd.DataFrame = None,
        index_weights_zz500: pd.DataFrame = None,
        index_weights_zz1000: pd.DataFrame = None,
        opens: pd.DataFrame = None,
        closes: pd.DataFrame = None,
        hs300_closes: pd.DataFrame = None,
        zz500_closes: pd.DataFrame = None,
        zz1000_closes: pd.DataFrame = None,
    ) -> None:
        """线性规划求解，目标为预期收益率最大（即因子方向为负时，组合因子值最小）
        条件为，严格控制市值中性（数据：总市值的对数；含义：组合在市值上的暴露与指数在市值上的暴露相等）
        严格控制行业中性（数据：使用中信一级行业哑变量），个股偏离在1%以内，成分股权重之和在80%以上
        分别在沪深300、中证500、中证1000上优化求解

        Parameters
        ----------
        facs : pd.DataFrame
            因子值，index为时间，columns为股票代码，values为因子值
        total_caps : pd.DataFrame, optional
            总市值数据，index为时间，columns为股票代码，values为总市值, by default None
        indu_dummys : pd.DataFrame, optional
            行业哑变量，包含两列名为date的时间和code的股票代码，以及30+列行业哑变量, by default None
        index_weights_hs300 : pd.DataFrame, optional
            沪深300指数成分股权重，月频数据, by default None
        index_weights_zz500 : pd.DataFrame, optional
            中证500指数成分股权重，月频数据, by default None
        index_weights_zz1000 : pd.DataFrame, optional
            中证1000指数成分股权重，月频数据, by default None
        opens : pd.DataFrame, optional
            每月月初开盘价数据, by default None
        closes : pd.DataFrame, optional
            每月月末收盘价数据, by default None
        hs300_closes : pd.DataFrame, optional
            沪深300每月收盘价数据, by default None
        zz500_closes : pd.DataFrame, optional
            中证500每月收盘价数据,, by default None
        zz1000_closes : pd.DataFrame, optional
            中证1000每月收盘价数据,, by default None
        """
        self.facs = facs.resample("M").last()
        if total_caps is None:
            total_caps = standardlize(
                np.log(read_daily(total_cap=1).resample("M").last())
            )
        if indu_dummys is None:
            indu_dummys = read_daily(zxindustry_dummy_code=1)
        if index_weights_hs300 is None:
            index_weights_hs300 = read_daily(hs300_member_weight=1)
        if index_weights_zz500 is None:
            index_weights_zz500 = read_daily(zz500_member_weight=1)
        if index_weights_zz1000 is None:
            index_weights_zz1000 = read_daily(zz1000_member_weight=1)
        if opens is None:
            opens = read_daily(open=1).resample("M").first()
        if closes is None:
            closes = read_daily(close=1).resample("M").last()
        if hs300_closes is None:
            hs300_closes = read_index_single("000300.SH").resample("M").last()
        if zz500_closes is None:
            zz500_closes = read_index_single("000905.SH").resample("M").last()
        if zz1000_closes is None:
            zz1000_closes = read_index_single("000852.SH").resample("M").last()
        self.total_caps = total_caps
        self.indu_dummys = indu_dummys
        self.index_weights_hs300 = index_weights_hs300
        self.index_weights_zz500 = index_weights_zz500
        self.index_weights_zz1000 = index_weights_zz1000
        self.hs300_weights = []
        self.zz500_weights = []
        self.zz1000_weights = []
        self.ret_next = closes / opens - 1
        self.ret_hs300 = hs300_closes.pct_change()
        self.ret_zz500 = zz500_closes.pct_change()
        self.ret_zz1000 = zz1000_closes.pct_change()

    def optimize_one_day(
        self,
        fac: pd.DataFrame,
        flow_cap: pd.DataFrame,
        indu_dummy: pd.DataFrame,
        index_weight: pd.DataFrame,
        name: str,
    ) -> pd.DataFrame:
        """优化单期求解

        Parameters
        ----------
        fac : pd.DataFrame
            单期因子值，index为code，columns为date，values为因子值
        flow_cap : pd.DataFrame
            流通市值，index为code，columns为date，values为截面标准化的流通市值
        indu_dummy : pd.DataFrame
            行业哑变量，index为code，columns为行业代码，values为哑变量
        index_weight : pd.DataFrame
            指数成分股权重，index为code，columns为date，values为权重

        Returns
        -------
        pd.DataFrame
            当期最佳权重
        """
        if fac.shape[0] > 0 and index_weight.shape[1] > 0:
            date = fac.columns.tolist()[0]
            codes = list(
                set(fac.index)
                | set(flow_cap.index)
                | set(indu_dummy.index)
                | set(index_weight.index)
            )
            fac, flow_cap, indu_dummy, index_weight = list(
                map(
                    lambda x: x.reindex(codes).fillna(0).to_numpy(),
                    [fac, flow_cap, indu_dummy, index_weight],
                )
            )
            sign_index_weight = np.sign(index_weight)
            # 个股权重大于零、偏离1%
            bounds = list(
                zip(
                    select_max(index_weight - 0.01, 0).flatten(),
                    select_min(index_weight + 0.01, 1).flatten(),
                )
            )
            # 市值中性+行业中性+权重和为1
            huge = np.vstack([flow_cap.T, indu_dummy.T, np.array([1] * len(codes))])
            target = (
                list(flow_cap.T @ index_weight.flatten())
                + list((indu_dummy.T @ index_weight).flatten())
                + [np.sum(index_weight)]
            )
            # 写线性条件
            c = fac.T.flatten().tolist()
            a = sign_index_weight.reshape((1, -1)).tolist()
            b = [0.8]
            # 优化求解
            res = linprog(c, a, b, huge, target, bounds)
            if res.success:
                return pd.DataFrame({date: res.x.tolist()}, index=codes)
            else:
                # raise NotImplementedError(f"{date}这一期的优化失败，请检查")
                logger.warning(f"{name}在{date}这一期的优化失败，请检查")
                return None
        else:
            return None

    def optimize_many_days(self, startdate: int = STATES["START"]):
        dates = [i for i in self.facs.index if i >= pd.Timestamp(str(startdate))]
        for date in tqdm.auto.tqdm(dates):
            fac = self.facs[self.facs.index == date].T.dropna()
            total_cap = self.total_caps[self.total_caps.index == date].T.dropna()
            indu_dummy = self.indu_dummys[self.indu_dummys.date <= date]
            indu_dummy = (
                indu_dummy[indu_dummy.date == indu_dummy.date.max()]
                .drop(columns=["date"])
                .set_index("code")
            )
            index_weight_hs300 = self.index_weights_hs300[
                self.index_weights_hs300.index == date
            ].T.dropna()
            index_weight_zz500 = self.index_weights_zz500[
                self.index_weights_zz500.index == date
            ].T.dropna()
            index_weight_zz1000 = self.index_weights_zz1000[
                self.index_weights_zz1000.index == date
            ].T.dropna()
            weight_hs300 = self.optimize_one_day(
                fac, total_cap, indu_dummy, index_weight_hs300, "hs300"
            )
            weight_zz500 = self.optimize_one_day(
                fac, total_cap, indu_dummy, index_weight_zz500, "zz500"
            )
            weight_zz1000 = self.optimize_one_day(
                fac, total_cap, indu_dummy, index_weight_zz1000, "zz1000"
            )
            self.hs300_weights.append(weight_hs300)
            self.zz500_weights.append(weight_zz500)
            self.zz1000_weights.append(weight_zz1000)
        self.hs300_weights = pd.concat(self.hs300_weights, axis=1).T
        self.zz500_weights = pd.concat(self.zz500_weights, axis=1).T
        self.zz1000_weights = pd.concat(self.zz1000_weights, axis=1).T

    def make_contrast(self, weight, index, name) -> list[pd.DataFrame]:
        ret = (weight.shift(1) * self.ret_next).sum(axis=1)
        abret = ret - index
        rets = pd.concat([ret, index, abret], axis=1).dropna()
        rets.columns = [f"{name}增强组合净值", f"{name}指数净值", f"{name}增强组合超额净值"]
        rets = (rets + 1).cumprod()
        rets = rets.apply(lambda x: x / x.iloc[0])
        comments = comments_on_twins(rets[f"{name}增强组合超额净值"], abret.dropna())
        return comments, rets

    def run(self, startdate: int = STATES["START"]) -> pd.DataFrame:
        """运行规划求解

        Parameters
        ----------
        startdate : int, optional
            起始日期, by default 20130101

        Returns
        -------
        pd.DataFrame
            超额绩效指标
        """
        self.optimize_many_days(startdate=startdate)
        self.hs300_comments, self.hs300_nets = self.make_contrast(
            self.hs300_weights, self.ret_hs300, "沪深300"
        )
        self.zz500_comments, self.zz500_nets = self.make_contrast(
            self.zz500_weights, self.ret_zz500, "中证500"
        )
        self.zz1000_comments, self.zz1000_nets = self.make_contrast(
            self.zz1000_weights, self.ret_zz1000, "中证1000"
        )

        figs = cf.figures(
            pd.concat([self.hs300_nets, self.zz500_nets, self.zz1000_nets]),
            [
                dict(kind="line", y=list(self.hs300_nets.columns)),
                dict(kind="line", y=list(self.zz500_nets.columns)),
                dict(kind="line", y=list(self.zz1000_nets.columns)),
            ],
            asList=True,
        )
        base_layout = cf.tools.get_base_layout(figs)

        sp = cf.subplots(
            figs,
            shape=(1, 3),
            base_layout=base_layout,
            vertical_spacing=0.15,
            horizontal_spacing=0.03,
            shared_yaxes=False,
            subplot_titles=["沪深300增强", "中证500增强", "中证1000增强"],
        )
        sp["layout"].update(showlegend=True)
        cf.iplot(sp)

        self.comments = pd.concat(
            [self.hs300_comments, self.zz500_comments, self.zz1000_comments], axis=1
        )
        self.comments.columns = ["沪深300超额", "中证500超额", "中证1000超额"]

        from pure_ocean_breeze.state.states import COMMENTS_WRITER, NET_VALUES_WRITER

        comments_writer = COMMENTS_WRITER
        net_values_writer = NET_VALUES_WRITER
        if comments_writer is not None:
            self.hs300_comments.to_excel(comments_writer, sheet_name="沪深300组合优化超额绩效")
            self.zz500_comments.to_excel(comments_writer, sheet_name="中证500组合优化超额绩效")
            self.zz1000_comments.to_excel(comments_writer, sheet_name="中证1000组合优化超额绩效")
        if net_values_writer is not None:
            self.hs300_nets.to_excel(net_values_writer, sheet_name="沪深300组合优化净值")
            self.zz500_nets.to_excel(net_values_writer, sheet_name="中证500组合优化净值")
            self.zz1000_nets.to_excel(net_values_writer, sheet_name="中证1000组合优化净值")

        return self.comments.T

__init__(facs, total_caps=None, indu_dummys=None, index_weights_hs300=None, index_weights_zz500=None, index_weights_zz1000=None, opens=None, closes=None, hs300_closes=None, zz500_closes=None, zz1000_closes=None)

线性规划求解，目标为预期收益率最大（即因子方向为负时，组合因子值最小） 条件为，严格控制市值中性（数据：总市值的对数；含义：组合在市值上的暴露与指数在市值上的暴露相等） 严格控制行业中性（数据：使用中信一级行业哑变量），个股偏离在1%以内，成分股权重之和在80%以上 分别在沪深300、中证500、中证1000上优化求解

Parameters

facs : pd.DataFrame 因子值，index为时间，columns为股票代码，values为因子值 total_caps : pd.DataFrame, optional 总市值数据，index为时间，columns为股票代码，values为总市值, by default None indu_dummys : pd.DataFrame, optional 行业哑变量，包含两列名为date的时间和code的股票代码，以及30+列行业哑变量, by default None index_weights_hs300 : pd.DataFrame, optional 沪深300指数成分股权重，月频数据, by default None index_weights_zz500 : pd.DataFrame, optional 中证500指数成分股权重，月频数据, by default None index_weights_zz1000 : pd.DataFrame, optional 中证1000指数成分股权重，月频数据, by default None opens : pd.DataFrame, optional 每月月初开盘价数据, by default None closes : pd.DataFrame, optional 每月月末收盘价数据, by default None hs300_closes : pd.DataFrame, optional 沪深300每月收盘价数据, by default None zz500_closes : pd.DataFrame, optional 中证500每月收盘价数据,, by default None zz1000_closes : pd.DataFrame, optional 中证1000每月收盘价数据,, by default None

Source code in pure_ocean_breeze/labor/process.py
Python
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
def __init__(
    self,
    facs: pd.DataFrame,
    total_caps: pd.DataFrame = None,
    indu_dummys: pd.DataFrame = None,
    index_weights_hs300: pd.DataFrame = None,
    index_weights_zz500: pd.DataFrame = None,
    index_weights_zz1000: pd.DataFrame = None,
    opens: pd.DataFrame = None,
    closes: pd.DataFrame = None,
    hs300_closes: pd.DataFrame = None,
    zz500_closes: pd.DataFrame = None,
    zz1000_closes: pd.DataFrame = None,
) -> None:
    """线性规划求解，目标为预期收益率最大（即因子方向为负时，组合因子值最小）
    条件为，严格控制市值中性（数据：总市值的对数；含义：组合在市值上的暴露与指数在市值上的暴露相等）
    严格控制行业中性（数据：使用中信一级行业哑变量），个股偏离在1%以内，成分股权重之和在80%以上
    分别在沪深300、中证500、中证1000上优化求解

    Parameters
    ----------
    facs : pd.DataFrame
        因子值，index为时间，columns为股票代码，values为因子值
    total_caps : pd.DataFrame, optional
        总市值数据，index为时间，columns为股票代码，values为总市值, by default None
    indu_dummys : pd.DataFrame, optional
        行业哑变量，包含两列名为date的时间和code的股票代码，以及30+列行业哑变量, by default None
    index_weights_hs300 : pd.DataFrame, optional
        沪深300指数成分股权重，月频数据, by default None
    index_weights_zz500 : pd.DataFrame, optional
        中证500指数成分股权重，月频数据, by default None
    index_weights_zz1000 : pd.DataFrame, optional
        中证1000指数成分股权重，月频数据, by default None
    opens : pd.DataFrame, optional
        每月月初开盘价数据, by default None
    closes : pd.DataFrame, optional
        每月月末收盘价数据, by default None
    hs300_closes : pd.DataFrame, optional
        沪深300每月收盘价数据, by default None
    zz500_closes : pd.DataFrame, optional
        中证500每月收盘价数据,, by default None
    zz1000_closes : pd.DataFrame, optional
        中证1000每月收盘价数据,, by default None
    """
    self.facs = facs.resample("M").last()
    if total_caps is None:
        total_caps = standardlize(
            np.log(read_daily(total_cap=1).resample("M").last())
        )
    if indu_dummys is None:
        indu_dummys = read_daily(zxindustry_dummy_code=1)
    if index_weights_hs300 is None:
        index_weights_hs300 = read_daily(hs300_member_weight=1)
    if index_weights_zz500 is None:
        index_weights_zz500 = read_daily(zz500_member_weight=1)
    if index_weights_zz1000 is None:
        index_weights_zz1000 = read_daily(zz1000_member_weight=1)
    if opens is None:
        opens = read_daily(open=1).resample("M").first()
    if closes is None:
        closes = read_daily(close=1).resample("M").last()
    if hs300_closes is None:
        hs300_closes = read_index_single("000300.SH").resample("M").last()
    if zz500_closes is None:
        zz500_closes = read_index_single("000905.SH").resample("M").last()
    if zz1000_closes is None:
        zz1000_closes = read_index_single("000852.SH").resample("M").last()
    self.total_caps = total_caps
    self.indu_dummys = indu_dummys
    self.index_weights_hs300 = index_weights_hs300
    self.index_weights_zz500 = index_weights_zz500
    self.index_weights_zz1000 = index_weights_zz1000
    self.hs300_weights = []
    self.zz500_weights = []
    self.zz1000_weights = []
    self.ret_next = closes / opens - 1
    self.ret_hs300 = hs300_closes.pct_change()
    self.ret_zz500 = zz500_closes.pct_change()
    self.ret_zz1000 = zz1000_closes.pct_change()

optimize_one_day(fac, flow_cap, indu_dummy, index_weight, name)

优化单期求解

Parameters

fac : pd.DataFrame 单期因子值，index为code，columns为date，values为因子值 flow_cap : pd.DataFrame 流通市值，index为code，columns为date，values为截面标准化的流通市值 indu_dummy : pd.DataFrame 行业哑变量，index为code，columns为行业代码，values为哑变量 index_weight : pd.DataFrame 指数成分股权重，index为code，columns为date，values为权重

Returns

pd.DataFrame 当期最佳权重

Source code in pure_ocean_breeze/labor/process.py
Python
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
def optimize_one_day(
    self,
    fac: pd.DataFrame,
    flow_cap: pd.DataFrame,
    indu_dummy: pd.DataFrame,
    index_weight: pd.DataFrame,
    name: str,
) -> pd.DataFrame:
    """优化单期求解

    Parameters
    ----------
    fac : pd.DataFrame
        单期因子值，index为code，columns为date，values为因子值
    flow_cap : pd.DataFrame
        流通市值，index为code，columns为date，values为截面标准化的流通市值
    indu_dummy : pd.DataFrame
        行业哑变量，index为code，columns为行业代码，values为哑变量
    index_weight : pd.DataFrame
        指数成分股权重，index为code，columns为date，values为权重

    Returns
    -------
    pd.DataFrame
        当期最佳权重
    """
    if fac.shape[0] > 0 and index_weight.shape[1] > 0:
        date = fac.columns.tolist()[0]
        codes = list(
            set(fac.index)
            | set(flow_cap.index)
            | set(indu_dummy.index)
            | set(index_weight.index)
        )
        fac, flow_cap, indu_dummy, index_weight = list(
            map(
                lambda x: x.reindex(codes).fillna(0).to_numpy(),
                [fac, flow_cap, indu_dummy, index_weight],
            )
        )
        sign_index_weight = np.sign(index_weight)
        # 个股权重大于零、偏离1%
        bounds = list(
            zip(
                select_max(index_weight - 0.01, 0).flatten(),
                select_min(index_weight + 0.01, 1).flatten(),
            )
        )
        # 市值中性+行业中性+权重和为1
        huge = np.vstack([flow_cap.T, indu_dummy.T, np.array([1] * len(codes))])
        target = (
            list(flow_cap.T @ index_weight.flatten())
            + list((indu_dummy.T @ index_weight).flatten())
            + [np.sum(index_weight)]
        )
        # 写线性条件
        c = fac.T.flatten().tolist()
        a = sign_index_weight.reshape((1, -1)).tolist()
        b = [0.8]
        # 优化求解
        res = linprog(c, a, b, huge, target, bounds)
        if res.success:
            return pd.DataFrame({date: res.x.tolist()}, index=codes)
        else:
            # raise NotImplementedError(f"{date}这一期的优化失败，请检查")
            logger.warning(f"{name}在{date}这一期的优化失败，请检查")
            return None
    else:
        return None

run(startdate=STATES['START'])

运行规划求解

Parameters

startdate : int, optional 起始日期, by default 20130101

Returns

pd.DataFrame 超额绩效指标

Source code in pure_ocean_breeze/labor/process.py
Python
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
def run(self, startdate: int = STATES["START"]) -> pd.DataFrame:
    """运行规划求解

    Parameters
    ----------
    startdate : int, optional
        起始日期, by default 20130101

    Returns
    -------
    pd.DataFrame
        超额绩效指标
    """
    self.optimize_many_days(startdate=startdate)
    self.hs300_comments, self.hs300_nets = self.make_contrast(
        self.hs300_weights, self.ret_hs300, "沪深300"
    )
    self.zz500_comments, self.zz500_nets = self.make_contrast(
        self.zz500_weights, self.ret_zz500, "中证500"
    )
    self.zz1000_comments, self.zz1000_nets = self.make_contrast(
        self.zz1000_weights, self.ret_zz1000, "中证1000"
    )

    figs = cf.figures(
        pd.concat([self.hs300_nets, self.zz500_nets, self.zz1000_nets]),
        [
            dict(kind="line", y=list(self.hs300_nets.columns)),
            dict(kind="line", y=list(self.zz500_nets.columns)),
            dict(kind="line", y=list(self.zz1000_nets.columns)),
        ],
        asList=True,
    )
    base_layout = cf.tools.get_base_layout(figs)

    sp = cf.subplots(
        figs,
        shape=(1, 3),
        base_layout=base_layout,
        vertical_spacing=0.15,
        horizontal_spacing=0.03,
        shared_yaxes=False,
        subplot_titles=["沪深300增强", "中证500增强", "中证1000增强"],
    )
    sp["layout"].update(showlegend=True)
    cf.iplot(sp)

    self.comments = pd.concat(
        [self.hs300_comments, self.zz500_comments, self.zz1000_comments], axis=1
    )
    self.comments.columns = ["沪深300超额", "中证500超额", "中证1000超额"]

    from pure_ocean_breeze.state.states import COMMENTS_WRITER, NET_VALUES_WRITER

    comments_writer = COMMENTS_WRITER
    net_values_writer = NET_VALUES_WRITER
    if comments_writer is not None:
        self.hs300_comments.to_excel(comments_writer, sheet_name="沪深300组合优化超额绩效")
        self.zz500_comments.to_excel(comments_writer, sheet_name="中证500组合优化超额绩效")
        self.zz1000_comments.to_excel(comments_writer, sheet_name="中证1000组合优化超额绩效")
    if net_values_writer is not None:
        self.hs300_nets.to_excel(net_values_writer, sheet_name="沪深300组合优化净值")
        self.zz500_nets.to_excel(net_values_writer, sheet_name="中证500组合优化净值")
        self.zz1000_nets.to_excel(net_values_writer, sheet_name="中证1000组合优化净值")

    return self.comments.T

pure_moon

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
class pure_moon(object):
    __slots__ = [
        "homeplace",
        "sts_monthly_file",
        "states_monthly_file",
        "factors",
        "codes",
        "tradedays",
        "ages",
        "amounts",
        "closes",
        "opens",
        "capital",
        "states",
        "sts",
        "turnovers",
        "sts_monthly",
        "states_monthly",
        "ages_monthly",
        "tris_monthly",
        "opens_monthly",
        "closes_monthly",
        "rets_monthly",
        "opens_monthly_shift",
        "rets_monthly_begin",
        "limit_ups",
        "limit_downs",
        "data",
        "ic_icir_and_rank",
        "rets_monthly_limit_downs",
        "group_rets",
        "long_short_rets",
        "long_short_net_values",
        "group_net_values",
        "long_short_ret_yearly",
        "long_short_vol_yearly",
        "long_short_info_ratio",
        "long_short_win_times",
        "long_short_win_ratio",
        "retreats",
        "max_retreat",
        "long_short_comments",
        "total_comments",
        "square_rets",
        "cap",
        "cap_value",
        "industry_dummy",
        "industry_codes",
        "industry_codes_str",
        "industry_ws",
        "__factors_out",
        "ics",
        "rankics",
        "factor_turnover_rates",
        "factor_turnover_rate",
        "group_rets_std",
        "group_rets_stds",
        "group_rets_skews",
        "group_rets_skew",
        "wind_out",
        "swindustry_dummy",
        "zxindustry_dummy",
        "closes2_monthly",
        "rets_monthly_last",
        "freq_ctrl",
        "freq",
        "factor_cover",
        "factor_cross_skew",
        "factor_cross_skew_after_neu",
        "pos_neg_rate",
        "corr_itself",
        "factor_cross_stds",
        "corr_itself_shift2",
        "rets_all",
        "inner_long_ret_yearly",
        "inner_short_ret_yearly",
        "inner_long_net_values",
        "inner_short_net_values",
        "group_mean_rets_monthly"
    ]

    @classmethod
    @lru_cache(maxsize=None)
    def __init__(
        cls,
        freq: str = "M",
        no_read_indu: bool = 0,
        swindustry_dummy: pd.DataFrame = None,
        zxindustry_dummy: pd.DataFrame = None,
        read_in_swindustry_dummy: bool = 0,
    ):
        cls.homeplace = HomePlace()
        cls.freq = freq
        cls.freq_ctrl = frequency_controller(freq)
        # 已经算好的月度st状态文件
        # week_here
        cls.sts_monthly_file = cls.freq_ctrl.sts_files
        # 已经算好的月度交易状态文件
        # week_here
        cls.states_monthly_file = cls.freq_ctrl.states_files

        if swindustry_dummy is not None:
            cls.swindustry_dummy = swindustry_dummy
        if zxindustry_dummy is not None:
            cls.zxindustry_dummy = zxindustry_dummy

        def deal_dummy(industry_dummy):
            industry_dummy = industry_dummy.drop(columns=["code"]).reset_index()
            industry_ws = [f"w{i}" for i in range(1, industry_dummy.shape[1] - 1)]
            col = ["code", "date"] + industry_ws
            industry_dummy.columns = col
            industry_dummy = industry_dummy[
                industry_dummy.date >= pd.Timestamp("20100101")
            ]
            return industry_dummy

        if (swindustry_dummy is None) and (zxindustry_dummy is None):
            if not no_read_indu:
                if read_in_swindustry_dummy:
                    # week_here
                    cls.swindustry_dummy = (
                        pd.read_parquet(
                            cls.homeplace.daily_data_file + "申万行业2021版哑变量.parquet"
                        )
                        .fillna(0)
                        .set_index("date")
                        .groupby("code")
                        .resample(freq)
                        .last()
                    )
                    cls.swindustry_dummy = deal_dummy(cls.swindustry_dummy)
                # week_here
                cls.zxindustry_dummy = (
                    pd.read_parquet(
                        cls.homeplace.daily_data_file + "中信一级行业哑变量代码版.parquet"
                    )
                    .fillna(0)
                    .set_index("date")
                    .groupby("code")
                    .resample(freq)
                    .last()
                    .fillna(0)
                )

                cls.zxindustry_dummy = deal_dummy(cls.zxindustry_dummy)

    @property
    def factors_out(self):
        return self.__factors_out

    def __call__(self):
        """调用对象则返回因子值"""
        return self.factors_out

    @classmethod
    @lru_cache(maxsize=None)
    def set_basic_data(
        cls,
        ages: pd.DataFrame = None,
        sts: pd.DataFrame = None,
        states: pd.DataFrame = None,
        opens: pd.DataFrame = None,
        closes: pd.DataFrame = None,
        capitals: pd.DataFrame = None,
        opens_average_first_day: bool = 0,
        total_cap: bool = 0,
    ):
        if ages is None:
            ages = read_daily(age=1, start=20100101)
        if sts is None:
            sts = read_daily(st=1, start=20100101)
        if states is None:
            states = read_daily(state=1, start=20100101)
        if opens is None:
            if opens_average_first_day:
                opens = read_daily(vwap=1, start=20100101)
            else:
                opens = read_daily(open=1, start=20100101)
        if closes is None:
            closes = read_daily(close=1, start=20100101)
        if capitals is None:
            if total_cap:
                capitals = (
                    read_daily(total_cap=1, start=20100101).resample(cls.freq).last()
                )
            else:
                capitals = (
                    read_daily(flow_cap=1, start=20100101).resample(cls.freq).last()
                )
        # 上市天数文件
        cls.ages = ages
        # st日子标志文件
        cls.sts = sts.fillna(0)
        # cls.sts = 1 - cls.sts.fillna(0)
        # 交易状态文件
        cls.states = states
        # 复权开盘价数据文件
        cls.opens = opens
        # 复权收盘价数据文件
        cls.closes = closes
        # 月底流通市值数据
        cls.capital = capitals
        if cls.opens is not None:
            cls.opens = cls.opens.replace(0, np.nan)
        if cls.closes is not None:
            cls.closes = cls.closes.replace(0, np.nan)

    def set_factor_df_date_as_index(self, df: pd.DataFrame):
        """设置因子数据的dataframe，因子表列名应为股票代码，索引应为时间"""
        # week_here
        self.factors = df.resample(self.freq).last().dropna(how="all")
        self.factor_cover = np.sign(self.factors.abs() + 1).sum().sum()
        opens = self.opens[self.opens.index >= self.factors.index.min()]
        total = np.sign(opens.resample(self.freq).last()).sum().sum()
        self.factor_cover = min(self.factor_cover / total, 1)
        self.factor_cross_skew = self.factors.skew(axis=1).mean()
        pos_num = ((self.factors > 0) + 0).sum().sum()
        neg_num = ((self.factors < 0) + 0).sum().sum()
        self.pos_neg_rate = pos_num / (neg_num + pos_num)
        self.corr_itself = show_corr(self.factors, self.factors.shift(1), plt_plot=0)
        self.corr_itself_shift2 = show_corr(
            self.factors, self.factors.shift(2), plt_plot=0
        )
        self.factor_cross_stds = self.factors.std(axis=1)

    @classmethod
    def judge_month_st(cls, df):
        """比较一个月内st的天数，如果st天数多，就删除本月，如果正常多，就保留本月"""
        st_count = len(df[df == 1])
        normal_count = len(df[df != 1])
        if st_count >= normal_count:
            return 0
        else:
            return 1

    @classmethod
    def judge_month_state(cls, df):
        """比较一个月内非正常交易的天数，如果非正常交易天数多，就删除本月，否则保留本月"""
        abnormal_count = len(df[df == 0])
        normal_count = len(df[df == 1])
        if abnormal_count >= normal_count:
            return 0
        else:
            return 1

    @classmethod
    def read_add(cls, pridf, df, func):
        """由于数据更新，过去计算的月度状态可能需要追加"""
        if pridf.index.max() > df.index.max():
            df_add = pridf[pridf.index > df.index.max()]
            if df_add.shape[0] > int(cls.freq_ctrl.days_in / 2):
                df_1 = df_add.index.max()
                year = df_1.year
                month = df_1.month
                last = tt.date.get_close(year=year, m=month).pd_date()
                if (last == df_1)[0]:
                    # week_here
                    df_add = df_add.resample(cls.freq).apply(func)
                    df = pd.concat([df, df_add])
                    return df
                else:
                    df_add = df_add[
                        df_add.index < pd.Timestamp(year=year, month=month, day=1)
                    ]
                    if df_add.shape[0] > 0:
                        df_add = df_add.resample(cls.freq).apply(func)
                        df = pd.concat([df, df_add])
                        return df
                    else:
                        return df
            else:
                return df
        else:
            return df

    @classmethod
    def daily_to_monthly(cls, pridf, path, func):
        """把日度的交易状态、st、上市天数，转化为月度的，并生成能否交易的判断
        读取本地已经算好的文件，并追加新的时间段部分，如果本地没有就直接全部重新算"""
        try:
            month_df = pd.read_parquet(path)
            month_df = cls.read_add(pridf, month_df, func)
            month_df.to_parquet(path)
        except Exception as e:
            if not STATES["NO_LOG"]:
                logger.error("error occurs when read state files")
                logger.error(e)
            print("state file rewriting……")
            # week_here
            df_1 = pridf.index.max()
            year = df_1.year
            month = df_1.month
            last = tt.date.get_close(year=year, m=month).pd_date()
            if not (last == df_1)[0]:
                pridf = pridf[pridf.index < pd.Timestamp(year=year, month=month, day=1)]
            month_df = pridf.resample(cls.freq).apply(func)
            month_df.to_parquet(path)
        return month_df

    @classmethod
    @lru_cache(maxsize=None)
    def judge_month(cls):
        """生成一个月综合判断的表格"""
        if cls.freq == "M":
            cls.sts_monthly = cls.daily_to_monthly(
                cls.sts, cls.sts_monthly_file, cls.judge_month_st
            )
            cls.states_monthly = cls.daily_to_monthly(
                cls.states, cls.states_monthly_file, cls.judge_month_state
            )
            # week_here
            cls.ages_monthly = (cls.ages.resample(cls.freq).last() > 60) + 0
            cls.tris_monthly = cls.sts_monthly * cls.states_monthly * cls.ages_monthly
            cls.tris_monthly = cls.tris_monthly.replace(0, np.nan)
        else:
            cls.tris_monthly = (
                (1 - cls.sts).resample(cls.freq).last().ffill(limit=2)
                * cls.states.resample(cls.freq).last().ffill(limit=2)
                * ((cls.ages.resample(cls.freq).last() > 60) + 0)
            ).replace(0, np.nan)

    @classmethod
    @lru_cache(maxsize=None)
    def get_rets_month(cls):
        """计算每月的收益率，并根据每月做出交易状态，做出删减"""
        # week_here
        cls.opens_monthly = cls.opens.resample(cls.freq).first()
        # week_here
        cls.closes_monthly = cls.closes.resample(cls.freq).last()
        cls.rets_monthly = (cls.closes_monthly - cls.opens_monthly) / cls.opens_monthly
        cls.rets_monthly = cls.rets_monthly * cls.tris_monthly
        cls.rets_monthly = cls.rets_monthly.stack().reset_index()
        cls.rets_monthly.columns = ["date", "code", "ret"]

    @classmethod
    def neutralize_factors(cls, df):
        """组内对因子进行市值中性化"""
        industry_codes = list(df.columns)
        industry_codes = [i for i in industry_codes if i.startswith("w")]
        industry_codes_str = "+".join(industry_codes)
        if len(industry_codes_str) > 0:
            ols_result = smf.ols("fac~cap_size+" + industry_codes_str, data=df).fit()
        else:
            ols_result = smf.ols("fac~cap_size", data=df).fit()
        ols_w = ols_result.params["cap_size"]
        ols_b = ols_result.params["Intercept"]
        ols_bs = {}
        for ind in industry_codes:
            ols_bs[ind] = ols_result.params[ind]
        df.fac = df.fac - ols_w * df.cap_size - ols_b
        for k, v in ols_bs.items():
            df.fac = df.fac - v * df[k]
        df = df[["fac"]]
        return df

    @classmethod
    @lru_cache(maxsize=None)
    def get_log_cap(cls, boxcox=True):
        """获得对数市值"""
        cls.cap = cls.capital.stack().reset_index()
        cls.cap.columns = ["date", "code", "cap_size"]
        if boxcox:

            def single(x):
                x.cap_size = ss.boxcox(x.cap_size)[0]
                return x

            cls.cap = cls.cap.groupby(["date"]).apply(single)
        else:
            cls.cap["cap_size"] = np.log(cls.cap["cap_size"])

    def get_neutral_factors(
        self, zxindustry_dummies=0, swindustry_dummies=0, only_cap=0
    ):
        """对因子进行行业市值中性化"""
        # week_here
        self.factors.index = self.factors.index + self.freq_ctrl.time_shift
        # week_here
        self.factors = self.factors.resample(self.freq).last()
        # week_here
        last_date = self.freq_ctrl.next_end(self.tris_monthly.index.max())
        add_tail = pd.DataFrame(1, index=[last_date], columns=self.tris_monthly.columns)
        tris_monthly = pd.concat([self.tris_monthly, add_tail])
        self.factors = self.factors * tris_monthly
        # week_here
        self.factors.index = self.factors.index - self.freq_ctrl.time_shift
        # week_here
        self.factors = self.factors.resample(self.freq).last()
        self.factors = self.factors.stack().reset_index()
        self.factors.columns = ["date", "code", "fac"]
        self.factors = pd.merge(
            self.factors, self.cap, how="inner", on=["date", "code"]
        )
        if not only_cap:
            if swindustry_dummies:
                self.factors = pd.merge(
                    self.factors, self.swindustry_dummy, on=["date", "code"]
                )
            else:
                self.factors = pd.merge(
                    self.factors, self.zxindustry_dummy, on=["date", "code"]
                )
        self.factors = self.factors.set_index(["date", "code"])
        self.factors = self.factors.groupby(["date"]).apply(self.neutralize_factors)
        self.factors = self.factors.reset_index()

    def deal_with_factors(self):
        """删除不符合交易条件的因子数据"""
        self.__factors_out = self.factors.copy()
        # week_here
        self.factors.index = self.factors.index + self.freq_ctrl.time_shift
        # week_here
        self.factors = self.factors.resample(self.freq).last()
        self.factors = self.factors * self.tris_monthly
        self.factors = self.factors.stack().reset_index()
        self.factors.columns = ["date", "code", "fac"]

    def deal_with_factors_after_neutralize(self):
        """中性化之后的因子处理方法"""
        self.factors = self.factors.set_index(["date", "code"])
        self.factors = self.factors.unstack()
        self.__factors_out = self.factors.copy()
        self.__factors_out.columns = [i[1] for i in list(self.__factors_out.columns)]
        # week_here
        self.factors.index = self.factors.index + self.freq_ctrl.time_shift
        # week_here
        self.factors = self.factors.resample(self.freq).last()
        self.factors.columns = list(map(lambda x: x[1], list(self.factors.columns)))
        self.factors = self.factors.stack().reset_index()
        self.factors.columns = ["date", "code", "fac"]

    @classmethod
    def find_limit(cls, df, up=1):
        """计算涨跌幅超过9.8%的股票，并将其存储进一个长列表里
        其中时间列，为某月的最后一天；涨停日虽然为下月初第一天，但这里标注的时间统一为上月最后一天"""
        limit_df = np.sign(df.applymap(lambda x: x - up * 0.098)).replace(
            -1 * up, np.nan
        )
        limit_df = limit_df.stack().reset_index()
        limit_df.columns = ["date", "code", "limit_up_signal"]
        limit_df = limit_df[["date", "code"]]
        return limit_df

    @classmethod
    @lru_cache(maxsize=None)
    def get_limit_ups_downs(cls):
        """找月初第一天就涨停"""
        """或者是月末跌停的股票"""
        cls.opens_monthly_shift = cls.opens_monthly.copy()
        cls.opens_monthly_shift = cls.opens_monthly_shift.shift(-1)
        cls.rets_monthly_begin = (
            cls.opens_monthly_shift - cls.closes_monthly
        ) / cls.closes_monthly
        # week_here
        cls.closes2_monthly = cls.closes.shift(1).resample(cls.freq).last()
        cls.rets_monthly_last = (
            cls.closes_monthly - cls.closes2_monthly
        ) / cls.closes2_monthly
        cls.limit_ups = cls.find_limit(cls.rets_monthly_begin, up=1)
        cls.limit_downs = cls.find_limit(cls.rets_monthly_last, up=-1)

    def get_ic_rankic(cls, df):
        """计算IC和RankIC"""
        df1 = df[["ret", "fac"]]
        ic = df1.corr(method="pearson").iloc[0, 1]
        rankic = df1.rank().corr().iloc[0, 1]
        df2 = pd.DataFrame({"ic": [ic], "rankic": [rankic]})
        return df2

    def get_icir_rankicir(cls, df):
        """计算ICIR和RankICIR"""
        ic = df.ic.mean()
        rankic = df.rankic.mean()
        # week_here
        icir = ic / np.std(df.ic) * (cls.freq_ctrl.counts_one_year ** (0.5))
        # week_here
        rankicir = rankic / np.std(df.rankic) * (cls.freq_ctrl.counts_one_year ** (0.5))
        return pd.DataFrame(
            {"IC": [ic], "ICIR": [icir], "RankIC": [rankic], "RankICIR": [rankicir]},
            index=["评价指标"],
        )

    def get_ic_icir_and_rank(cls, df):
        """计算IC、ICIR、RankIC、RankICIR"""
        df1 = df.groupby("date").apply(cls.get_ic_rankic)
        cls.ics = df1.ic
        cls.rankics = df1.rankic
        cls.ics = cls.ics.reset_index(drop=True, level=1).to_frame()
        cls.rankics = cls.rankics.reset_index(drop=True, level=1).to_frame()
        df2 = cls.get_icir_rankicir(df1)
        df2 = df2.T
        dura = (df.date.max() - df.date.min()).days / 365
        t_value = df2.iloc[3, 0] * (dura ** (1 / 2))
        df3 = pd.DataFrame({"评价指标": [t_value]}, index=["RankIC.t"])
        df4 = pd.concat([df2, df3])
        return df4

    @classmethod
    def get_groups(cls, df, groups_num):
        """依据因子值，判断是在第几组"""
        if "group" in list(df.columns):
            df = df.drop(columns=["group"])
        df = df.sort_values(["fac"], ascending=True)
        each_group = round(df.shape[0] / groups_num)
        l = list(
            map(
                lambda x, y: [x] * y,
                list(range(1, groups_num + 1)),
                [each_group] * groups_num,
            )
        )
        l = reduce(lambda x, y: x + y, l)
        if len(l) < df.shape[0]:
            l = l + [groups_num] * (df.shape[0] - len(l))
        l = l[: df.shape[0]]
        df.insert(0, "group", l)
        return df

    @classmethod
    def limit_old_to_new(cls, limit, data):
        """获取跌停股在旧月的组号，然后将日期调整到新月里
        涨停股则获得新月里涨停股的代码和时间，然后直接删去"""
        data1 = data.copy()
        data1 = data1.reset_index()
        data1.columns = ["data_index"] + list(data1.columns)[1:]
        old = pd.merge(limit, data1, how="inner", on=["date", "code"])
        old = old.set_index("data_index")
        old = old[["group", "date", "code"]]
        # week_here
        old.date = list(map(cls.freq_ctrl.next_end, list(old.date)))
        return old

    def get_data(self, groups_num):
        """拼接因子数据和每月收益率数据，并对涨停和跌停股加以处理"""
        self.data = pd.merge(
            self.rets_monthly, self.factors, how="inner", on=["date", "code"]
        )
        self.ic_icir_and_rank = self.get_ic_icir_and_rank(self.data)
        self.data = self.data.groupby("date").apply(
            lambda x: self.get_groups(x, groups_num)
        )
        self.wind_out = self.data.copy()
        self.factor_turnover_rates = self.data.pivot(
            index="date", columns="code", values="group"
        )
        rates = []
        for i in range(1, groups_num + 1):
            son = (self.factor_turnover_rates == i) + 0
            son1 = son.diff()
            # self.factor_turnover_rates = self.factor_turnover_rates.diff()
            change = ((np.abs(np.sign(son1)) == 1) + 0).sum(axis=1)
            still = (((son1 == 0) + 0) * son).sum(axis=1)
            rate = change / (change + still)
            rates.append(rate.to_frame(f"group{i}"))
        rates = pd.concat(rates, axis=1).fillna(0)
        self.factor_turnover_rates = rates
        self.data = self.data.reset_index(drop=True)
        limit_ups_object = self.limit_old_to_new(self.limit_ups, self.data)
        limit_downs_object = self.limit_old_to_new(self.limit_downs, self.data)
        self.data = self.data.drop(limit_ups_object.index)
        rets_monthly_limit_downs = pd.merge(
            self.rets_monthly, limit_downs_object, how="inner", on=["date", "code"]
        )
        self.data = pd.concat([self.data, rets_monthly_limit_downs])

    def make_start_to_one(self, l):
        """让净值序列的第一个数变成1"""
        min_date = self.factors.date.min()
        add_date = min_date - relativedelta(days=min_date.day)
        add_l = pd.Series([1], index=[add_date])
        l = pd.concat([add_l, l])
        return l

    def to_group_ret(self, l):
        """每一组的年化收益率"""
        # week_here
        ret = l[-1] ** (self.freq_ctrl.counts_one_year / len(l)) - 1
        return ret

    def get_group_rets_net_values(
        self, groups_num=10, value_weighted=False, trade_cost_double_side=0
    ):
        """计算组内每一期的平均收益，生成每日收益率序列和净值序列"""
        if value_weighted:
            cap_value = self.capital.copy()
            # week_here
            cap_value = cap_value.resample(self.freq).last().shift(1)
            cap_value = cap_value * self.tris_monthly
            # cap_value=np.log(cap_value)
            cap_value = cap_value.stack().reset_index()
            cap_value.columns = ["date", "code", "cap_value"]
            self.data = pd.merge(self.data, cap_value, on=["date", "code"])

            def in_g(df):
                df.cap_value = df.cap_value / df.cap_value.sum()
                df.ret = df.ret * df.cap_value
                return df.ret.sum()

            self.group_rets = self.data.groupby(["date", "group"]).apply(in_g)
            self.rets_all = self.data.groupby(["date"]).apply(in_g)
            self.group_rets_std = "市值加权暂未设置该功能，敬请期待🌙"
        else:
            self.group_rets = self.data.groupby(["date", "group"]).apply(
                lambda x: x.ret.mean()
            )
            self.rets_all = self.data.groupby(["date"]).apply(lambda x: x.ret.mean())
            self.group_rets_stds = self.data.groupby(["date", "group"]).ret.std()
            self.group_rets_std = (
                self.group_rets_stds.reset_index().groupby("group").mean()
            )
            self.group_rets_skews = self.data.groupby(["date", "group"]).ret.skew()
            self.group_rets_skew = (
                self.group_rets_skews.reset_index().groupby("group").mean()
            )
        # dropna是因为如果股票行情数据比因子数据的截止日期晚，而最后一个月发生月初跌停时，会造成最后某组多出一个月的数据
        self.group_rets = self.group_rets.unstack()
        self.group_rets = self.group_rets[
            self.group_rets.index <= self.factors.date.max()
        ]
        self.group_rets.columns = list(map(str, list(self.group_rets.columns)))
        self.group_rets = self.group_rets.add_prefix("group")
        self.group_rets = (
            self.group_rets - self.factor_turnover_rates * trade_cost_double_side
        )
        self.rets_all = (
            self.rets_all
            - self.factor_turnover_rates.mean(axis=1) * trade_cost_double_side
        ).dropna()
        self.long_short_rets = (
            self.group_rets["group1"] - self.group_rets["group" + str(groups_num)]
        )
        self.inner_rets_long = self.group_rets.group1 - self.rets_all
        self.inner_rets_short = (
            self.rets_all - self.group_rets["group" + str(groups_num)]
        )
        self.long_short_net_values = self.make_start_to_one(
            (self.long_short_rets + 1).cumprod()
        )
        if self.long_short_net_values[-1] <= self.long_short_net_values[0]:
            self.long_short_rets = (
                self.group_rets["group" + str(groups_num)] - self.group_rets["group1"]
            )
            self.long_short_net_values = self.make_start_to_one(
                (self.long_short_rets + 1).cumprod()
            )
            self.inner_rets_long = (
                self.group_rets["group" + str(groups_num)] - self.rets_all
            )
            self.inner_rets_short = self.rets_all - self.group_rets.group1
        self.inner_long_net_values = self.make_start_to_one(
            (self.inner_rets_long + 1).cumprod()
        )
        self.inner_short_net_values = self.make_start_to_one(
            (self.inner_rets_short + 1).cumprod()
        )
        self.group_rets = self.group_rets.assign(long_short=self.long_short_rets)
        self.group_net_values = self.group_rets.applymap(lambda x: x + 1)
        self.group_net_values = self.group_net_values.cumprod()
        self.group_net_values = self.group_net_values.apply(self.make_start_to_one)
        a = groups_num ** (0.5)
        # 判断是否要两个因子画表格
        if a == int(a):
            self.square_rets = (
                self.group_net_values.iloc[:, :-1].apply(self.to_group_ret).to_numpy()
            )
            self.square_rets = self.square_rets.reshape((int(a), int(a)))
            self.square_rets = pd.DataFrame(
                self.square_rets,
                columns=list(range(1, int(a) + 1)),
                index=list(range(1, int(a) + 1)),
            )
            print("这是self.square_rets", self.square_rets)

    def get_long_short_comments(self, on_paper=False):
        """计算多空对冲的相关评价指标
        包括年化收益率、年化波动率、信息比率、月度胜率、最大回撤率"""
        # week_here
        self.long_short_ret_yearly = (
            self.long_short_net_values[-1]
            ** (self.freq_ctrl.counts_one_year / len(self.long_short_net_values))
            - 1
        )
        self.inner_long_ret_yearly = (
            self.inner_long_net_values[-1]
            ** (self.freq_ctrl.counts_one_year / len(self.inner_long_net_values))
            - 1
        )
        self.inner_short_ret_yearly = (
            self.inner_short_net_values[-1]
            ** (self.freq_ctrl.counts_one_year / len(self.inner_short_net_values))
            - 1
        )
        # week_here
        self.long_short_vol_yearly = np.std(self.long_short_rets) * (
            self.freq_ctrl.counts_one_year**0.5
        )
        self.long_short_info_ratio = (
            self.long_short_ret_yearly / self.long_short_vol_yearly
        )
        self.long_short_win_times = len(self.long_short_rets[self.long_short_rets > 0])
        self.long_short_win_ratio = self.long_short_win_times / len(
            self.long_short_rets
        )
        self.max_retreat = -(
            self.long_short_net_values / self.long_short_net_values.expanding(1).max()
            - 1
        ).min()
        if on_paper:
            self.long_short_comments = pd.DataFrame(
                {
                    "评价指标": [
                        self.long_short_ret_yearly,
                        self.long_short_vol_yearly,
                        self.long_short_info_ratio,
                        self.long_short_win_ratio,
                        self.max_retreat,
                    ]
                },
                # week_here
                index=[
                    "年化收益率",
                    "年化波动率",
                    "收益波动比",
                    f"{self.freq_ctrl.comment_name}度胜率",
                    "最大回撤率",
                ],
            )
        else:
            self.long_short_comments = pd.DataFrame(
                {
                    "评价指标": [
                        self.long_short_ret_yearly,
                        self.long_short_vol_yearly,
                        self.long_short_info_ratio,
                        self.long_short_win_ratio,
                        self.max_retreat,
                    ]
                },
                # week_here
                index=[
                    "年化收益率",
                    "年化波动率",
                    "信息比率",
                    f"{self.freq_ctrl.comment_name}度胜率",
                    "最大回撤率",
                ],
            )

    def get_total_comments(self, groups_num):
        """综合IC、ICIR、RankIC、RankICIR,年化收益率、年化波动率、信息比率、胜率、最大回撤率"""
        rankic = self.rankics.mean()
        rankic_win = self.rankics[self.rankics * rankic > 0]
        rankic_win_ratio = rankic_win.dropna().shape[0] / self.rankics.dropna().shape[0]
        self.factor_cross_skew_after_neu = self.__factors_out.skew(axis=1).mean()
        if self.ic_icir_and_rank.iloc[2, 0] > 0:
            self.factor_turnover_rate = self.factor_turnover_rates[
                f"group{groups_num}"
            ].mean()
        else:
            self.factor_turnover_rate = self.factor_turnover_rates["group1"].mean()
        self.total_comments = pd.concat(
            [
                self.ic_icir_and_rank,
                pd.DataFrame(
                    {"评价指标": [rankic_win_ratio]},
                    index=["RankIC胜率"],
                ),
                self.long_short_comments,
                # week_here
                pd.DataFrame(
                    {
                        "评价指标": [
                            self.factor_turnover_rate,
                            self.factor_cover,
                            self.pos_neg_rate,
                            self.factor_cross_skew,
                            self.inner_long_ret_yearly,
                            self.inner_long_ret_yearly
                            / (
                                self.inner_long_ret_yearly + self.inner_short_ret_yearly
                            ),
                            self.corr_itself,
                        ]
                    },
                    index=[
                        f"多头{self.freq_ctrl.comment_name}均换手",
                        "因子覆盖率",
                        "因子正值占比",
                        "因子截面偏度",
                        "多头超均收益",
                        "多头收益占比",
                        "一阶自相关性",
                    ],
                ),
            ]
        )
        self.group_mean_rets_monthly=self.group_rets.drop(columns=['long_short']).mean()
        self.group_mean_rets_monthly=self.group_mean_rets_monthly-self.group_mean_rets_monthly.mean()

    def plot_net_values(self, y2, filename, iplot=1, ilegend=1, without_breakpoint=0):
        """使用matplotlib来画图，y2为是否对多空组合采用双y轴"""
        if not iplot:
            fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(33, 8))
            self.group_net_values.plot(secondary_y=y2, rot=60, ax=ax[0])
            self.group_net_values.plot(secondary_y=y2, ax=ax[0])
            b = self.rankics.copy()
            b.index = [int(i.year) if i.month == 1 else "" for i in list(b.index)]
            b.plot(kind="bar", rot=60, ax=ax[1])
            self.factor_cross_stds.plot(rot=60, ax=ax[2])

            filename_path = filename + ".png"
            if not STATES["NO_SAVE"]:
                plt.savefig(filename_path)
        else:

            tris = self.group_net_values
            if without_breakpoint:
                tris = tris.dropna()
            figs = cf.figures(
                tris,
                [
                    dict(kind="line", y=list(self.group_net_values.columns)),
                    # dict(kind="bar", y="各组月均超均收益"),
                    # dict(kind="bar", y="rankic"),
                ],
                asList=True,
            )
            comments = (
                self.total_comments.applymap(lambda x: round(x, 4))
                .rename(index={"RankIC均值t值": "RankIC.t"})
                .reset_index()
            )
            here = pd.concat(
                [
                    comments.iloc[:6, :].reset_index(drop=True),
                    comments.iloc[6:12, :].reset_index(drop=True),
                    comments.iloc[12:, :].reset_index(drop=True),
                ],
                axis=1,
            )
            here.columns = ["信息系数", "结果", "绩效指标", "结果", "其他指标", "结果"]
            # here=here.to_numpy().tolist()+[['信息系数','结果','绩效指标','结果']]
            table = FF.create_table(here.iloc[::-1])
            table.update_yaxes(matches=None)
            pic2=go.Figure(go.Bar(y=list(self.group_mean_rets_monthly),x=[i.replace('roup','') for i in list(self.group_mean_rets_monthly.index)]))
            # table=go.Figure([go.Table(header=dict(values=list(here.columns)),cells=dict(values=here.to_numpy().tolist()))])
            pic3_data=go.Bar(y=list(self.rankics.rankic),x=list(self.rankics.index))
            pic3=go.Figure(data=[pic3_data])
            pic4_data=go.Line(y=list(self.rankics.rankic.cumsum()),x=list(self.rankics.index),name='y2',yaxis='y2')
            pic4_layout=go.Layout(yaxis2=dict(title='y2',side='right'))
            pic4=go.Figure(data=[pic4_data],layout=pic4_layout)
            figs.append(table)
            figs = [figs[-1]] + figs[:-1]
            figs.append(pic2)
            figs = [figs[0],figs[1],figs[-1],pic3]
            figs[1].update_layout(
                legend=dict(yanchor="top", y=0.99, xanchor="left", x=0.01)
            )
            figs[3].update_layout(yaxis2=dict(title='y2',side='right'))
            base_layout = cf.tools.get_base_layout(figs)

            sp = cf.subplots(
                figs,
                shape=(2, 11),
                base_layout=base_layout,
                vertical_spacing=0.15,
                horizontal_spacing=0.03,
                shared_yaxes=False,

                specs=[
                    [
                        {"rowspan": 2, "colspan": 4},
                        None,
                        None,
                        None,
                        {"rowspan": 2, "colspan": 4},
                        None,
                        None,
                        None,
                        {"colspan": 3},
                        None,
                        None,
                    ],
                    [
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        {"colspan": 3},
                        None,
                        None,
                    ],
                ],
                subplot_titles=["净值曲线", "各组月均超均收益", "Rank IC时序图", "绩效指标"],
            )
            sp["layout"].update(showlegend=ilegend)
            # los=sp['layout']['annotations']
            # los[0]['font']['color']='#000000'
            # los[1]['font']['color']='#000000'
            # los[2]['font']['color']='#000000'
            # los[3]['font']['color']='#000000'
            # los[-1]['font']['color']='#ffffff'
            # los[-2]['font']['color']='#ffffff'
            # los[-3]['font']['color']='#ffffff'
            # los[-4]['font']['color']='#ffffff'
            # los[0]['text']=los[0]['text'][3:-4]
            # los[1]['text']=los[1]['text'][3:-4]
            # los[2]['text']=los[2]['text'][3:-4]
            # los[3]['text']=los[3]['text'][3:-4]
            # los[-1]['text']='<b>'+los[-1]['text']+'</b>'
            # los[-2]['text']='<b>'+los[-2]['text']+'</b>'
            # los[-3]['text']='<b>'+los[-3]['text']+'</b>'
            # los[-4]['text']='<b>'+los[-4]['text']+'</b>'
            # sp['layout']['annotations']=los
            # print(sp['layout']['annotations'])
            # sp['layout']['annotations'][0]['yanchor']='top'

            cf.iplot(sp)
            # tris=pd.concat([self.group_net_values,self.rankics,self.factor_turnover_rates],axis=1).rename(columns={0:'turnover_rate'})
            # sp=plyoo.make_subplots(rows=2,cols=8,vertical_spacing=.15,horizontal_spacing=.03,
            #                specs=[[{'rowspan':2,'colspan':2,'type':'domain'},None,{'rowspan':2,'colspan':4,'type':'xy'},None,None,None,{'colspan':2,'type':'xy'},None],
            #                       [None,None,None,None,None,None,{'colspan':2,'type':'xy'},None]],
            #                subplot_titles=['净值曲线','Rank IC时序图','月换手率','绩效指标'])
            # comments=self.total_comments.applymap(lambda x:round(x,4)).rename(index={'RankIC均值t值':'RankIC.t'}).reset_index()
            # here=pd.concat([comments.iloc[:5,:].reset_index(drop=True),comments.iloc[5:,:].reset_index(drop=True)],axis=1)
            # here.columns=['信息系数','结果','绩效指标','结果']
            # table=FF.create_table(here)
            # sp.add_trace(table)

    def plotly_net_values(self, filename):
        """使用plotly.express画图"""
        fig = pe.line(self.group_net_values)
        filename_path = filename + ".html"
        pio.write_html(fig, filename_path, auto_open=True)

    @classmethod
    @lru_cache(maxsize=None)
    def prerpare(cls):
        """通用数据准备"""
        cls.judge_month()
        cls.get_rets_month()

    def run(
        self,
        groups_num=10,
        neutralize=False,
        boxcox=False,
        trade_cost_double_side=0,
        value_weighted=False,
        y2=False,
        plt_plot=True,
        plotly_plot=False,
        filename="分组净值图",
        print_comments=True,
        comments_writer=None,
        net_values_writer=None,
        rets_writer=None,
        comments_sheetname=None,
        net_values_sheetname=None,
        rets_sheetname=None,
        on_paper=False,
        sheetname=None,
        zxindustry_dummies=0,
        swindustry_dummies=0,
        only_cap=0,
        iplot=1,
        ilegend=0,
        without_breakpoint=0,
        beauty_comments=0,
    ):
        """运行回测部分"""
        if comments_writer and not (comments_sheetname or sheetname):
            raise IOError("把total_comments输出到excel中时，必须指定sheetname🤒")
        if net_values_writer and not (net_values_sheetname or sheetname):
            raise IOError("把group_net_values输出到excel中时，必须指定sheetname🤒")
        if rets_writer and not (rets_sheetname or sheetname):
            raise IOError("把group_rets输出到excel中时，必须指定sheetname🤒")
        if neutralize:
            self.get_log_cap()
            self.get_neutral_factors(
                swindustry_dummies=swindustry_dummies,
                zxindustry_dummies=zxindustry_dummies,
            )
            self.deal_with_factors_after_neutralize()
        elif boxcox:
            self.get_log_cap(boxcox=True)
            self.get_neutral_factors(
                swindustry_dummies=swindustry_dummies,
                zxindustry_dummies=zxindustry_dummies,
                only_cap=only_cap,
            )
            self.deal_with_factors_after_neutralize()
        else:
            self.deal_with_factors()
        self.get_limit_ups_downs()
        self.get_data(groups_num)
        self.get_group_rets_net_values(
            groups_num=groups_num,
            value_weighted=value_weighted,
            trade_cost_double_side=trade_cost_double_side,
        )
        self.get_long_short_comments(on_paper=on_paper)
        self.get_total_comments(groups_num=groups_num)
        if on_paper:
            group1_ttest = ss.ttest_1samp(self.group_rets.group1, 0).pvalue
            group10_ttest = ss.ttest_1samp(
                self.group_rets[f"group{groups_num}"], 0
            ).pvalue
            group_long_short_ttest = ss.ttest_1samp(self.long_short_rets, 0).pvalue
            group1_ret = self.group_rets.group1.mean()
            group10_ret = self.group_rets[f"group{groups_num}"].mean()
            group_long_short_ret = self.long_short_rets.mean()
            papers = pd.DataFrame(
                {
                    "评价指标": [
                        group1_ttest,
                        group10_ttest,
                        group_long_short_ttest,
                        group1_ret,
                        group10_ret,
                        group_long_short_ret,
                    ]
                },
                index=[
                    "分组1p值",
                    f"分组{groups_num}p值",
                    f"分组1-分组{groups_num}p值",
                    "分组1收益率",
                    f"分组{groups_num}收益率",
                    f"分组1-分组{groups_num}收益率",
                ],
            )
            self.total_comments = pd.concat([papers, self.total_comments])

        if plt_plot:
            if not STATES["NO_PLOT"]:
                if filename:
                    self.plot_net_values(
                        y2=y2,
                        filename=filename,
                        iplot=iplot,
                        ilegend=bool(ilegend),
                        without_breakpoint=without_breakpoint,
                    )
                else:
                    self.plot_net_values(
                        y2=y2,
                        filename=self.factors_file.split(".")[-2].split("/")[-1]
                        + str(groups_num)
                        + "分组",
                        iplot=iplot,
                        ilegend=bool(ilegend),
                        without_breakpoint=without_breakpoint,
                    )
                plt.show()
        if plotly_plot:
            if not STATES["NO_PLOT"]:
                if filename:
                    self.plotly_net_values(filename=filename)
                else:
                    self.plotly_net_values(
                        filename=self.factors_file.split(".")[-2].split("/")[-1]
                        + str(groups_num)
                        + "分组"
                    )
        if print_comments:
            if not STATES["NO_COMMENT"]:
                tb = Texttable()
                tb.set_cols_width(
                    [8] * 5 + [9] + [8] * 2 + [7] * 2 + [8] + [8] + [9] + [10] * 5
                )
                tb.set_cols_dtype(["f"] * 18)
                tb.header(list(self.total_comments.T.columns))
                tb.add_rows(self.total_comments.T.to_numpy(), header=False)
                print(tb.draw())
        if sheetname:
            if comments_writer:
                if not on_paper:
                    total_comments = self.total_comments.copy()
                    tc = list(total_comments.评价指标)
                    if beauty_comments:
                        tc[0] = str(round(tc[0] * 100, 2)) + "%"
                        tc[1] = str(round(tc[1], 2))
                        tc[2] = str(round(tc[2] * 100, 2)) + "%"
                        tc[3] = str(round(tc[3], 2))
                        tc[4] = str(round(tc[4], 2))
                        tc[5] = str(round(tc[5] * 100, 2)) + "%"
                        tc[6] = str(round(tc[6] * 100, 2)) + "%"
                        tc[7] = str(round(tc[7] * 100, 2)) + "%"
                        tc[8] = str(round(tc[8], 2))
                        tc[9] = str(round(tc[9] * 100, 2)) + "%"
                        tc[10] = str(round(tc[10] * 100, 2)) + "%"
                        tc[11] = str(round(tc[11] * 100, 2)) + "%"
                        tc[12] = str(round(tc[12] * 100, 2)) + "%"
                        tc[13] = str(round(tc[13] * 100, 2)) + "%"
                        tc[14] = str(round(tc[14], 2))
                        tc[15] = str(round(tc[15], 2))
                        tc[16] = str(round(tc[16] * 100, 2)) + "%"
                        tc[17] = str(round(tc[17] * 100, 2)) + "%"
                    tc=tc+list(self.group_mean_rets_monthly)
                    new_total_comments = pd.DataFrame(
                        {sheetname: tc}, index=list(total_comments.index)+[f'第{i}组' for i in range(1,groups_num+1)]
                    )
                    new_total_comments.to_excel(comments_writer, sheet_name=sheetname)
                    rankic_twins=pd.concat([self.rankics.rankic,self.rankics.rankic.cumsum()],axis=1)
                    rankic_twins.columns=['RankIC','RankIC累积']
                    rankic_twins.to_excel(comments_writer,sheet_name=sheetname+'RankIC')
                else:
                    self.total_comments.rename(columns={"评价指标": sheetname}).to_excel(
                        comments_writer, sheet_name=sheetname
                    )
            if net_values_writer:
                groups_net_values = self.group_net_values.copy()
                groups_net_values.index = groups_net_values.index.strftime("%Y/%m/%d")
                groups_net_values.columns = [
                    f"分组{i}" for i in range(1, len(list(groups_net_values.columns)))
                ] + ["多空对冲（右轴）"]
                groups_net_values.to_excel(net_values_writer, sheet_name=sheetname)
            if rets_writer:
                group_rets = self.group_rets.copy()
                group_rets.index = group_rets.index.strftime("%Y/%m/%d")
                group_rets.columns = [
                    f"分组{i}" for i in range(1, len(list(group_rets.columns)))
                ] + ["多空对冲（右轴）"]
                group_rets.to_excel(rets_writer, sheet_name=sheetname)
        else:
            if comments_writer and comments_sheetname:
                total_comments = self.total_comments.copy()
                tc = list(total_comments.评价指标)
                tc[0] = str(round(tc[0] * 100, 2)) + "%"
                tc[1] = str(round(tc[1], 2))
                tc[2] = str(round(tc[2] * 100, 2)) + "%"
                tc[3] = str(round(tc[3], 2))
                tc[4] = str(round(tc[4], 2))
                tc[5] = str(round(tc[5] * 100, 2)) + "%"
                tc[6] = str(round(tc[6] * 100, 2)) + "%"
                tc[7] = str(round(tc[7] * 100, 2)) + "%"
                tc[8] = str(round(tc[8], 2))
                tc[9] = str(round(tc[9] * 100, 2)) + "%"
                tc[10] = str(round(tc[10] * 100, 2)) + "%"
                tc[11] = str(round(tc[11] * 100, 2)) + "%"
                tc[12] = str(round(tc[12] * 100, 2)) + "%"
                tc[13] = str(round(tc[13] * 100, 2)) + "%"
                tc[14] = str(round(tc[14], 2))
                tc[15] = str(round(tc[15], 2))
                tc[16] = str(round(tc[16] * 100, 2)) + "%"
                tc[17] = str(round(tc[17] * 100, 2)) + "%"
                new_total_comments = pd.DataFrame(
                    {comments_sheetname: tc}, index=total_comments.index
                )
                new_total_comments.T.to_excel(
                    comments_writer, sheet_name=comments_sheetname
                )
            if net_values_writer and net_values_sheetname:
                groups_net_values = self.group_net_values.copy()
                groups_net_values.index = groups_net_values.index.strftime("%Y/%m/%d")
                groups_net_values.columns = [
                    f"分组{i}" for i in range(1, len(list(groups_net_values.columns)))
                ] + ["多空对冲（右轴）"]
                groups_net_values.to_excel(
                    net_values_writer, sheet_name=net_values_sheetname
                )
            if rets_writer and rets_sheetname:
                group_rets = self.group_rets.copy()
                group_rets.index = group_rets.index.strftime("%Y/%m/%d")
                group_rets.columns = [
                    f"分组{i}" for i in range(1, len(list(group_rets.columns)))
                ] + ["多空对冲（右轴）"]
                group_rets.to_excel(rets_writer, sheet_name=rets_sheetname)

__call__()

调用对象则返回因子值

Source code in pure_ocean_breeze/labor/process.py
Python
1586
1587
1588
def __call__(self):
    """调用对象则返回因子值"""
    return self.factors_out

daily_to_monthly(pridf, path, func) classmethod

把日度的交易状态、st、上市天数，转化为月度的，并生成能否交易的判断 读取本地已经算好的文件，并追加新的时间段部分，如果本地没有就直接全部重新算

Source code in pure_ocean_breeze/labor/process.py
Python
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
@classmethod
def daily_to_monthly(cls, pridf, path, func):
    """把日度的交易状态、st、上市天数，转化为月度的，并生成能否交易的判断
    读取本地已经算好的文件，并追加新的时间段部分，如果本地没有就直接全部重新算"""
    try:
        month_df = pd.read_parquet(path)
        month_df = cls.read_add(pridf, month_df, func)
        month_df.to_parquet(path)
    except Exception as e:
        if not STATES["NO_LOG"]:
            logger.error("error occurs when read state files")
            logger.error(e)
        print("state file rewriting……")
        # week_here
        df_1 = pridf.index.max()
        year = df_1.year
        month = df_1.month
        last = tt.date.get_close(year=year, m=month).pd_date()
        if not (last == df_1)[0]:
            pridf = pridf[pridf.index < pd.Timestamp(year=year, month=month, day=1)]
        month_df = pridf.resample(cls.freq).apply(func)
        month_df.to_parquet(path)
    return month_df

deal_with_factors()

删除不符合交易条件的因子数据

Source code in pure_ocean_breeze/labor/process.py
Python
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
def deal_with_factors(self):
    """删除不符合交易条件的因子数据"""
    self.__factors_out = self.factors.copy()
    # week_here
    self.factors.index = self.factors.index + self.freq_ctrl.time_shift
    # week_here
    self.factors = self.factors.resample(self.freq).last()
    self.factors = self.factors * self.tris_monthly
    self.factors = self.factors.stack().reset_index()
    self.factors.columns = ["date", "code", "fac"]

deal_with_factors_after_neutralize()

中性化之后的因子处理方法

Source code in pure_ocean_breeze/labor/process.py
Python
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
def deal_with_factors_after_neutralize(self):
    """中性化之后的因子处理方法"""
    self.factors = self.factors.set_index(["date", "code"])
    self.factors = self.factors.unstack()
    self.__factors_out = self.factors.copy()
    self.__factors_out.columns = [i[1] for i in list(self.__factors_out.columns)]
    # week_here
    self.factors.index = self.factors.index + self.freq_ctrl.time_shift
    # week_here
    self.factors = self.factors.resample(self.freq).last()
    self.factors.columns = list(map(lambda x: x[1], list(self.factors.columns)))
    self.factors = self.factors.stack().reset_index()
    self.factors.columns = ["date", "code", "fac"]

find_limit(df, up=1) classmethod

计算涨跌幅超过9.8%的股票，并将其存储进一个长列表里 其中时间列，为某月的最后一天；涨停日虽然为下月初第一天，但这里标注的时间统一为上月最后一天

Source code in pure_ocean_breeze/labor/process.py
Python
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
@classmethod
def find_limit(cls, df, up=1):
    """计算涨跌幅超过9.8%的股票，并将其存储进一个长列表里
    其中时间列，为某月的最后一天；涨停日虽然为下月初第一天，但这里标注的时间统一为上月最后一天"""
    limit_df = np.sign(df.applymap(lambda x: x - up * 0.098)).replace(
        -1 * up, np.nan
    )
    limit_df = limit_df.stack().reset_index()
    limit_df.columns = ["date", "code", "limit_up_signal"]
    limit_df = limit_df[["date", "code"]]
    return limit_df

get_data(groups_num)

拼接因子数据和每月收益率数据，并对涨停和跌停股加以处理

Source code in pure_ocean_breeze/labor/process.py
Python
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
def get_data(self, groups_num):
    """拼接因子数据和每月收益率数据，并对涨停和跌停股加以处理"""
    self.data = pd.merge(
        self.rets_monthly, self.factors, how="inner", on=["date", "code"]
    )
    self.ic_icir_and_rank = self.get_ic_icir_and_rank(self.data)
    self.data = self.data.groupby("date").apply(
        lambda x: self.get_groups(x, groups_num)
    )
    self.wind_out = self.data.copy()
    self.factor_turnover_rates = self.data.pivot(
        index="date", columns="code", values="group"
    )
    rates = []
    for i in range(1, groups_num + 1):
        son = (self.factor_turnover_rates == i) + 0
        son1 = son.diff()
        # self.factor_turnover_rates = self.factor_turnover_rates.diff()
        change = ((np.abs(np.sign(son1)) == 1) + 0).sum(axis=1)
        still = (((son1 == 0) + 0) * son).sum(axis=1)
        rate = change / (change + still)
        rates.append(rate.to_frame(f"group{i}"))
    rates = pd.concat(rates, axis=1).fillna(0)
    self.factor_turnover_rates = rates
    self.data = self.data.reset_index(drop=True)
    limit_ups_object = self.limit_old_to_new(self.limit_ups, self.data)
    limit_downs_object = self.limit_old_to_new(self.limit_downs, self.data)
    self.data = self.data.drop(limit_ups_object.index)
    rets_monthly_limit_downs = pd.merge(
        self.rets_monthly, limit_downs_object, how="inner", on=["date", "code"]
    )
    self.data = pd.concat([self.data, rets_monthly_limit_downs])

get_group_rets_net_values(groups_num=10, value_weighted=False, trade_cost_double_side=0)

计算组内每一期的平均收益，生成每日收益率序列和净值序列

Source code in pure_ocean_breeze/labor/process.py
Python
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
def get_group_rets_net_values(
    self, groups_num=10, value_weighted=False, trade_cost_double_side=0
):
    """计算组内每一期的平均收益，生成每日收益率序列和净值序列"""
    if value_weighted:
        cap_value = self.capital.copy()
        # week_here
        cap_value = cap_value.resample(self.freq).last().shift(1)
        cap_value = cap_value * self.tris_monthly
        # cap_value=np.log(cap_value)
        cap_value = cap_value.stack().reset_index()
        cap_value.columns = ["date", "code", "cap_value"]
        self.data = pd.merge(self.data, cap_value, on=["date", "code"])

        def in_g(df):
            df.cap_value = df.cap_value / df.cap_value.sum()
            df.ret = df.ret * df.cap_value
            return df.ret.sum()

        self.group_rets = self.data.groupby(["date", "group"]).apply(in_g)
        self.rets_all = self.data.groupby(["date"]).apply(in_g)
        self.group_rets_std = "市值加权暂未设置该功能，敬请期待🌙"
    else:
        self.group_rets = self.data.groupby(["date", "group"]).apply(
            lambda x: x.ret.mean()
        )
        self.rets_all = self.data.groupby(["date"]).apply(lambda x: x.ret.mean())
        self.group_rets_stds = self.data.groupby(["date", "group"]).ret.std()
        self.group_rets_std = (
            self.group_rets_stds.reset_index().groupby("group").mean()
        )
        self.group_rets_skews = self.data.groupby(["date", "group"]).ret.skew()
        self.group_rets_skew = (
            self.group_rets_skews.reset_index().groupby("group").mean()
        )
    # dropna是因为如果股票行情数据比因子数据的截止日期晚，而最后一个月发生月初跌停时，会造成最后某组多出一个月的数据
    self.group_rets = self.group_rets.unstack()
    self.group_rets = self.group_rets[
        self.group_rets.index <= self.factors.date.max()
    ]
    self.group_rets.columns = list(map(str, list(self.group_rets.columns)))
    self.group_rets = self.group_rets.add_prefix("group")
    self.group_rets = (
        self.group_rets - self.factor_turnover_rates * trade_cost_double_side
    )
    self.rets_all = (
        self.rets_all
        - self.factor_turnover_rates.mean(axis=1) * trade_cost_double_side
    ).dropna()
    self.long_short_rets = (
        self.group_rets["group1"] - self.group_rets["group" + str(groups_num)]
    )
    self.inner_rets_long = self.group_rets.group1 - self.rets_all
    self.inner_rets_short = (
        self.rets_all - self.group_rets["group" + str(groups_num)]
    )
    self.long_short_net_values = self.make_start_to_one(
        (self.long_short_rets + 1).cumprod()
    )
    if self.long_short_net_values[-1] <= self.long_short_net_values[0]:
        self.long_short_rets = (
            self.group_rets["group" + str(groups_num)] - self.group_rets["group1"]
        )
        self.long_short_net_values = self.make_start_to_one(
            (self.long_short_rets + 1).cumprod()
        )
        self.inner_rets_long = (
            self.group_rets["group" + str(groups_num)] - self.rets_all
        )
        self.inner_rets_short = self.rets_all - self.group_rets.group1
    self.inner_long_net_values = self.make_start_to_one(
        (self.inner_rets_long + 1).cumprod()
    )
    self.inner_short_net_values = self.make_start_to_one(
        (self.inner_rets_short + 1).cumprod()
    )
    self.group_rets = self.group_rets.assign(long_short=self.long_short_rets)
    self.group_net_values = self.group_rets.applymap(lambda x: x + 1)
    self.group_net_values = self.group_net_values.cumprod()
    self.group_net_values = self.group_net_values.apply(self.make_start_to_one)
    a = groups_num ** (0.5)
    # 判断是否要两个因子画表格
    if a == int(a):
        self.square_rets = (
            self.group_net_values.iloc[:, :-1].apply(self.to_group_ret).to_numpy()
        )
        self.square_rets = self.square_rets.reshape((int(a), int(a)))
        self.square_rets = pd.DataFrame(
            self.square_rets,
            columns=list(range(1, int(a) + 1)),
            index=list(range(1, int(a) + 1)),
        )
        print("这是self.square_rets", self.square_rets)

get_groups(df, groups_num) classmethod

依据因子值，判断是在第几组

Source code in pure_ocean_breeze/labor/process.py
Python
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
@classmethod
def get_groups(cls, df, groups_num):
    """依据因子值，判断是在第几组"""
    if "group" in list(df.columns):
        df = df.drop(columns=["group"])
    df = df.sort_values(["fac"], ascending=True)
    each_group = round(df.shape[0] / groups_num)
    l = list(
        map(
            lambda x, y: [x] * y,
            list(range(1, groups_num + 1)),
            [each_group] * groups_num,
        )
    )
    l = reduce(lambda x, y: x + y, l)
    if len(l) < df.shape[0]:
        l = l + [groups_num] * (df.shape[0] - len(l))
    l = l[: df.shape[0]]
    df.insert(0, "group", l)
    return df

get_ic_icir_and_rank(df)

计算IC、ICIR、RankIC、RankICIR

Source code in pure_ocean_breeze/labor/process.py
Python
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
def get_ic_icir_and_rank(cls, df):
    """计算IC、ICIR、RankIC、RankICIR"""
    df1 = df.groupby("date").apply(cls.get_ic_rankic)
    cls.ics = df1.ic
    cls.rankics = df1.rankic
    cls.ics = cls.ics.reset_index(drop=True, level=1).to_frame()
    cls.rankics = cls.rankics.reset_index(drop=True, level=1).to_frame()
    df2 = cls.get_icir_rankicir(df1)
    df2 = df2.T
    dura = (df.date.max() - df.date.min()).days / 365
    t_value = df2.iloc[3, 0] * (dura ** (1 / 2))
    df3 = pd.DataFrame({"评价指标": [t_value]}, index=["RankIC.t"])
    df4 = pd.concat([df2, df3])
    return df4

get_ic_rankic(df)

计算IC和RankIC

Source code in pure_ocean_breeze/labor/process.py
Python
1897
1898
1899
1900
1901
1902
1903
def get_ic_rankic(cls, df):
    """计算IC和RankIC"""
    df1 = df[["ret", "fac"]]
    ic = df1.corr(method="pearson").iloc[0, 1]
    rankic = df1.rank().corr().iloc[0, 1]
    df2 = pd.DataFrame({"ic": [ic], "rankic": [rankic]})
    return df2

get_icir_rankicir(df)

计算ICIR和RankICIR

Source code in pure_ocean_breeze/labor/process.py
Python
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
def get_icir_rankicir(cls, df):
    """计算ICIR和RankICIR"""
    ic = df.ic.mean()
    rankic = df.rankic.mean()
    # week_here
    icir = ic / np.std(df.ic) * (cls.freq_ctrl.counts_one_year ** (0.5))
    # week_here
    rankicir = rankic / np.std(df.rankic) * (cls.freq_ctrl.counts_one_year ** (0.5))
    return pd.DataFrame(
        {"IC": [ic], "ICIR": [icir], "RankIC": [rankic], "RankICIR": [rankicir]},
        index=["评价指标"],
    )

get_limit_ups_downs() cached classmethod

找月初第一天就涨停

Source code in pure_ocean_breeze/labor/process.py
Python
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
@classmethod
@lru_cache(maxsize=None)
def get_limit_ups_downs(cls):
    """找月初第一天就涨停"""
    """或者是月末跌停的股票"""
    cls.opens_monthly_shift = cls.opens_monthly.copy()
    cls.opens_monthly_shift = cls.opens_monthly_shift.shift(-1)
    cls.rets_monthly_begin = (
        cls.opens_monthly_shift - cls.closes_monthly
    ) / cls.closes_monthly
    # week_here
    cls.closes2_monthly = cls.closes.shift(1).resample(cls.freq).last()
    cls.rets_monthly_last = (
        cls.closes_monthly - cls.closes2_monthly
    ) / cls.closes2_monthly
    cls.limit_ups = cls.find_limit(cls.rets_monthly_begin, up=1)
    cls.limit_downs = cls.find_limit(cls.rets_monthly_last, up=-1)

get_log_cap(boxcox=True) cached classmethod

获得对数市值

Source code in pure_ocean_breeze/labor/process.py
Python
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
@classmethod
@lru_cache(maxsize=None)
def get_log_cap(cls, boxcox=True):
    """获得对数市值"""
    cls.cap = cls.capital.stack().reset_index()
    cls.cap.columns = ["date", "code", "cap_size"]
    if boxcox:

        def single(x):
            x.cap_size = ss.boxcox(x.cap_size)[0]
            return x

        cls.cap = cls.cap.groupby(["date"]).apply(single)
    else:
        cls.cap["cap_size"] = np.log(cls.cap["cap_size"])

get_long_short_comments(on_paper=False)

计算多空对冲的相关评价指标 包括年化收益率、年化波动率、信息比率、月度胜率、最大回撤率

Source code in pure_ocean_breeze/labor/process.py
Python
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
def get_long_short_comments(self, on_paper=False):
    """计算多空对冲的相关评价指标
    包括年化收益率、年化波动率、信息比率、月度胜率、最大回撤率"""
    # week_here
    self.long_short_ret_yearly = (
        self.long_short_net_values[-1]
        ** (self.freq_ctrl.counts_one_year / len(self.long_short_net_values))
        - 1
    )
    self.inner_long_ret_yearly = (
        self.inner_long_net_values[-1]
        ** (self.freq_ctrl.counts_one_year / len(self.inner_long_net_values))
        - 1
    )
    self.inner_short_ret_yearly = (
        self.inner_short_net_values[-1]
        ** (self.freq_ctrl.counts_one_year / len(self.inner_short_net_values))
        - 1
    )
    # week_here
    self.long_short_vol_yearly = np.std(self.long_short_rets) * (
        self.freq_ctrl.counts_one_year**0.5
    )
    self.long_short_info_ratio = (
        self.long_short_ret_yearly / self.long_short_vol_yearly
    )
    self.long_short_win_times = len(self.long_short_rets[self.long_short_rets > 0])
    self.long_short_win_ratio = self.long_short_win_times / len(
        self.long_short_rets
    )
    self.max_retreat = -(
        self.long_short_net_values / self.long_short_net_values.expanding(1).max()
        - 1
    ).min()
    if on_paper:
        self.long_short_comments = pd.DataFrame(
            {
                "评价指标": [
                    self.long_short_ret_yearly,
                    self.long_short_vol_yearly,
                    self.long_short_info_ratio,
                    self.long_short_win_ratio,
                    self.max_retreat,
                ]
            },
            # week_here
            index=[
                "年化收益率",
                "年化波动率",
                "收益波动比",
                f"{self.freq_ctrl.comment_name}度胜率",
                "最大回撤率",
            ],
        )
    else:
        self.long_short_comments = pd.DataFrame(
            {
                "评价指标": [
                    self.long_short_ret_yearly,
                    self.long_short_vol_yearly,
                    self.long_short_info_ratio,
                    self.long_short_win_ratio,
                    self.max_retreat,
                ]
            },
            # week_here
            index=[
                "年化收益率",
                "年化波动率",
                "信息比率",
                f"{self.freq_ctrl.comment_name}度胜率",
                "最大回撤率",
            ],
        )

get_neutral_factors(zxindustry_dummies=0, swindustry_dummies=0, only_cap=0)

对因子进行行业市值中性化

Source code in pure_ocean_breeze/labor/process.py
Python
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
def get_neutral_factors(
    self, zxindustry_dummies=0, swindustry_dummies=0, only_cap=0
):
    """对因子进行行业市值中性化"""
    # week_here
    self.factors.index = self.factors.index + self.freq_ctrl.time_shift
    # week_here
    self.factors = self.factors.resample(self.freq).last()
    # week_here
    last_date = self.freq_ctrl.next_end(self.tris_monthly.index.max())
    add_tail = pd.DataFrame(1, index=[last_date], columns=self.tris_monthly.columns)
    tris_monthly = pd.concat([self.tris_monthly, add_tail])
    self.factors = self.factors * tris_monthly
    # week_here
    self.factors.index = self.factors.index - self.freq_ctrl.time_shift
    # week_here
    self.factors = self.factors.resample(self.freq).last()
    self.factors = self.factors.stack().reset_index()
    self.factors.columns = ["date", "code", "fac"]
    self.factors = pd.merge(
        self.factors, self.cap, how="inner", on=["date", "code"]
    )
    if not only_cap:
        if swindustry_dummies:
            self.factors = pd.merge(
                self.factors, self.swindustry_dummy, on=["date", "code"]
            )
        else:
            self.factors = pd.merge(
                self.factors, self.zxindustry_dummy, on=["date", "code"]
            )
    self.factors = self.factors.set_index(["date", "code"])
    self.factors = self.factors.groupby(["date"]).apply(self.neutralize_factors)
    self.factors = self.factors.reset_index()

get_rets_month() cached classmethod

计算每月的收益率，并根据每月做出交易状态，做出删减

Source code in pure_ocean_breeze/labor/process.py
Python
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
@classmethod
@lru_cache(maxsize=None)
def get_rets_month(cls):
    """计算每月的收益率，并根据每月做出交易状态，做出删减"""
    # week_here
    cls.opens_monthly = cls.opens.resample(cls.freq).first()
    # week_here
    cls.closes_monthly = cls.closes.resample(cls.freq).last()
    cls.rets_monthly = (cls.closes_monthly - cls.opens_monthly) / cls.opens_monthly
    cls.rets_monthly = cls.rets_monthly * cls.tris_monthly
    cls.rets_monthly = cls.rets_monthly.stack().reset_index()
    cls.rets_monthly.columns = ["date", "code", "ret"]

get_total_comments(groups_num)

综合IC、ICIR、RankIC、RankICIR,年化收益率、年化波动率、信息比率、胜率、最大回撤率

Source code in pure_ocean_breeze/labor/process.py
Python
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
def get_total_comments(self, groups_num):
    """综合IC、ICIR、RankIC、RankICIR,年化收益率、年化波动率、信息比率、胜率、最大回撤率"""
    rankic = self.rankics.mean()
    rankic_win = self.rankics[self.rankics * rankic > 0]
    rankic_win_ratio = rankic_win.dropna().shape[0] / self.rankics.dropna().shape[0]
    self.factor_cross_skew_after_neu = self.__factors_out.skew(axis=1).mean()
    if self.ic_icir_and_rank.iloc[2, 0] > 0:
        self.factor_turnover_rate = self.factor_turnover_rates[
            f"group{groups_num}"
        ].mean()
    else:
        self.factor_turnover_rate = self.factor_turnover_rates["group1"].mean()
    self.total_comments = pd.concat(
        [
            self.ic_icir_and_rank,
            pd.DataFrame(
                {"评价指标": [rankic_win_ratio]},
                index=["RankIC胜率"],
            ),
            self.long_short_comments,
            # week_here
            pd.DataFrame(
                {
                    "评价指标": [
                        self.factor_turnover_rate,
                        self.factor_cover,
                        self.pos_neg_rate,
                        self.factor_cross_skew,
                        self.inner_long_ret_yearly,
                        self.inner_long_ret_yearly
                        / (
                            self.inner_long_ret_yearly + self.inner_short_ret_yearly
                        ),
                        self.corr_itself,
                    ]
                },
                index=[
                    f"多头{self.freq_ctrl.comment_name}均换手",
                    "因子覆盖率",
                    "因子正值占比",
                    "因子截面偏度",
                    "多头超均收益",
                    "多头收益占比",
                    "一阶自相关性",
                ],
            ),
        ]
    )
    self.group_mean_rets_monthly=self.group_rets.drop(columns=['long_short']).mean()
    self.group_mean_rets_monthly=self.group_mean_rets_monthly-self.group_mean_rets_monthly.mean()

judge_month() cached classmethod

生成一个月综合判断的表格

Source code in pure_ocean_breeze/labor/process.py
Python
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
@classmethod
@lru_cache(maxsize=None)
def judge_month(cls):
    """生成一个月综合判断的表格"""
    if cls.freq == "M":
        cls.sts_monthly = cls.daily_to_monthly(
            cls.sts, cls.sts_monthly_file, cls.judge_month_st
        )
        cls.states_monthly = cls.daily_to_monthly(
            cls.states, cls.states_monthly_file, cls.judge_month_state
        )
        # week_here
        cls.ages_monthly = (cls.ages.resample(cls.freq).last() > 60) + 0
        cls.tris_monthly = cls.sts_monthly * cls.states_monthly * cls.ages_monthly
        cls.tris_monthly = cls.tris_monthly.replace(0, np.nan)
    else:
        cls.tris_monthly = (
            (1 - cls.sts).resample(cls.freq).last().ffill(limit=2)
            * cls.states.resample(cls.freq).last().ffill(limit=2)
            * ((cls.ages.resample(cls.freq).last() > 60) + 0)
        ).replace(0, np.nan)

judge_month_st(df) classmethod

比较一个月内st的天数，如果st天数多，就删除本月，如果正常多，就保留本月

Source code in pure_ocean_breeze/labor/process.py
Python
1661
1662
1663
1664
1665
1666
1667
1668
1669
@classmethod
def judge_month_st(cls, df):
    """比较一个月内st的天数，如果st天数多，就删除本月，如果正常多，就保留本月"""
    st_count = len(df[df == 1])
    normal_count = len(df[df != 1])
    if st_count >= normal_count:
        return 0
    else:
        return 1

judge_month_state(df) classmethod

比较一个月内非正常交易的天数，如果非正常交易天数多，就删除本月，否则保留本月

Source code in pure_ocean_breeze/labor/process.py
Python
1671
1672
1673
1674
1675
1676
1677
1678
1679
@classmethod
def judge_month_state(cls, df):
    """比较一个月内非正常交易的天数，如果非正常交易天数多，就删除本月，否则保留本月"""
    abnormal_count = len(df[df == 0])
    normal_count = len(df[df == 1])
    if abnormal_count >= normal_count:
        return 0
    else:
        return 1

limit_old_to_new(limit, data) classmethod

获取跌停股在旧月的组号，然后将日期调整到新月里 涨停股则获得新月里涨停股的代码和时间，然后直接删去

Source code in pure_ocean_breeze/labor/process.py
Python
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
@classmethod
def limit_old_to_new(cls, limit, data):
    """获取跌停股在旧月的组号，然后将日期调整到新月里
    涨停股则获得新月里涨停股的代码和时间，然后直接删去"""
    data1 = data.copy()
    data1 = data1.reset_index()
    data1.columns = ["data_index"] + list(data1.columns)[1:]
    old = pd.merge(limit, data1, how="inner", on=["date", "code"])
    old = old.set_index("data_index")
    old = old[["group", "date", "code"]]
    # week_here
    old.date = list(map(cls.freq_ctrl.next_end, list(old.date)))
    return old

make_start_to_one(l)

让净值序列的第一个数变成1

Source code in pure_ocean_breeze/labor/process.py
Python
2001
2002
2003
2004
2005
2006
2007
def make_start_to_one(self, l):
    """让净值序列的第一个数变成1"""
    min_date = self.factors.date.min()
    add_date = min_date - relativedelta(days=min_date.day)
    add_l = pd.Series([1], index=[add_date])
    l = pd.concat([add_l, l])
    return l

neutralize_factors(df) classmethod

组内对因子进行市值中性化

Source code in pure_ocean_breeze/labor/process.py
Python
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
@classmethod
def neutralize_factors(cls, df):
    """组内对因子进行市值中性化"""
    industry_codes = list(df.columns)
    industry_codes = [i for i in industry_codes if i.startswith("w")]
    industry_codes_str = "+".join(industry_codes)
    if len(industry_codes_str) > 0:
        ols_result = smf.ols("fac~cap_size+" + industry_codes_str, data=df).fit()
    else:
        ols_result = smf.ols("fac~cap_size", data=df).fit()
    ols_w = ols_result.params["cap_size"]
    ols_b = ols_result.params["Intercept"]
    ols_bs = {}
    for ind in industry_codes:
        ols_bs[ind] = ols_result.params[ind]
    df.fac = df.fac - ols_w * df.cap_size - ols_b
    for k, v in ols_bs.items():
        df.fac = df.fac - v * df[k]
    df = df[["fac"]]
    return df

plot_net_values(y2, filename, iplot=1, ilegend=1, without_breakpoint=0)

使用matplotlib来画图，y2为是否对多空组合采用双y轴

Source code in pure_ocean_breeze/labor/process.py
Python
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
def plot_net_values(self, y2, filename, iplot=1, ilegend=1, without_breakpoint=0):
    """使用matplotlib来画图，y2为是否对多空组合采用双y轴"""
    if not iplot:
        fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(33, 8))
        self.group_net_values.plot(secondary_y=y2, rot=60, ax=ax[0])
        self.group_net_values.plot(secondary_y=y2, ax=ax[0])
        b = self.rankics.copy()
        b.index = [int(i.year) if i.month == 1 else "" for i in list(b.index)]
        b.plot(kind="bar", rot=60, ax=ax[1])
        self.factor_cross_stds.plot(rot=60, ax=ax[2])

        filename_path = filename + ".png"
        if not STATES["NO_SAVE"]:
            plt.savefig(filename_path)
    else:

        tris = self.group_net_values
        if without_breakpoint:
            tris = tris.dropna()
        figs = cf.figures(
            tris,
            [
                dict(kind="line", y=list(self.group_net_values.columns)),
                # dict(kind="bar", y="各组月均超均收益"),
                # dict(kind="bar", y="rankic"),
            ],
            asList=True,
        )
        comments = (
            self.total_comments.applymap(lambda x: round(x, 4))
            .rename(index={"RankIC均值t值": "RankIC.t"})
            .reset_index()
        )
        here = pd.concat(
            [
                comments.iloc[:6, :].reset_index(drop=True),
                comments.iloc[6:12, :].reset_index(drop=True),
                comments.iloc[12:, :].reset_index(drop=True),
            ],
            axis=1,
        )
        here.columns = ["信息系数", "结果", "绩效指标", "结果", "其他指标", "结果"]
        # here=here.to_numpy().tolist()+[['信息系数','结果','绩效指标','结果']]
        table = FF.create_table(here.iloc[::-1])
        table.update_yaxes(matches=None)
        pic2=go.Figure(go.Bar(y=list(self.group_mean_rets_monthly),x=[i.replace('roup','') for i in list(self.group_mean_rets_monthly.index)]))
        # table=go.Figure([go.Table(header=dict(values=list(here.columns)),cells=dict(values=here.to_numpy().tolist()))])
        pic3_data=go.Bar(y=list(self.rankics.rankic),x=list(self.rankics.index))
        pic3=go.Figure(data=[pic3_data])
        pic4_data=go.Line(y=list(self.rankics.rankic.cumsum()),x=list(self.rankics.index),name='y2',yaxis='y2')
        pic4_layout=go.Layout(yaxis2=dict(title='y2',side='right'))
        pic4=go.Figure(data=[pic4_data],layout=pic4_layout)
        figs.append(table)
        figs = [figs[-1]] + figs[:-1]
        figs.append(pic2)
        figs = [figs[0],figs[1],figs[-1],pic3]
        figs[1].update_layout(
            legend=dict(yanchor="top", y=0.99, xanchor="left", x=0.01)
        )
        figs[3].update_layout(yaxis2=dict(title='y2',side='right'))
        base_layout = cf.tools.get_base_layout(figs)

        sp = cf.subplots(
            figs,
            shape=(2, 11),
            base_layout=base_layout,
            vertical_spacing=0.15,
            horizontal_spacing=0.03,
            shared_yaxes=False,

            specs=[
                [
                    {"rowspan": 2, "colspan": 4},
                    None,
                    None,
                    None,
                    {"rowspan": 2, "colspan": 4},
                    None,
                    None,
                    None,
                    {"colspan": 3},
                    None,
                    None,
                ],
                [
                    None,
                    None,
                    None,
                    None,
                    None,
                    None,
                    None,
                    None,
                    {"colspan": 3},
                    None,
                    None,
                ],
            ],
            subplot_titles=["净值曲线", "各组月均超均收益", "Rank IC时序图", "绩效指标"],
        )
        sp["layout"].update(showlegend=ilegend)
        # los=sp['layout']['annotations']
        # los[0]['font']['color']='#000000'
        # los[1]['font']['color']='#000000'
        # los[2]['font']['color']='#000000'
        # los[3]['font']['color']='#000000'
        # los[-1]['font']['color']='#ffffff'
        # los[-2]['font']['color']='#ffffff'
        # los[-3]['font']['color']='#ffffff'
        # los[-4]['font']['color']='#ffffff'
        # los[0]['text']=los[0]['text'][3:-4]
        # los[1]['text']=los[1]['text'][3:-4]
        # los[2]['text']=los[2]['text'][3:-4]
        # los[3]['text']=los[3]['text'][3:-4]
        # los[-1]['text']='<b>'+los[-1]['text']+'</b>'
        # los[-2]['text']='<b>'+los[-2]['text']+'</b>'
        # los[-3]['text']='<b>'+los[-3]['text']+'</b>'
        # los[-4]['text']='<b>'+los[-4]['text']+'</b>'
        # sp['layout']['annotations']=los
        # print(sp['layout']['annotations'])
        # sp['layout']['annotations'][0]['yanchor']='top'

        cf.iplot(sp)

plotly_net_values(filename)

使用plotly.express画图

Source code in pure_ocean_breeze/labor/process.py
Python
2369
2370
2371
2372
2373
def plotly_net_values(self, filename):
    """使用plotly.express画图"""
    fig = pe.line(self.group_net_values)
    filename_path = filename + ".html"
    pio.write_html(fig, filename_path, auto_open=True)

prerpare() cached classmethod

通用数据准备

Source code in pure_ocean_breeze/labor/process.py
Python
2375
2376
2377
2378
2379
2380
@classmethod
@lru_cache(maxsize=None)
def prerpare(cls):
    """通用数据准备"""
    cls.judge_month()
    cls.get_rets_month()

read_add(pridf, df, func) classmethod

由于数据更新，过去计算的月度状态可能需要追加

Source code in pure_ocean_breeze/labor/process.py
Python
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
@classmethod
def read_add(cls, pridf, df, func):
    """由于数据更新，过去计算的月度状态可能需要追加"""
    if pridf.index.max() > df.index.max():
        df_add = pridf[pridf.index > df.index.max()]
        if df_add.shape[0] > int(cls.freq_ctrl.days_in / 2):
            df_1 = df_add.index.max()
            year = df_1.year
            month = df_1.month
            last = tt.date.get_close(year=year, m=month).pd_date()
            if (last == df_1)[0]:
                # week_here
                df_add = df_add.resample(cls.freq).apply(func)
                df = pd.concat([df, df_add])
                return df
            else:
                df_add = df_add[
                    df_add.index < pd.Timestamp(year=year, month=month, day=1)
                ]
                if df_add.shape[0] > 0:
                    df_add = df_add.resample(cls.freq).apply(func)
                    df = pd.concat([df, df_add])
                    return df
                else:
                    return df
        else:
            return df
    else:
        return df

run(groups_num=10, neutralize=False, boxcox=False, trade_cost_double_side=0, value_weighted=False, y2=False, plt_plot=True, plotly_plot=False, filename='分组净值图', print_comments=True, comments_writer=None, net_values_writer=None, rets_writer=None, comments_sheetname=None, net_values_sheetname=None, rets_sheetname=None, on_paper=False, sheetname=None, zxindustry_dummies=0, swindustry_dummies=0, only_cap=0, iplot=1, ilegend=0, without_breakpoint=0, beauty_comments=0)

运行回测部分

Source code in pure_ocean_breeze/labor/process.py
Python
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
def run(
    self,
    groups_num=10,
    neutralize=False,
    boxcox=False,
    trade_cost_double_side=0,
    value_weighted=False,
    y2=False,
    plt_plot=True,
    plotly_plot=False,
    filename="分组净值图",
    print_comments=True,
    comments_writer=None,
    net_values_writer=None,
    rets_writer=None,
    comments_sheetname=None,
    net_values_sheetname=None,
    rets_sheetname=None,
    on_paper=False,
    sheetname=None,
    zxindustry_dummies=0,
    swindustry_dummies=0,
    only_cap=0,
    iplot=1,
    ilegend=0,
    without_breakpoint=0,
    beauty_comments=0,
):
    """运行回测部分"""
    if comments_writer and not (comments_sheetname or sheetname):
        raise IOError("把total_comments输出到excel中时，必须指定sheetname🤒")
    if net_values_writer and not (net_values_sheetname or sheetname):
        raise IOError("把group_net_values输出到excel中时，必须指定sheetname🤒")
    if rets_writer and not (rets_sheetname or sheetname):
        raise IOError("把group_rets输出到excel中时，必须指定sheetname🤒")
    if neutralize:
        self.get_log_cap()
        self.get_neutral_factors(
            swindustry_dummies=swindustry_dummies,
            zxindustry_dummies=zxindustry_dummies,
        )
        self.deal_with_factors_after_neutralize()
    elif boxcox:
        self.get_log_cap(boxcox=True)
        self.get_neutral_factors(
            swindustry_dummies=swindustry_dummies,
            zxindustry_dummies=zxindustry_dummies,
            only_cap=only_cap,
        )
        self.deal_with_factors_after_neutralize()
    else:
        self.deal_with_factors()
    self.get_limit_ups_downs()
    self.get_data(groups_num)
    self.get_group_rets_net_values(
        groups_num=groups_num,
        value_weighted=value_weighted,
        trade_cost_double_side=trade_cost_double_side,
    )
    self.get_long_short_comments(on_paper=on_paper)
    self.get_total_comments(groups_num=groups_num)
    if on_paper:
        group1_ttest = ss.ttest_1samp(self.group_rets.group1, 0).pvalue
        group10_ttest = ss.ttest_1samp(
            self.group_rets[f"group{groups_num}"], 0
        ).pvalue
        group_long_short_ttest = ss.ttest_1samp(self.long_short_rets, 0).pvalue
        group1_ret = self.group_rets.group1.mean()
        group10_ret = self.group_rets[f"group{groups_num}"].mean()
        group_long_short_ret = self.long_short_rets.mean()
        papers = pd.DataFrame(
            {
                "评价指标": [
                    group1_ttest,
                    group10_ttest,
                    group_long_short_ttest,
                    group1_ret,
                    group10_ret,
                    group_long_short_ret,
                ]
            },
            index=[
                "分组1p值",
                f"分组{groups_num}p值",
                f"分组1-分组{groups_num}p值",
                "分组1收益率",
                f"分组{groups_num}收益率",
                f"分组1-分组{groups_num}收益率",
            ],
        )
        self.total_comments = pd.concat([papers, self.total_comments])

    if plt_plot:
        if not STATES["NO_PLOT"]:
            if filename:
                self.plot_net_values(
                    y2=y2,
                    filename=filename,
                    iplot=iplot,
                    ilegend=bool(ilegend),
                    without_breakpoint=without_breakpoint,
                )
            else:
                self.plot_net_values(
                    y2=y2,
                    filename=self.factors_file.split(".")[-2].split("/")[-1]
                    + str(groups_num)
                    + "分组",
                    iplot=iplot,
                    ilegend=bool(ilegend),
                    without_breakpoint=without_breakpoint,
                )
            plt.show()
    if plotly_plot:
        if not STATES["NO_PLOT"]:
            if filename:
                self.plotly_net_values(filename=filename)
            else:
                self.plotly_net_values(
                    filename=self.factors_file.split(".")[-2].split("/")[-1]
                    + str(groups_num)
                    + "分组"
                )
    if print_comments:
        if not STATES["NO_COMMENT"]:
            tb = Texttable()
            tb.set_cols_width(
                [8] * 5 + [9] + [8] * 2 + [7] * 2 + [8] + [8] + [9] + [10] * 5
            )
            tb.set_cols_dtype(["f"] * 18)
            tb.header(list(self.total_comments.T.columns))
            tb.add_rows(self.total_comments.T.to_numpy(), header=False)
            print(tb.draw())
    if sheetname:
        if comments_writer:
            if not on_paper:
                total_comments = self.total_comments.copy()
                tc = list(total_comments.评价指标)
                if beauty_comments:
                    tc[0] = str(round(tc[0] * 100, 2)) + "%"
                    tc[1] = str(round(tc[1], 2))
                    tc[2] = str(round(tc[2] * 100, 2)) + "%"
                    tc[3] = str(round(tc[3], 2))
                    tc[4] = str(round(tc[4], 2))
                    tc[5] = str(round(tc[5] * 100, 2)) + "%"
                    tc[6] = str(round(tc[6] * 100, 2)) + "%"
                    tc[7] = str(round(tc[7] * 100, 2)) + "%"
                    tc[8] = str(round(tc[8], 2))
                    tc[9] = str(round(tc[9] * 100, 2)) + "%"
                    tc[10] = str(round(tc[10] * 100, 2)) + "%"
                    tc[11] = str(round(tc[11] * 100, 2)) + "%"
                    tc[12] = str(round(tc[12] * 100, 2)) + "%"
                    tc[13] = str(round(tc[13] * 100, 2)) + "%"
                    tc[14] = str(round(tc[14], 2))
                    tc[15] = str(round(tc[15], 2))
                    tc[16] = str(round(tc[16] * 100, 2)) + "%"
                    tc[17] = str(round(tc[17] * 100, 2)) + "%"
                tc=tc+list(self.group_mean_rets_monthly)
                new_total_comments = pd.DataFrame(
                    {sheetname: tc}, index=list(total_comments.index)+[f'第{i}组' for i in range(1,groups_num+1)]
                )
                new_total_comments.to_excel(comments_writer, sheet_name=sheetname)
                rankic_twins=pd.concat([self.rankics.rankic,self.rankics.rankic.cumsum()],axis=1)
                rankic_twins.columns=['RankIC','RankIC累积']
                rankic_twins.to_excel(comments_writer,sheet_name=sheetname+'RankIC')
            else:
                self.total_comments.rename(columns={"评价指标": sheetname}).to_excel(
                    comments_writer, sheet_name=sheetname
                )
        if net_values_writer:
            groups_net_values = self.group_net_values.copy()
            groups_net_values.index = groups_net_values.index.strftime("%Y/%m/%d")
            groups_net_values.columns = [
                f"分组{i}" for i in range(1, len(list(groups_net_values.columns)))
            ] + ["多空对冲（右轴）"]
            groups_net_values.to_excel(net_values_writer, sheet_name=sheetname)
        if rets_writer:
            group_rets = self.group_rets.copy()
            group_rets.index = group_rets.index.strftime("%Y/%m/%d")
            group_rets.columns = [
                f"分组{i}" for i in range(1, len(list(group_rets.columns)))
            ] + ["多空对冲（右轴）"]
            group_rets.to_excel(rets_writer, sheet_name=sheetname)
    else:
        if comments_writer and comments_sheetname:
            total_comments = self.total_comments.copy()
            tc = list(total_comments.评价指标)
            tc[0] = str(round(tc[0] * 100, 2)) + "%"
            tc[1] = str(round(tc[1], 2))
            tc[2] = str(round(tc[2] * 100, 2)) + "%"
            tc[3] = str(round(tc[3], 2))
            tc[4] = str(round(tc[4], 2))
            tc[5] = str(round(tc[5] * 100, 2)) + "%"
            tc[6] = str(round(tc[6] * 100, 2)) + "%"
            tc[7] = str(round(tc[7] * 100, 2)) + "%"
            tc[8] = str(round(tc[8], 2))
            tc[9] = str(round(tc[9] * 100, 2)) + "%"
            tc[10] = str(round(tc[10] * 100, 2)) + "%"
            tc[11] = str(round(tc[11] * 100, 2)) + "%"
            tc[12] = str(round(tc[12] * 100, 2)) + "%"
            tc[13] = str(round(tc[13] * 100, 2)) + "%"
            tc[14] = str(round(tc[14], 2))
            tc[15] = str(round(tc[15], 2))
            tc[16] = str(round(tc[16] * 100, 2)) + "%"
            tc[17] = str(round(tc[17] * 100, 2)) + "%"
            new_total_comments = pd.DataFrame(
                {comments_sheetname: tc}, index=total_comments.index
            )
            new_total_comments.T.to_excel(
                comments_writer, sheet_name=comments_sheetname
            )
        if net_values_writer and net_values_sheetname:
            groups_net_values = self.group_net_values.copy()
            groups_net_values.index = groups_net_values.index.strftime("%Y/%m/%d")
            groups_net_values.columns = [
                f"分组{i}" for i in range(1, len(list(groups_net_values.columns)))
            ] + ["多空对冲（右轴）"]
            groups_net_values.to_excel(
                net_values_writer, sheet_name=net_values_sheetname
            )
        if rets_writer and rets_sheetname:
            group_rets = self.group_rets.copy()
            group_rets.index = group_rets.index.strftime("%Y/%m/%d")
            group_rets.columns = [
                f"分组{i}" for i in range(1, len(list(group_rets.columns)))
            ] + ["多空对冲（右轴）"]
            group_rets.to_excel(rets_writer, sheet_name=rets_sheetname)

set_factor_df_date_as_index(df)

设置因子数据的dataframe，因子表列名应为股票代码，索引应为时间

Source code in pure_ocean_breeze/labor/process.py
Python
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
def set_factor_df_date_as_index(self, df: pd.DataFrame):
    """设置因子数据的dataframe，因子表列名应为股票代码，索引应为时间"""
    # week_here
    self.factors = df.resample(self.freq).last().dropna(how="all")
    self.factor_cover = np.sign(self.factors.abs() + 1).sum().sum()
    opens = self.opens[self.opens.index >= self.factors.index.min()]
    total = np.sign(opens.resample(self.freq).last()).sum().sum()
    self.factor_cover = min(self.factor_cover / total, 1)
    self.factor_cross_skew = self.factors.skew(axis=1).mean()
    pos_num = ((self.factors > 0) + 0).sum().sum()
    neg_num = ((self.factors < 0) + 0).sum().sum()
    self.pos_neg_rate = pos_num / (neg_num + pos_num)
    self.corr_itself = show_corr(self.factors, self.factors.shift(1), plt_plot=0)
    self.corr_itself_shift2 = show_corr(
        self.factors, self.factors.shift(2), plt_plot=0
    )
    self.factor_cross_stds = self.factors.std(axis=1)

to_group_ret(l)

每一组的年化收益率

Source code in pure_ocean_breeze/labor/process.py
Python
2009
2010
2011
2012
2013
def to_group_ret(self, l):
    """每一组的年化收益率"""
    # week_here
    ret = l[-1] ** (self.freq_ctrl.counts_one_year / len(l)) - 1
    return ret

pure_moonnight

Bases: object

封装选股框架

Source code in pure_ocean_breeze/labor/process.py
Python
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
@do_on_dfs
class pure_moonnight(object):
    """封装选股框架"""

    __slots__ = ["shen"]

    def __init__(
        self,
        factors: pd.DataFrame,
        groups_num: int = 10,
        freq: str = "M",
        neutralize: bool = 0,
        boxcox: bool = 1,
        trade_cost_double_side: float = 0,
        value_weighted: bool = 0,
        y2: bool = 0,
        plt_plot: bool = 1,
        plotly_plot: bool = 0,
        filename: str = "分组净值图",
        time_start: int = None,
        time_end: int = None,
        print_comments: bool = 1,
        comments_writer: pd.ExcelWriter = None,
        net_values_writer: pd.ExcelWriter = None,
        rets_writer: pd.ExcelWriter = None,
        comments_sheetname: str = None,
        net_values_sheetname: str = None,
        rets_sheetname: str = None,
        on_paper: bool = 0,
        sheetname: str = None,
        zxindustry_dummies: bool = 0,
        swindustry_dummies: bool = 0,
        ages: pd.DataFrame = None,
        sts: pd.DataFrame = None,
        states: pd.DataFrame = None,
        opens: pd.DataFrame = None,
        closes: pd.DataFrame = None,
        capitals: pd.DataFrame = None,
        swindustry_dummy: pd.DataFrame = None,
        zxindustry_dummy: pd.DataFrame = None,
        no_read_indu: bool = 0,
        only_cap: bool = 0,
        iplot: bool = 1,
        ilegend: bool = 0,
        without_breakpoint: bool = 0,
        opens_average_first_day: bool = 0,
        total_cap: bool = 0,
    ) -> None:
        """一键回测框架，测试单因子的月频调仓的分组表现
        每月月底计算因子值，月初第一天开盘时买入，月末收盘最后一天收盘时卖出
        剔除上市不足60天的，停牌天数超过一半的，st天数超过一半的
        月末收盘跌停的不卖出，月初开盘涨停的不买入
        由最好组和最差组的多空组合构成多空对冲组

        Parameters
        ----------
        factors : pd.DataFrame
            要用于检测的因子值，index是时间，columns是股票代码
        groups_num : int, optional
            分组数量, by default 10
        freq : str, optional
            回测频率, by default 'M'
        neutralize : bool, optional
            对流通市值取自然对数，以完成行业市值中性化, by default 0
        boxcox : bool, optional
            对流通市值做截面boxcox变换，以完成行业市值中性化, by default 1
        trade_cost_double_side : float, optional
            交易的双边手续费率, by default 0
        value_weighted : bool, optional
            是否用流通市值加权, by default 0
        y2 : bool, optional
            画图时是否启用第二y轴, by default 0
        plt_plot : bool, optional
            将分组净值曲线用matplotlib画出来, by default 1
        plotly_plot : bool, optional
            将分组净值曲线用plotly画出来, by default 0
        filename : str, optional
            分组净值曲线的图保存的名称, by default "分组净值图"
        time_start : int, optional
            回测起始时间, by default None
        time_end : int, optional
            回测终止时间, by default None
        print_comments : bool, optional
            是否打印出评价指标, by default 1
        comments_writer : pd.ExcelWriter, optional
            用于记录评价指标的xlsx文件, by default None
        net_values_writer : pd.ExcelWriter, optional
            用于记录净值序列的xlsx文件, by default None
        rets_writer : pd.ExcelWriter, optional
            用于记录收益率序列的xlsx文件, by default None
        comments_sheetname : str, optional
            在记录评价指标的xlsx文件中，该工作表的名称, by default None
        net_values_sheetname : str, optional
            在记录净值序列的xlsx文件中，该工作表的名称, by default None
        rets_sheetname : str, optional
            在记录收益率序列的xlsx文件中，该工作表的名称, by default None
        on_paper : bool, optional
            使用学术化评价指标, by default 0
        sheetname : str, optional
            各个pd.Excelwriter中工作表的统一名称, by default None
        zxindustry_dummies : bool, optional
            行业中性化时，选用中信一级行业, by default 0
        swindustry_dummies : bool, optional
            行业中性化时，选用申万一级行业, by default 0
        ages : pd.DataFrame, optional
            输入股票上市天数的数据，index是时间，columns是股票代码，values是天数, by default None
        sts : pd.DataFrame, optional
            输入股票每天是否st的数据，是st股即为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None
        states : pd.DataFrame, optional
            输入股票每天交易状态的数据，正常交易为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None
        opens : pd.DataFrame, optional
            输入股票的复权开盘价数据，index是时间，columns是股票代码，values是价格, by default None
        closes : pd.DataFrame, optional
            输入股票的复权收盘价数据，index是时间，columns是股票代码，values是价格, by default None
        capitals : pd.DataFrame, optional
            输入股票的每月月末流通市值数据，index是时间，columns是股票代码，values是流通市值, by default None
        swindustry_dummy : pd.DataFrame, optioanl
            熟人股票的每月月末的申万一级行业哑变量，表包含33列，第一列为股票代码，名为`code`，第二列为月末最后一天的日期，名为`date`
            其余31列，为各个行业的哑变量，名为`w1`、`w2`、`w3`……`w31`, by default None
        zxindustry_dummy : pd.DataFrame, optioanl
            熟人股票的每月月末的中信一级行业哑变量，表包含32列，第一列为股票代码，名为`code`，第二列为月末最后一天的日期，名为`date`
            其余30列，为各个行业的哑变量，名为`w1`、`w2`、`w3`……`w30`, by default None
        no_read_indu : bool, optional
            不读入行业数据, by default 0
        only_cap : bool, optional
            仅做市值中性化, by default 0
        iplot : bool, optional
            使用cufflinks呈现回测结果, by default 1
        ilegend : bool, optional
            使用cufflinks绘图时，是否显示图例, by default 1
        without_breakpoint : bool, optional
            画图的时候是否去除间断点, by default 0
        opens_average_first_day : bool, optional
            买入时使用第一天的平均价格, by default 0
        total_cap : bool, optional
            加权和行业市值中性化时使用总市值, by default 0
        """

        if not isinstance(factors, pd.DataFrame):
            factors = factors()
        if comments_writer is None and sheetname is not None:
            from pure_ocean_breeze.state.states import COMMENTS_WRITER

            comments_writer = COMMENTS_WRITER
        if net_values_writer is None and sheetname is not None:
            from pure_ocean_breeze.state.states import NET_VALUES_WRITER

            net_values_writer = NET_VALUES_WRITER
        if not on_paper:
            from pure_ocean_breeze.state.states import ON_PAPER

            on_paper = ON_PAPER
        if time_start is None:
            from pure_ocean_breeze.state.states import MOON_START

            if MOON_START is not None:
                factors = factors[factors.index >= pd.Timestamp(str(MOON_START))]
        else:
            factors = factors[factors.index >= pd.Timestamp(str(time_start))]
        if time_end is None:
            from pure_ocean_breeze.state.states import MOON_END

            if MOON_END is not None:
                factors = factors[factors.index <= pd.Timestamp(str(MOON_END))]
        else:
            factors = factors[factors.index <= pd.Timestamp(str(time_end))]
        if boxcox + neutralize == 0:
            no_read_indu = 1
        if only_cap + no_read_indu > 0:
            only_cap = no_read_indu = 1
        if iplot:
            print_comments = 0
        if total_cap:
            if opens_average_first_day:
                if freq == "M":
                    self.shen = pure_moon_b(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
                elif freq == "W":
                    self.shen = pure_week_b(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
            else:
                if freq == "M":
                    self.shen = pure_moon_c(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
                elif freq == "W":
                    self.shen = pure_week_c(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
        else:
            if opens_average_first_day:
                if freq == "M":
                    self.shen = pure_moon(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
                elif freq == "W":
                    self.shen = pure_week(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
            else:
                if freq == "M":
                    self.shen = pure_moon_a(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
                elif freq == "W":
                    self.shen = pure_week_a(
                        freq=freq,
                        no_read_indu=no_read_indu,
                        swindustry_dummy=swindustry_dummy,
                        zxindustry_dummy=zxindustry_dummy,
                        read_in_swindustry_dummy=swindustry_dummies,
                    )
        self.shen.set_basic_data(
            ages=ages,
            sts=sts,
            states=states,
            opens=opens,
            closes=closes,
            capitals=capitals,
            opens_average_first_day=opens_average_first_day,
            total_cap=total_cap,
        )
        self.shen.set_factor_df_date_as_index(factors)
        self.shen.prerpare()
        self.shen.run(
            groups_num=groups_num,
            neutralize=neutralize,
            boxcox=boxcox,
            trade_cost_double_side=trade_cost_double_side,
            value_weighted=value_weighted,
            y2=y2,
            plt_plot=plt_plot,
            plotly_plot=plotly_plot,
            filename=filename,
            print_comments=print_comments,
            comments_writer=comments_writer,
            net_values_writer=net_values_writer,
            rets_writer=rets_writer,
            comments_sheetname=comments_sheetname,
            net_values_sheetname=net_values_sheetname,
            rets_sheetname=rets_sheetname,
            on_paper=on_paper,
            sheetname=sheetname,
            swindustry_dummies=swindustry_dummies,
            zxindustry_dummies=zxindustry_dummies,
            only_cap=only_cap,
            iplot=iplot,
            ilegend=ilegend,
            without_breakpoint=without_breakpoint,
        )

    def __call__(self) -> pd.DataFrame:
        """如果做了行业市值中性化，则返回行业市值中性化之后的因子数据

        Returns
        -------
        `pd.DataFrame`
            如果做了行业市值中性化，则行业市值中性化之后的因子数据，否则返回原因子数据
        """
        return self.shen.factors_out

    def comments_ten(self) -> pd.DataFrame:
        """对回测的十分组结果分别给出评价

        Returns
        -------
        `pd.DataFrame`
            评价指标包括年化收益率、总收益率、年化波动率、年化夏普比率、最大回撤率、胜率
        """
        rets_cols = list(self.shen.group_rets.columns)
        rets_cols = rets_cols[:-1]
        coms = []
        for i in rets_cols:
            ret = self.shen.group_rets[i]
            net = self.shen.group_net_values[i]
            com = comments_on_twins(net, ret)
            com = com.to_frame(i)
            coms.append(com)
        df = pd.concat(coms, axis=1)
        return df.T

    def comment_yearly(self) -> pd.DataFrame:
        """对回测的每年表现给出评价

        Returns
        -------
        pd.DataFrame
            各年度的收益率
        """
        df = self.shen.group_net_values.resample("Y").last().pct_change()
        df.index = df.index.year
        return df

__call__()

如果做了行业市值中性化，则返回行业市值中性化之后的因子数据

Returns

pd.DataFrame 如果做了行业市值中性化，则行业市值中性化之后的因子数据，否则返回原因子数据

Source code in pure_ocean_breeze/labor/process.py
Python
2892
2893
2894
2895
2896
2897
2898
2899
2900
def __call__(self) -> pd.DataFrame:
    """如果做了行业市值中性化，则返回行业市值中性化之后的因子数据

    Returns
    -------
    `pd.DataFrame`
        如果做了行业市值中性化，则行业市值中性化之后的因子数据，否则返回原因子数据
    """
    return self.shen.factors_out

__init__(factors, groups_num=10, freq='M', neutralize=0, boxcox=1, trade_cost_double_side=0, value_weighted=0, y2=0, plt_plot=1, plotly_plot=0, filename='分组净值图', time_start=None, time_end=None, print_comments=1, comments_writer=None, net_values_writer=None, rets_writer=None, comments_sheetname=None, net_values_sheetname=None, rets_sheetname=None, on_paper=0, sheetname=None, zxindustry_dummies=0, swindustry_dummies=0, ages=None, sts=None, states=None, opens=None, closes=None, capitals=None, swindustry_dummy=None, zxindustry_dummy=None, no_read_indu=0, only_cap=0, iplot=1, ilegend=0, without_breakpoint=0, opens_average_first_day=0, total_cap=0)

一键回测框架，测试单因子的月频调仓的分组表现 每月月底计算因子值，月初第一天开盘时买入，月末收盘最后一天收盘时卖出 剔除上市不足60天的，停牌天数超过一半的，st天数超过一半的 月末收盘跌停的不卖出，月初开盘涨停的不买入 由最好组和最差组的多空组合构成多空对冲组

Parameters

factors : pd.DataFrame 要用于检测的因子值，index是时间，columns是股票代码 groups_num : int, optional 分组数量, by default 10 freq : str, optional 回测频率, by default 'M' neutralize : bool, optional 对流通市值取自然对数，以完成行业市值中性化, by default 0 boxcox : bool, optional 对流通市值做截面boxcox变换，以完成行业市值中性化, by default 1 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 value_weighted : bool, optional 是否用流通市值加权, by default 0 y2 : bool, optional 画图时是否启用第二y轴, by default 0 plt_plot : bool, optional 将分组净值曲线用matplotlib画出来, by default 1 plotly_plot : bool, optional 将分组净值曲线用plotly画出来, by default 0 filename : str, optional 分组净值曲线的图保存的名称, by default "分组净值图" time_start : int, optional 回测起始时间, by default None time_end : int, optional 回测终止时间, by default None print_comments : bool, optional 是否打印出评价指标, by default 1 comments_writer : pd.ExcelWriter, optional 用于记录评价指标的xlsx文件, by default None net_values_writer : pd.ExcelWriter, optional 用于记录净值序列的xlsx文件, by default None rets_writer : pd.ExcelWriter, optional 用于记录收益率序列的xlsx文件, by default None comments_sheetname : str, optional 在记录评价指标的xlsx文件中，该工作表的名称, by default None net_values_sheetname : str, optional 在记录净值序列的xlsx文件中，该工作表的名称, by default None rets_sheetname : str, optional 在记录收益率序列的xlsx文件中，该工作表的名称, by default None on_paper : bool, optional 使用学术化评价指标, by default 0 sheetname : str, optional 各个pd.Excelwriter中工作表的统一名称, by default None zxindustry_dummies : bool, optional 行业中性化时，选用中信一级行业, by default 0 swindustry_dummies : bool, optional 行业中性化时，选用申万一级行业, by default 0 ages : pd.DataFrame, optional 输入股票上市天数的数据，index是时间，columns是股票代码，values是天数, by default None sts : pd.DataFrame, optional 输入股票每天是否st的数据，是st股即为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None states : pd.DataFrame, optional 输入股票每天交易状态的数据，正常交易为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None opens : pd.DataFrame, optional 输入股票的复权开盘价数据，index是时间，columns是股票代码，values是价格, by default None closes : pd.DataFrame, optional 输入股票的复权收盘价数据，index是时间，columns是股票代码，values是价格, by default None capitals : pd.DataFrame, optional 输入股票的每月月末流通市值数据，index是时间，columns是股票代码，values是流通市值, by default None swindustry_dummy : pd.DataFrame, optioanl 熟人股票的每月月末的申万一级行业哑变量，表包含33列，第一列为股票代码，名为code，第二列为月末最后一天的日期，名为date 其余31列，为各个行业的哑变量，名为w1、w2、w3……w31, by default None zxindustry_dummy : pd.DataFrame, optioanl 熟人股票的每月月末的中信一级行业哑变量，表包含32列，第一列为股票代码，名为code，第二列为月末最后一天的日期，名为date 其余30列，为各个行业的哑变量，名为w1、w2、w3……w30, by default None no_read_indu : bool, optional 不读入行业数据, by default 0 only_cap : bool, optional 仅做市值中性化, by default 0 iplot : bool, optional 使用cufflinks呈现回测结果, by default 1 ilegend : bool, optional 使用cufflinks绘图时，是否显示图例, by default 1 without_breakpoint : bool, optional 画图的时候是否去除间断点, by default 0 opens_average_first_day : bool, optional 买入时使用第一天的平均价格, by default 0 total_cap : bool, optional 加权和行业市值中性化时使用总市值, by default 0

Source code in pure_ocean_breeze/labor/process.py
Python
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
def __init__(
    self,
    factors: pd.DataFrame,
    groups_num: int = 10,
    freq: str = "M",
    neutralize: bool = 0,
    boxcox: bool = 1,
    trade_cost_double_side: float = 0,
    value_weighted: bool = 0,
    y2: bool = 0,
    plt_plot: bool = 1,
    plotly_plot: bool = 0,
    filename: str = "分组净值图",
    time_start: int = None,
    time_end: int = None,
    print_comments: bool = 1,
    comments_writer: pd.ExcelWriter = None,
    net_values_writer: pd.ExcelWriter = None,
    rets_writer: pd.ExcelWriter = None,
    comments_sheetname: str = None,
    net_values_sheetname: str = None,
    rets_sheetname: str = None,
    on_paper: bool = 0,
    sheetname: str = None,
    zxindustry_dummies: bool = 0,
    swindustry_dummies: bool = 0,
    ages: pd.DataFrame = None,
    sts: pd.DataFrame = None,
    states: pd.DataFrame = None,
    opens: pd.DataFrame = None,
    closes: pd.DataFrame = None,
    capitals: pd.DataFrame = None,
    swindustry_dummy: pd.DataFrame = None,
    zxindustry_dummy: pd.DataFrame = None,
    no_read_indu: bool = 0,
    only_cap: bool = 0,
    iplot: bool = 1,
    ilegend: bool = 0,
    without_breakpoint: bool = 0,
    opens_average_first_day: bool = 0,
    total_cap: bool = 0,
) -> None:
    """一键回测框架，测试单因子的月频调仓的分组表现
    每月月底计算因子值，月初第一天开盘时买入，月末收盘最后一天收盘时卖出
    剔除上市不足60天的，停牌天数超过一半的，st天数超过一半的
    月末收盘跌停的不卖出，月初开盘涨停的不买入
    由最好组和最差组的多空组合构成多空对冲组

    Parameters
    ----------
    factors : pd.DataFrame
        要用于检测的因子值，index是时间，columns是股票代码
    groups_num : int, optional
        分组数量, by default 10
    freq : str, optional
        回测频率, by default 'M'
    neutralize : bool, optional
        对流通市值取自然对数，以完成行业市值中性化, by default 0
    boxcox : bool, optional
        对流通市值做截面boxcox变换，以完成行业市值中性化, by default 1
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    value_weighted : bool, optional
        是否用流通市值加权, by default 0
    y2 : bool, optional
        画图时是否启用第二y轴, by default 0
    plt_plot : bool, optional
        将分组净值曲线用matplotlib画出来, by default 1
    plotly_plot : bool, optional
        将分组净值曲线用plotly画出来, by default 0
    filename : str, optional
        分组净值曲线的图保存的名称, by default "分组净值图"
    time_start : int, optional
        回测起始时间, by default None
    time_end : int, optional
        回测终止时间, by default None
    print_comments : bool, optional
        是否打印出评价指标, by default 1
    comments_writer : pd.ExcelWriter, optional
        用于记录评价指标的xlsx文件, by default None
    net_values_writer : pd.ExcelWriter, optional
        用于记录净值序列的xlsx文件, by default None
    rets_writer : pd.ExcelWriter, optional
        用于记录收益率序列的xlsx文件, by default None
    comments_sheetname : str, optional
        在记录评价指标的xlsx文件中，该工作表的名称, by default None
    net_values_sheetname : str, optional
        在记录净值序列的xlsx文件中，该工作表的名称, by default None
    rets_sheetname : str, optional
        在记录收益率序列的xlsx文件中，该工作表的名称, by default None
    on_paper : bool, optional
        使用学术化评价指标, by default 0
    sheetname : str, optional
        各个pd.Excelwriter中工作表的统一名称, by default None
    zxindustry_dummies : bool, optional
        行业中性化时，选用中信一级行业, by default 0
    swindustry_dummies : bool, optional
        行业中性化时，选用申万一级行业, by default 0
    ages : pd.DataFrame, optional
        输入股票上市天数的数据，index是时间，columns是股票代码，values是天数, by default None
    sts : pd.DataFrame, optional
        输入股票每天是否st的数据，是st股即为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None
    states : pd.DataFrame, optional
        输入股票每天交易状态的数据，正常交易为1，否则为0，index是时间，columns是股票代码，values是0或1, by default None
    opens : pd.DataFrame, optional
        输入股票的复权开盘价数据，index是时间，columns是股票代码，values是价格, by default None
    closes : pd.DataFrame, optional
        输入股票的复权收盘价数据，index是时间，columns是股票代码，values是价格, by default None
    capitals : pd.DataFrame, optional
        输入股票的每月月末流通市值数据，index是时间，columns是股票代码，values是流通市值, by default None
    swindustry_dummy : pd.DataFrame, optioanl
        熟人股票的每月月末的申万一级行业哑变量，表包含33列，第一列为股票代码，名为`code`，第二列为月末最后一天的日期，名为`date`
        其余31列，为各个行业的哑变量，名为`w1`、`w2`、`w3`……`w31`, by default None
    zxindustry_dummy : pd.DataFrame, optioanl
        熟人股票的每月月末的中信一级行业哑变量，表包含32列，第一列为股票代码，名为`code`，第二列为月末最后一天的日期，名为`date`
        其余30列，为各个行业的哑变量，名为`w1`、`w2`、`w3`……`w30`, by default None
    no_read_indu : bool, optional
        不读入行业数据, by default 0
    only_cap : bool, optional
        仅做市值中性化, by default 0
    iplot : bool, optional
        使用cufflinks呈现回测结果, by default 1
    ilegend : bool, optional
        使用cufflinks绘图时，是否显示图例, by default 1
    without_breakpoint : bool, optional
        画图的时候是否去除间断点, by default 0
    opens_average_first_day : bool, optional
        买入时使用第一天的平均价格, by default 0
    total_cap : bool, optional
        加权和行业市值中性化时使用总市值, by default 0
    """

    if not isinstance(factors, pd.DataFrame):
        factors = factors()
    if comments_writer is None and sheetname is not None:
        from pure_ocean_breeze.state.states import COMMENTS_WRITER

        comments_writer = COMMENTS_WRITER
    if net_values_writer is None and sheetname is not None:
        from pure_ocean_breeze.state.states import NET_VALUES_WRITER

        net_values_writer = NET_VALUES_WRITER
    if not on_paper:
        from pure_ocean_breeze.state.states import ON_PAPER

        on_paper = ON_PAPER
    if time_start is None:
        from pure_ocean_breeze.state.states import MOON_START

        if MOON_START is not None:
            factors = factors[factors.index >= pd.Timestamp(str(MOON_START))]
    else:
        factors = factors[factors.index >= pd.Timestamp(str(time_start))]
    if time_end is None:
        from pure_ocean_breeze.state.states import MOON_END

        if MOON_END is not None:
            factors = factors[factors.index <= pd.Timestamp(str(MOON_END))]
    else:
        factors = factors[factors.index <= pd.Timestamp(str(time_end))]
    if boxcox + neutralize == 0:
        no_read_indu = 1
    if only_cap + no_read_indu > 0:
        only_cap = no_read_indu = 1
    if iplot:
        print_comments = 0
    if total_cap:
        if opens_average_first_day:
            if freq == "M":
                self.shen = pure_moon_b(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
            elif freq == "W":
                self.shen = pure_week_b(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
        else:
            if freq == "M":
                self.shen = pure_moon_c(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
            elif freq == "W":
                self.shen = pure_week_c(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
    else:
        if opens_average_first_day:
            if freq == "M":
                self.shen = pure_moon(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
            elif freq == "W":
                self.shen = pure_week(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
        else:
            if freq == "M":
                self.shen = pure_moon_a(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
            elif freq == "W":
                self.shen = pure_week_a(
                    freq=freq,
                    no_read_indu=no_read_indu,
                    swindustry_dummy=swindustry_dummy,
                    zxindustry_dummy=zxindustry_dummy,
                    read_in_swindustry_dummy=swindustry_dummies,
                )
    self.shen.set_basic_data(
        ages=ages,
        sts=sts,
        states=states,
        opens=opens,
        closes=closes,
        capitals=capitals,
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
    )
    self.shen.set_factor_df_date_as_index(factors)
    self.shen.prerpare()
    self.shen.run(
        groups_num=groups_num,
        neutralize=neutralize,
        boxcox=boxcox,
        trade_cost_double_side=trade_cost_double_side,
        value_weighted=value_weighted,
        y2=y2,
        plt_plot=plt_plot,
        plotly_plot=plotly_plot,
        filename=filename,
        print_comments=print_comments,
        comments_writer=comments_writer,
        net_values_writer=net_values_writer,
        rets_writer=rets_writer,
        comments_sheetname=comments_sheetname,
        net_values_sheetname=net_values_sheetname,
        rets_sheetname=rets_sheetname,
        on_paper=on_paper,
        sheetname=sheetname,
        swindustry_dummies=swindustry_dummies,
        zxindustry_dummies=zxindustry_dummies,
        only_cap=only_cap,
        iplot=iplot,
        ilegend=ilegend,
        without_breakpoint=without_breakpoint,
    )

comment_yearly()

对回测的每年表现给出评价

Returns

pd.DataFrame 各年度的收益率

Source code in pure_ocean_breeze/labor/process.py
Python
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
def comment_yearly(self) -> pd.DataFrame:
    """对回测的每年表现给出评价

    Returns
    -------
    pd.DataFrame
        各年度的收益率
    """
    df = self.shen.group_net_values.resample("Y").last().pct_change()
    df.index = df.index.year
    return df

comments_ten()

对回测的十分组结果分别给出评价

Returns

pd.DataFrame 评价指标包括年化收益率、总收益率、年化波动率、年化夏普比率、最大回撤率、胜率

Source code in pure_ocean_breeze/labor/process.py
Python
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
def comments_ten(self) -> pd.DataFrame:
    """对回测的十分组结果分别给出评价

    Returns
    -------
    `pd.DataFrame`
        评价指标包括年化收益率、总收益率、年化波动率、年化夏普比率、最大回撤率、胜率
    """
    rets_cols = list(self.shen.group_rets.columns)
    rets_cols = rets_cols[:-1]
    coms = []
    for i in rets_cols:
        ret = self.shen.group_rets[i]
        net = self.shen.group_net_values[i]
        com = comments_on_twins(net, ret)
        com = com.to_frame(i)
        coms.append(com)
    df = pd.concat(coms, axis=1)
    return df.T

pure_newyear

Bases: object

转为生成25分组和百分组的收益矩阵而封装

Source code in pure_ocean_breeze/labor/process.py
Python
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
class pure_newyear(object):
    """转为生成25分组和百分组的收益矩阵而封装"""

    def __init__(
        self,
        facx: pd.DataFrame,
        facy: pd.DataFrame,
        group_num_single: int,
        trade_cost_double_side: float = 0,
        namex: str = "主",
        namey: str = "次",
    ) -> None:
        """条件双变量排序法，先对所有股票，依照因子facx进行排序
        然后在每个组内，依照facy进行排序，最后统计各个组内的平均收益率

        Parameters
        ----------
        facx : pd.DataFrame
            首先进行排序的因子，通常为控制变量，相当于正交化中的自变量
            index为时间，columns为股票代码，values为因子值
        facy : pd.DataFrame
            在facx的各个组内，依照facy进行排序，为主要要研究的因子
            index为时间，columns为股票代码，values为因子值
        group_num_single : int
            单个因子分成几组，通常为5或10
        trade_cost_double_side : float, optional
            交易的双边手续费率, by default 0
        namex : str, optional
            facx这一因子的名字, by default "主"
        namey : str, optional
            facy这一因子的名字, by default "次"
        """
        homex = pure_fallmount(facx)
        homey = pure_fallmount(facy)
        if group_num_single == 5:
            homexy = homex > homey
        elif group_num_single == 10:
            homexy = homex >> homey
        shen = pure_moonnight(
            homexy(),
            group_num_single**2,
            trade_cost_double_side=trade_cost_double_side,
            plt_plot=False,
            print_comments=False,
        )
        sq = shen.shen.square_rets.copy()
        sq.index = [namex + str(i) for i in list(sq.index)]
        sq.columns = [namey + str(i) for i in list(sq.columns)]
        self.square_rets = sq

    def __call__(self) -> pd.DataFrame:
        """调用对象时，返回最终结果，正方形的分组年化收益率表

        Returns
        -------
        `pd.DataFrame`
            每个组的年化收益率
        """
        return self.square_rets.copy()

__call__()

调用对象时，返回最终结果，正方形的分组年化收益率表

Returns

pd.DataFrame 每个组的年化收益率

Source code in pure_ocean_breeze/labor/process.py
Python
4068
4069
4070
4071
4072
4073
4074
4075
4076
def __call__(self) -> pd.DataFrame:
    """调用对象时，返回最终结果，正方形的分组年化收益率表

    Returns
    -------
    `pd.DataFrame`
        每个组的年化收益率
    """
    return self.square_rets.copy()

__init__(facx, facy, group_num_single, trade_cost_double_side=0, namex='主', namey='次')

条件双变量排序法，先对所有股票，依照因子facx进行排序 然后在每个组内，依照facy进行排序，最后统计各个组内的平均收益率

Parameters

facx : pd.DataFrame 首先进行排序的因子，通常为控制变量，相当于正交化中的自变量 index为时间，columns为股票代码，values为因子值 facy : pd.DataFrame 在facx的各个组内，依照facy进行排序，为主要要研究的因子 index为时间，columns为股票代码，values为因子值 group_num_single : int 单个因子分成几组，通常为5或10 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 namex : str, optional facx这一因子的名字, by default "主" namey : str, optional facy这一因子的名字, by default "次"

Source code in pure_ocean_breeze/labor/process.py
Python
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
def __init__(
    self,
    facx: pd.DataFrame,
    facy: pd.DataFrame,
    group_num_single: int,
    trade_cost_double_side: float = 0,
    namex: str = "主",
    namey: str = "次",
) -> None:
    """条件双变量排序法，先对所有股票，依照因子facx进行排序
    然后在每个组内，依照facy进行排序，最后统计各个组内的平均收益率

    Parameters
    ----------
    facx : pd.DataFrame
        首先进行排序的因子，通常为控制变量，相当于正交化中的自变量
        index为时间，columns为股票代码，values为因子值
    facy : pd.DataFrame
        在facx的各个组内，依照facy进行排序，为主要要研究的因子
        index为时间，columns为股票代码，values为因子值
    group_num_single : int
        单个因子分成几组，通常为5或10
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    namex : str, optional
        facx这一因子的名字, by default "主"
    namey : str, optional
        facy这一因子的名字, by default "次"
    """
    homex = pure_fallmount(facx)
    homey = pure_fallmount(facy)
    if group_num_single == 5:
        homexy = homex > homey
    elif group_num_single == 10:
        homexy = homex >> homey
    shen = pure_moonnight(
        homexy(),
        group_num_single**2,
        trade_cost_double_side=trade_cost_double_side,
        plt_plot=False,
        print_comments=False,
    )
    sq = shen.shen.square_rets.copy()
    sq.index = [namex + str(i) for i in list(sq.index)]
    sq.columns = [namey + str(i) for i in list(sq.columns)]
    self.square_rets = sq

pure_rollingols

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
class pure_rollingols(object):
    def __init__(
        self,
        y: pd.DataFrame,
        xs: Union[List[pd.DataFrame], pd.DataFrame],
        backsee: int = 20,
        factors_names: List[str] = None,
    ) -> None:
        """使用若干个dataframe，对应的股票进行指定窗口的时序滚动回归

        Parameters
        ----------
        y : pd.DataFrame
            滚动回归中的因变量y，index是时间，columns是股票代码
        xs : Union[List[pd.DataFrame], pd.DataFrame]
            滚动回归中的自变量xi，每一个dataframe，index是时间，columns是股票代码
        backsee : int, optional
            滚动回归的时间窗口, by default 20
        factors_names : List[str], optional
            xs中，每个因子的名字, by default None
        """
        self.backsee = backsee
        self.y = y
        if not isinstance(xs, list):
            xs = [xs]
        self.xs = xs
        y = y.stack().reset_index()
        xs = [i.stack().reset_index() for i in xs]
        y.columns = ["date", "code", "y"]
        xs = [
            i.rename(
                columns={list(i.columns)[1]: "code", list(i.columns)[2]: f"x{j+1}"}
            )
            for j, i in enumerate(xs)
        ]
        xs = [y] + xs
        xs = reduce(lambda x, y: pd.merge(x, y, on=["date", "code"]), xs)
        xs = xs.set_index("date")
        self.__data = xs
        self.haha = xs
        tqdm.auto.tqdm.pandas()
        self.__coefficients = (
            self.__data.groupby("code").progress_apply(self.ols_in).reset_index()
        )
        self.__coefficients = self.__coefficients.rename(
            columns={i: "co" + i for i in list(self.__coefficients.columns) if "x" in i}
        )
        self.__data = pd.merge(
            self.__data.reset_index(), self.__coefficients, on=["date", "code"]
        )
        betas = [
            self.__data[i] * self.__data["co" + i]
            for i in list(self.__data.columns)
            if i.startswith("x")
        ]
        betas = sum(betas)
        self.__data = self.__data.assign(
            residual=self.__data.y - self.__data.intercept - betas
        )
        self.__residual = self.__data.pivot(
            index="date", columns="code", values="residual"
        )
        self.__alphas = self.__data.pivot(
            index="date", columns="code", values="intercept"
        )
        if factors_names is None:
            self.__betas = {
                i: self.__data.pivot(index="date", columns="code", values=i)
                for i in list(self.__data.columns)
                if i.startswith("cox")
            }
        else:
            facs = [i for i in list(self.__data.columns) if i.startswith("cox")]
            self.__betas = {
                factors_names[num]: self.__data.pivot(
                    index="date", columns="code", values=i
                )
                for num, i in enumerate(facs)
            }
        if len(list(self.__betas)) == 1:
            self.__betas = list(self.__betas.values())[0]

    @property
    def residual(self):
        return self.__residual

    @property
    def data(self):
        return self.__data

    @property
    def alphas(self):
        return self.__alphas

    @property
    def betas(self):
        return self.__betas

    @property
    def coefficients(self):
        return self.__coefficients

    def ols_in(self, df):
        try:
            ols = po.PandasRollingOLS(
                y=df[["y"]],
                x=df[[f"x{i+1}" for i in range(len(self.xs))]],
                window=self.backsee,
            )
            betas = ols.beta
            alpha = ols.alpha
            return pd.concat([alpha, betas], axis=1)

        except Exception:
            # 有些数据总共不足，那就跳过
            ...

__init__(y, xs, backsee=20, factors_names=None)

使用若干个dataframe，对应的股票进行指定窗口的时序滚动回归

Parameters

y : pd.DataFrame 滚动回归中的因变量y，index是时间，columns是股票代码 xs : Union[List[pd.DataFrame], pd.DataFrame] 滚动回归中的自变量xi，每一个dataframe，index是时间，columns是股票代码 backsee : int, optional 滚动回归的时间窗口, by default 20 factors_names : List[str], optional xs中，每个因子的名字, by default None

Source code in pure_ocean_breeze/labor/process.py
Python
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
def __init__(
    self,
    y: pd.DataFrame,
    xs: Union[List[pd.DataFrame], pd.DataFrame],
    backsee: int = 20,
    factors_names: List[str] = None,
) -> None:
    """使用若干个dataframe，对应的股票进行指定窗口的时序滚动回归

    Parameters
    ----------
    y : pd.DataFrame
        滚动回归中的因变量y，index是时间，columns是股票代码
    xs : Union[List[pd.DataFrame], pd.DataFrame]
        滚动回归中的自变量xi，每一个dataframe，index是时间，columns是股票代码
    backsee : int, optional
        滚动回归的时间窗口, by default 20
    factors_names : List[str], optional
        xs中，每个因子的名字, by default None
    """
    self.backsee = backsee
    self.y = y
    if not isinstance(xs, list):
        xs = [xs]
    self.xs = xs
    y = y.stack().reset_index()
    xs = [i.stack().reset_index() for i in xs]
    y.columns = ["date", "code", "y"]
    xs = [
        i.rename(
            columns={list(i.columns)[1]: "code", list(i.columns)[2]: f"x{j+1}"}
        )
        for j, i in enumerate(xs)
    ]
    xs = [y] + xs
    xs = reduce(lambda x, y: pd.merge(x, y, on=["date", "code"]), xs)
    xs = xs.set_index("date")
    self.__data = xs
    self.haha = xs
    tqdm.auto.tqdm.pandas()
    self.__coefficients = (
        self.__data.groupby("code").progress_apply(self.ols_in).reset_index()
    )
    self.__coefficients = self.__coefficients.rename(
        columns={i: "co" + i for i in list(self.__coefficients.columns) if "x" in i}
    )
    self.__data = pd.merge(
        self.__data.reset_index(), self.__coefficients, on=["date", "code"]
    )
    betas = [
        self.__data[i] * self.__data["co" + i]
        for i in list(self.__data.columns)
        if i.startswith("x")
    ]
    betas = sum(betas)
    self.__data = self.__data.assign(
        residual=self.__data.y - self.__data.intercept - betas
    )
    self.__residual = self.__data.pivot(
        index="date", columns="code", values="residual"
    )
    self.__alphas = self.__data.pivot(
        index="date", columns="code", values="intercept"
    )
    if factors_names is None:
        self.__betas = {
            i: self.__data.pivot(index="date", columns="code", values=i)
            for i in list(self.__data.columns)
            if i.startswith("cox")
        }
    else:
        facs = [i for i in list(self.__data.columns) if i.startswith("cox")]
        self.__betas = {
            factors_names[num]: self.__data.pivot(
                index="date", columns="code", values=i
            )
            for num, i in enumerate(facs)
        }
    if len(list(self.__betas)) == 1:
        self.__betas = list(self.__betas.values())[0]

pure_snowtrain

Bases: object

直接返回纯净因子

Source code in pure_ocean_breeze/labor/process.py
Python
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
@do_on_dfs
class pure_snowtrain(object):
    """直接返回纯净因子"""

    def __init__(
        self,
        factors: pd.DataFrame,
        facs_dict: Dict = None,
        momentum: bool = 1,
        earningsyield: bool = 1,
        growth: bool = 1,
        liquidity: bool = 1,
        size: bool = 1,
        leverage: bool = 1,
        beta: bool = 1,
        nonlinearsize: bool = 1,
        residualvolatility: bool = 1,
        booktoprice: bool = 1,
    ) -> None:
        """计算因子值与10种常用风格因子之间的相关性，并进行纯净化，可以额外加入其他因子

        Parameters
        ----------
        factors : pd.DataFrame
            要考察的因子值，index为时间，columns为股票代码，values为因子值
        facs_dict : Dict, optional
            额外加入的因子，名字为key，因子矩阵为value，形如`{'反转': ret20, '换手': tr20}`, by default None
        momentum : bool, optional
            是否删去动量因子, by default 1
        earningsyield : bool, optional
            是否删去盈利因子, by default 1
        growth : bool, optional
            是否删去成长因子, by default 1
        liquidity : bool, optional
            是否删去流动性因子, by default 1
        size : bool, optional
            是否删去规模因子, by default 1
        leverage : bool, optional
            是否删去杠杆因子, by default 1
        beta : bool, optional
            是否删去贝塔因子, by default 1
        nonlinearsize : bool, optional
            是否删去非线性市值因子, by default 1
        residualvolatility : bool, optional
            是否删去残差波动率因子, by default 1
        booktoprice : bool, optional
            是否删去账面市值比因子, by default 1
        """
        self.winter = pure_coldwinter(
            momentum=momentum,
            earningsyield=earningsyield,
            growth=growth,
            liquidity=liquidity,
            size=size,
            leverage=leverage,
            beta=beta,
            nonlinearsize=nonlinearsize,
            residualvolatility=residualvolatility,
            booktoprice=booktoprice,
        )
        self.winter.set_factors_df_wide(factors, facs_dict)
        self.winter.get_snow_fac()
        self.corr = self.winter.corr

    def __call__(self) -> pd.DataFrame:
        """获得纯净化之后的因子值

        Returns
        -------
        pd.DataFrame
            纯净化之后的因子值
        """
        return self.winter.snow_fac.copy()

    def show_corr(self) -> pd.DataFrame:
        """展示因子与barra风格因子的相关系数

        Returns
        -------
        pd.DataFrame
            相关系数表格
        """
        return self.corr.applymap(lambda x: to_percent(x))

__call__()

获得纯净化之后的因子值

Returns

pd.DataFrame 纯净化之后的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
3997
3998
3999
4000
4001
4002
4003
4004
4005
def __call__(self) -> pd.DataFrame:
    """获得纯净化之后的因子值

    Returns
    -------
    pd.DataFrame
        纯净化之后的因子值
    """
    return self.winter.snow_fac.copy()

__init__(factors, facs_dict=None, momentum=1, earningsyield=1, growth=1, liquidity=1, size=1, leverage=1, beta=1, nonlinearsize=1, residualvolatility=1, booktoprice=1)

计算因子值与10种常用风格因子之间的相关性，并进行纯净化，可以额外加入其他因子

Parameters

factors : pd.DataFrame 要考察的因子值，index为时间，columns为股票代码，values为因子值 facs_dict : Dict, optional 额外加入的因子，名字为key，因子矩阵为value，形如{'反转': ret20, '换手': tr20}, by default None momentum : bool, optional 是否删去动量因子, by default 1 earningsyield : bool, optional 是否删去盈利因子, by default 1 growth : bool, optional 是否删去成长因子, by default 1 liquidity : bool, optional 是否删去流动性因子, by default 1 size : bool, optional 是否删去规模因子, by default 1 leverage : bool, optional 是否删去杠杆因子, by default 1 beta : bool, optional 是否删去贝塔因子, by default 1 nonlinearsize : bool, optional 是否删去非线性市值因子, by default 1 residualvolatility : bool, optional 是否删去残差波动率因子, by default 1 booktoprice : bool, optional 是否删去账面市值比因子, by default 1

Source code in pure_ocean_breeze/labor/process.py
Python
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
def __init__(
    self,
    factors: pd.DataFrame,
    facs_dict: Dict = None,
    momentum: bool = 1,
    earningsyield: bool = 1,
    growth: bool = 1,
    liquidity: bool = 1,
    size: bool = 1,
    leverage: bool = 1,
    beta: bool = 1,
    nonlinearsize: bool = 1,
    residualvolatility: bool = 1,
    booktoprice: bool = 1,
) -> None:
    """计算因子值与10种常用风格因子之间的相关性，并进行纯净化，可以额外加入其他因子

    Parameters
    ----------
    factors : pd.DataFrame
        要考察的因子值，index为时间，columns为股票代码，values为因子值
    facs_dict : Dict, optional
        额外加入的因子，名字为key，因子矩阵为value，形如`{'反转': ret20, '换手': tr20}`, by default None
    momentum : bool, optional
        是否删去动量因子, by default 1
    earningsyield : bool, optional
        是否删去盈利因子, by default 1
    growth : bool, optional
        是否删去成长因子, by default 1
    liquidity : bool, optional
        是否删去流动性因子, by default 1
    size : bool, optional
        是否删去规模因子, by default 1
    leverage : bool, optional
        是否删去杠杆因子, by default 1
    beta : bool, optional
        是否删去贝塔因子, by default 1
    nonlinearsize : bool, optional
        是否删去非线性市值因子, by default 1
    residualvolatility : bool, optional
        是否删去残差波动率因子, by default 1
    booktoprice : bool, optional
        是否删去账面市值比因子, by default 1
    """
    self.winter = pure_coldwinter(
        momentum=momentum,
        earningsyield=earningsyield,
        growth=growth,
        liquidity=liquidity,
        size=size,
        leverage=leverage,
        beta=beta,
        nonlinearsize=nonlinearsize,
        residualvolatility=residualvolatility,
        booktoprice=booktoprice,
    )
    self.winter.set_factors_df_wide(factors, facs_dict)
    self.winter.get_snow_fac()
    self.corr = self.winter.corr

show_corr()

展示因子与barra风格因子的相关系数

Returns

pd.DataFrame 相关系数表格

Source code in pure_ocean_breeze/labor/process.py
Python
4007
4008
4009
4010
4011
4012
4013
4014
4015
def show_corr(self) -> pd.DataFrame:
    """展示因子与barra风格因子的相关系数

    Returns
    -------
    pd.DataFrame
        相关系数表格
    """
    return self.corr.applymap(lambda x: to_percent(x))

pure_star

Bases: object

Source code in pure_ocean_breeze/labor/process.py
Python
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
@do_on_dfs
class pure_star(object):
    def __init__(
        self,
        fac: pd.Series,
        code: str = None,
        price_opens: pd.Series = None,
        iplot: bool = 1,
        comments_writer: pd.ExcelWriter = None,
        net_values_writer: pd.ExcelWriter = None,
        sheetname: str = None,
        questdb_host: str = "127.0.0.1",
    ):
        """择时回测框架，输入仓位比例或信号值，依据信号买入对应的股票或指数，并考察绝对收益、超额收益和基准收益
        回测方式为，t日收盘时获得信号，t+1日开盘时以开盘价买入，t+2开盘时以开盘价卖出

        Parameters
        ----------
        fac : pd.Series
            仓位比例序列，或信号序列，输入信号序列时即为0和1，输入仓位比例时，将每一期的收益按照对应比例缩小
        code : str, optional
            回测的资产代码，可以为股票代码或基金代码, by default None
        price_opens : pd.Series, optional
            资产的开盘价序列, by default None
        iplot : bool, optional
            使用cufflinks呈现回测绩效和走势图, by default 1
        comments_writer : pd.ExcelWriter, optional
            绩效评价的存储文件, by default None
        net_values_writer : pd.ExcelWriter, optional
            净值序列的存储文件, by default None
        sheetname : str, optional
            存储文件的工作表的名字, by default None
        questdb_host: str, optional
            questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
        """
        if code is not None:
            x1 = code.split(".")[0]
            x2 = code.split(".")[1]
            if (x1[0] == "0" or x1[:2] == "30") and x2 == "SZ":
                kind = "stock"
            elif x1[0] == "6" and x2 == "SH":
                kind = "stock"
            else:
                kind = "index"
            self.kind = kind
            if kind == "index":
                qdb = Questdb(host=questdb_host)
                price_opens = qdb.get_data(
                    f"select date,num,close from minute_data_{kind} where code='{code}'"
                )
                price_opens = price_opens[price_opens.num == "1"]
                price_opens = price_opens.set_index("date").close
                price_opens.index = pd.to_datetime(price_opens.index, format="%Y%m%d")
            else:
                price_opens = read_daily(open=1)[code]
        price_opens = price_opens[price_opens.index >= fac.index.min()]
        self.price_opens = price_opens
        self.price_rets = price_opens.pct_change()
        self.fac = fac
        self.fac_rets = (self.fac.shift(2) * self.price_rets).dropna()
        self.ab_rets = (self.fac_rets - self.price_rets).dropna()
        self.price_rets = self.price_rets.dropna()
        self.fac_nets = (1 + self.fac_rets).cumprod()
        self.ab_nets = (1 + self.ab_rets).cumprod()
        self.price_nets = (1 + self.price_rets).cumprod()
        self.fac_nets = self.fac_nets / self.fac_nets.iloc[0]
        self.ab_nets = self.ab_nets / self.ab_nets.iloc[0]
        self.price_nets = self.price_nets / self.price_nets.iloc[0]
        self.fac_comments = comments_on_twins(self.fac_nets, self.fac_rets)
        self.ab_comments = comments_on_twins(self.ab_nets, self.ab_rets)
        self.price_comments = comments_on_twins(self.price_nets, self.price_rets)
        self.total_comments = pd.concat(
            [self.fac_comments, self.ab_comments, self.price_comments], axis=1
        )
        self.total_nets = pd.concat(
            [self.fac_nets, self.ab_nets, self.price_nets], axis=1
        )
        self.total_rets = pd.concat(
            [self.fac_rets, self.ab_rets, self.price_rets], axis=1
        )
        self.total_comments.columns = (
            self.total_nets.columns
        ) = self.total_rets.columns = ["因子绝对", "因子超额", "买入持有"]
        self.total_comments = np.around(self.total_comments, 3)
        self.iplot = iplot
        self.plot()
        if comments_writer is None and sheetname is not None:
            from pure_ocean_breeze.state.states import COMMENTS_WRITER

            comments_writer = COMMENTS_WRITER
            self.total_comments.to_excel(comments_writer, sheet_name=sheetname)
        if net_values_writer is None and sheetname is not None:
            from pure_ocean_breeze.state.states import NET_VALUES_WRITER

            net_values_writer = NET_VALUES_WRITER
            self.total_nets.to_excel(net_values_writer, sheet_name=sheetname)

    def plot(self):
        coms = self.total_comments.copy().reset_index()
        if self.iplot:
            figs = cf.figures(
                self.total_nets,
                [dict(kind="line", y=list(self.total_nets.columns))],
                asList=True,
            )
            coms = coms.rename(columns={list(coms)[0]: "绩效指标"})
            table = FF.create_table(coms.iloc[::-1])
            table.update_yaxes(matches=None)
            figs.append(table)
            figs = [figs[-1]] + figs[:-1]
            figs[1].update_layout(
                legend=dict(yanchor="top", y=0.99, xanchor="left", x=0.01)
            )
            base_layout = cf.tools.get_base_layout(figs)
            sp = cf.subplots(
                figs,
                shape=(2, 10),
                base_layout=base_layout,
                vertical_spacing=0.15,
                horizontal_spacing=0.03,
                shared_yaxes=False,
                specs=[
                    [
                        None,
                        {"rowspan": 2, "colspan": 3},
                        None,
                        None,
                        {"rowspan": 2, "colspan": 6},
                        None,
                        None,
                        None,
                        None,
                        None,
                    ],
                    [
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                    ],
                ],
            )
            sp["layout"].update(showlegend=True)
            cf.iplot(sp)
        else:
            self.total_nets.plot()
            plt.show()
            tb = Texttable()
            tb.set_cols_width([8] + [7] + [8] * 2 + [7] * 2 + [8])
            tb.set_cols_dtype(["f"] * 7)
            tb.header(list(coms.T.reset_index().columns))
            tb.add_rows(coms.T.reset_index().to_numpy(), header=True)
            print(tb.draw())

__init__(fac, code=None, price_opens=None, iplot=1, comments_writer=None, net_values_writer=None, sheetname=None, questdb_host='127.0.0.1')

择时回测框架，输入仓位比例或信号值，依据信号买入对应的股票或指数，并考察绝对收益、超额收益和基准收益 回测方式为，t日收盘时获得信号，t+1日开盘时以开盘价买入，t+2开盘时以开盘价卖出

Parameters

fac : pd.Series 仓位比例序列，或信号序列，输入信号序列时即为0和1，输入仓位比例时，将每一期的收益按照对应比例缩小 code : str, optional 回测的资产代码，可以为股票代码或基金代码, by default None price_opens : pd.Series, optional 资产的开盘价序列, by default None iplot : bool, optional 使用cufflinks呈现回测绩效和走势图, by default 1 comments_writer : pd.ExcelWriter, optional 绩效评价的存储文件, by default None net_values_writer : pd.ExcelWriter, optional 净值序列的存储文件, by default None sheetname : str, optional 存储文件的工作表的名字, by default None questdb_host: str, optional questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'

Source code in pure_ocean_breeze/labor/process.py
Python
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
def __init__(
    self,
    fac: pd.Series,
    code: str = None,
    price_opens: pd.Series = None,
    iplot: bool = 1,
    comments_writer: pd.ExcelWriter = None,
    net_values_writer: pd.ExcelWriter = None,
    sheetname: str = None,
    questdb_host: str = "127.0.0.1",
):
    """择时回测框架，输入仓位比例或信号值，依据信号买入对应的股票或指数，并考察绝对收益、超额收益和基准收益
    回测方式为，t日收盘时获得信号，t+1日开盘时以开盘价买入，t+2开盘时以开盘价卖出

    Parameters
    ----------
    fac : pd.Series
        仓位比例序列，或信号序列，输入信号序列时即为0和1，输入仓位比例时，将每一期的收益按照对应比例缩小
    code : str, optional
        回测的资产代码，可以为股票代码或基金代码, by default None
    price_opens : pd.Series, optional
        资产的开盘价序列, by default None
    iplot : bool, optional
        使用cufflinks呈现回测绩效和走势图, by default 1
    comments_writer : pd.ExcelWriter, optional
        绩效评价的存储文件, by default None
    net_values_writer : pd.ExcelWriter, optional
        净值序列的存储文件, by default None
    sheetname : str, optional
        存储文件的工作表的名字, by default None
    questdb_host: str, optional
        questdb的host，使用NAS时改为'192.168.1.3', by default '127.0.0.1'
    """
    if code is not None:
        x1 = code.split(".")[0]
        x2 = code.split(".")[1]
        if (x1[0] == "0" or x1[:2] == "30") and x2 == "SZ":
            kind = "stock"
        elif x1[0] == "6" and x2 == "SH":
            kind = "stock"
        else:
            kind = "index"
        self.kind = kind
        if kind == "index":
            qdb = Questdb(host=questdb_host)
            price_opens = qdb.get_data(
                f"select date,num,close from minute_data_{kind} where code='{code}'"
            )
            price_opens = price_opens[price_opens.num == "1"]
            price_opens = price_opens.set_index("date").close
            price_opens.index = pd.to_datetime(price_opens.index, format="%Y%m%d")
        else:
            price_opens = read_daily(open=1)[code]
    price_opens = price_opens[price_opens.index >= fac.index.min()]
    self.price_opens = price_opens
    self.price_rets = price_opens.pct_change()
    self.fac = fac
    self.fac_rets = (self.fac.shift(2) * self.price_rets).dropna()
    self.ab_rets = (self.fac_rets - self.price_rets).dropna()
    self.price_rets = self.price_rets.dropna()
    self.fac_nets = (1 + self.fac_rets).cumprod()
    self.ab_nets = (1 + self.ab_rets).cumprod()
    self.price_nets = (1 + self.price_rets).cumprod()
    self.fac_nets = self.fac_nets / self.fac_nets.iloc[0]
    self.ab_nets = self.ab_nets / self.ab_nets.iloc[0]
    self.price_nets = self.price_nets / self.price_nets.iloc[0]
    self.fac_comments = comments_on_twins(self.fac_nets, self.fac_rets)
    self.ab_comments = comments_on_twins(self.ab_nets, self.ab_rets)
    self.price_comments = comments_on_twins(self.price_nets, self.price_rets)
    self.total_comments = pd.concat(
        [self.fac_comments, self.ab_comments, self.price_comments], axis=1
    )
    self.total_nets = pd.concat(
        [self.fac_nets, self.ab_nets, self.price_nets], axis=1
    )
    self.total_rets = pd.concat(
        [self.fac_rets, self.ab_rets, self.price_rets], axis=1
    )
    self.total_comments.columns = (
        self.total_nets.columns
    ) = self.total_rets.columns = ["因子绝对", "因子超额", "买入持有"]
    self.total_comments = np.around(self.total_comments, 3)
    self.iplot = iplot
    self.plot()
    if comments_writer is None and sheetname is not None:
        from pure_ocean_breeze.state.states import COMMENTS_WRITER

        comments_writer = COMMENTS_WRITER
        self.total_comments.to_excel(comments_writer, sheet_name=sheetname)
    if net_values_writer is None and sheetname is not None:
        from pure_ocean_breeze.state.states import NET_VALUES_WRITER

        net_values_writer = NET_VALUES_WRITER
        self.total_nets.to_excel(net_values_writer, sheet_name=sheetname)

add_cross_standardlize(*args)

将众多因子横截面做z-score标准化之后相加

Returns

pd.DataFrame 合成后的因子

Source code in pure_ocean_breeze/labor/process.py
Python
907
908
909
910
911
912
913
914
915
916
917
918
919
920
@do_on_dfs
def add_cross_standardlize(*args: list) -> pd.DataFrame:
    """将众多因子横截面做z-score标准化之后相加

    Returns
    -------
    `pd.DataFrame`
        合成后的因子
    """
    fms = [pure_fallmount(i) for i in args]
    one = fms[0]
    others = fms[1:]
    final = one + others
    return final()

boom_four(df, backsee=20, daily=0, min_periods=None)

生成20天均值，20天标准差，及二者正向z-score合成，正向排序合成，负向z-score合成，负向排序合成这6个因子

Parameters

df : pd.DataFrame 原日频因子 backsee : int, optional 回看天数, by default 20 daily : bool, optional 为1是每天都滚动，为0则仅保留月底值, by default 0 min_periods : int, optional rolling时的最小期, by default backsee的一半

Returns

Tuple[pd.DataFrame] 6个因子的元组

Source code in pure_ocean_breeze/labor/process.py
Python
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
@do_on_dfs
def boom_four(
    df: pd.DataFrame, backsee: int = 20, daily: bool = 0, min_periods: int = None
) -> Tuple[pd.DataFrame]:
    """生成20天均值，20天标准差，及二者正向z-score合成，正向排序合成，负向z-score合成，负向排序合成这6个因子

    Parameters
    ----------
    df : pd.DataFrame
        原日频因子
    backsee : int, optional
        回看天数, by default 20
    daily : bool, optional
        为1是每天都滚动，为0则仅保留月底值, by default 0
    min_periods : int, optional
        rolling时的最小期, by default backsee的一半

    Returns
    -------
    `Tuple[pd.DataFrame]`
        6个因子的元组
    """
    if min_periods is None:
        min_periods = int(backsee * 0.5)
    if not daily:
        df_mean = (
            df.rolling(backsee, min_periods=min_periods).mean().resample("M").last()
        )
        df_std = df.rolling(backsee, min_periods=min_periods).std().resample("M").last()
        twins_add = (pure_fallmount(df_mean) + (pure_fallmount(df_std),))()
        rtwins_add = df_mean.rank(axis=1) + df_std.rank(axis=1)
        twins_minus = (pure_fallmount(df_mean) + (pure_fallmount(-df_std),))()
        rtwins_minus = df_mean.rank(axis=1) - df_std.rank(axis=1)
    else:
        df_mean = df.rolling(backsee, min_periods=min_periods).mean()
        df_std = df.rolling(backsee, min_periods=min_periods).std()
        twins_add = (pure_fallmount(df_mean) + (pure_fallmount(df_std),))()
        rtwins_add = df_mean.rank(axis=1) + df_std.rank(axis=1)
        twins_minus = (pure_fallmount(df_mean) + (pure_fallmount(-df_std),))()
        rtwins_minus = df_mean.rank(axis=1) - df_std.rank(axis=1)
    return df_mean, df_std, twins_add, rtwins_add, twins_minus, rtwins_minus

boom_fours(dfs, backsee=20, daily=0, min_periods=None)

对多个因子，每个因子都进行boom_four的操作

Parameters

dfs : List[pd.DataFrame] 多个因子的dataframe组成的list backsee : Union[int,List[int]], optional 每个因子回看期数, by default 20 daily : Union[bool,List[bool]], optional 每个因子是否逐日计算, by default 0 min_periods : Union[int,List[int]], optional 每个因子计算的最小期, by default None

Returns

List[List[pd.DataFrame]] 每个因子进行boom_four后的结果

Source code in pure_ocean_breeze/labor/process.py
Python
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
def boom_fours(
    dfs: List[pd.DataFrame],
    backsee: Union[int, List[int]] = 20,
    daily: Union[bool, List[bool]] = 0,
    min_periods: Union[int, List[int]] = None,
) -> List[List[pd.DataFrame]]:
    """对多个因子，每个因子都进行boom_four的操作

    Parameters
    ----------
    dfs : List[pd.DataFrame]
        多个因子的dataframe组成的list
    backsee : Union[int,List[int]], optional
        每个因子回看期数, by default 20
    daily : Union[bool,List[bool]], optional
        每个因子是否逐日计算, by default 0
    min_periods : Union[int,List[int]], optional
        每个因子计算的最小期, by default None

    Returns
    -------
    List[List[pd.DataFrame]]
        每个因子进行boom_four后的结果
    """
    return boom_four(df=dfs, backsee=backsee, daily=daily, min_periods=min_periods)

daily_factor_on300500(fac, hs300=0, zz500=0, zz1000=0, gz2000=0, other=0)

输入日频或月频因子值，将其限定在某指数成分股的股票池内， 目前仅支持沪深300、中证500、中证1000、国证2000成分股，以及这四种指数成分股的组合叠加，和除沪深300、中证500、中证1000以外的股票的成分股

Parameters

fac : pd.DataFrame 未限定股票池的因子值，index为时间，columns为股票代码 hs300 : bool, optional 限定股票池为沪深300, by default 0 zz500 : bool, optional 限定股票池为中证500, by default 0 zz1000 : bool, optional 限定股票池为中证1000, by default 0 gz2000 : bool, optional 限定股票池为国证2000, by default 0 other : bool, optional 限定股票池为除沪深300、中证500、中证1000以外的股票的成分股, by default 0

Returns

pd.DataFrame 仅包含成分股后的因子值，非成分股的因子值为空

Raises

ValueError 如果未指定任何一种指数的成分股，将报错

Source code in pure_ocean_breeze/labor/process.py
Python
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
@do_on_dfs
def daily_factor_on300500(
    fac: pd.DataFrame,
    hs300: bool = 0,
    zz500: bool = 0,
    zz1000: bool = 0,
    gz2000: bool = 0,
    other: bool = 0,
) -> pd.DataFrame:
    """输入日频或月频因子值，将其限定在某指数成分股的股票池内，
    目前仅支持沪深300、中证500、中证1000、国证2000成分股，以及这四种指数成分股的组合叠加，和除沪深300、中证500、中证1000以外的股票的成分股

    Parameters
    ----------
    fac : pd.DataFrame
        未限定股票池的因子值，index为时间，columns为股票代码
    hs300 : bool, optional
        限定股票池为沪深300, by default 0
    zz500 : bool, optional
        限定股票池为中证500, by default 0
    zz1000 : bool, optional
        限定股票池为中证1000, by default 0
    gz2000 : bool, optional
        限定股票池为国证2000, by default 0
    other : bool, optional
        限定股票池为除沪深300、中证500、中证1000以外的股票的成分股, by default 0

    Returns
    -------
    `pd.DataFrame`
        仅包含成分股后的因子值，非成分股的因子值为空

    Raises
    ------
    `ValueError`
        如果未指定任何一种指数的成分股，将报错
    """
    last = fac.resample("M").last()
    homeplace = HomePlace()
    dummies = []
    if fac.shape[0] / last.shape[0] > 2:
        if hs300:
            df = pd.read_parquet(
                homeplace.daily_data_file + "沪深300日成分股.parquet"
            ).fillna(0)
            dummies.append(df)
        if zz500:
            df = pd.read_parquet(
                homeplace.daily_data_file + "中证500日成分股.parquet"
            ).fillna(0)
            dummies.append(df)
        if zz1000:
            df = pd.read_parquet(
                homeplace.daily_data_file + "中证1000日成分股.parquet"
            ).fillna(0)
            dummies.append(df)
        if gz2000:
            df = pd.read_parquet(
                homeplace.daily_data_file + "国证2000日成分股.parquet"
            ).fillna(0)
            dummies.append(df)
        if other:
            tr = read_daily(tr=1).fillna(0).replace(0, 1)
            tr = np.sign(tr)
            df1 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "沪深300日成分股.parquet")
            ).fillna(0)
            df2 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "中证500日成分股.parquet")
            ).fillna(0)
            df3 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "中证1000日成分股.parquet")
            ).fillna(0)
            df = (1 - df1) * (1 - df2) * (1 - df3) * tr
            df = df.replace(0, np.nan) * fac
            df = df.dropna(how="all")
        if (hs300 + zz500 + zz1000 + gz2000 + other) == 0:
            raise ValueError("总得指定一下是哪个成分股吧🤒")
    else:
        if hs300:
            df = pd.read_parquet(
                homeplace.daily_data_file + "沪深300日成分股.parquet"
            ).fillna(0)
            df = df.resample("M").last()
            dummies.append(df)
        if zz500:
            df = pd.read_parquet(
                homeplace.daily_data_file + "中证500日成分股.parquet"
            ).fillna(0)
            df = df.resample("M").last()
            dummies.append(df)
        if zz1000:
            df = pd.read_parquet(
                homeplace.daily_data_file + "中证1000日成分股.parquet"
            ).fillna(0)
            df = df.resample("M").last()
            dummies.append(df)
        if gz2000:
            df = pd.read_parquet(
                homeplace.daily_data_file + "国证2000日成分股.parquet"
            ).fillna(0)
            df = df.resample("M").last()
            dummies.append(df)
        if other:
            tr = read_daily(tr=1).fillna(0).replace(0, 1).resample("M").last()
            tr = np.sign(tr)
            df1 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "沪深300日成分股.parquet")
            ).fillna(0)
            df1 = df1.resample("M").last()
            df2 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "中证500日成分股.parquet")
            ).fillna(0)
            df2 = df2.resample("M").last()
            df3 = (
                tr * pd.read_parquet(homeplace.daily_data_file + "中证1000日成分股.parquet")
            ).fillna(0)
            df3 = df3.resample("M").last()
            df = (1 - df1) * (1 - df2) * (1 - df3)
            df = df.replace(0, np.nan) * fac
            df = df.dropna(how="all")
        if (hs300 + zz500 + zz1000 + gz2000 + other) == 0:
            raise ValueError("总得指定一下是哪个成分股吧🤒")
    if len(dummies) > 0:
        dummies = sum(dummies).replace(0, np.nan)
        df = (dummies * fac).dropna(how="all")
    return df

daily_factor_on_industry(df, swindustry=0, zxindustry=0)

将一个因子变为仅在某个申万一级行业上的股票

Parameters

df : pd.DataFrame 全市场的因子值，index是时间，columns是股票代码 swindustry : bool, optional 选择使用申万一级行业, by default 0 zxindustry : bool, optional 选择使用中信一级行业, by default 0

Returns

dict key为行业代码，value为对应的行业上的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
@do_on_dfs
def daily_factor_on_industry(
    df: pd.DataFrame, swindustry: bool = 0, zxindustry: bool = 0
) -> dict:
    """将一个因子变为仅在某个申万一级行业上的股票

    Parameters
    ----------
    df : pd.DataFrame
        全市场的因子值，index是时间，columns是股票代码
    swindustry : bool, optional
        选择使用申万一级行业, by default 0
    zxindustry : bool, optional
        选择使用中信一级行业, by default 0

    Returns
    -------
    dict
        key为行业代码，value为对应的行业上的因子值
    """
    df1 = df.resample("M").last()
    if df1.shape[0] * 2 > df.shape[0]:
        daily = 0
        monthly = 1
    else:
        daily = 1
        monthly = 0
    start = int(datetime.datetime.strftime(df.index.min(), "%Y%m%d"))
    ress = get_industry_dummies(
        daily=daily,
        monthly=monthly,
        start=start,
        swindustry=swindustry,
        zxindustry=zxindustry,
    )
    ress = {k: v * df for k, v in ress.items()}
    return ress

de_cross(y, xs)

使用若干因子对某个因子进行正交化处理

Parameters

y : pd.DataFrame 研究的目标，回归中的y xs : Union[List[pd.DataFrame],pd.DataFrame] 用于正交化的若干因子，回归中的x

Returns

pd.DataFrame 正交化之后的因子

Source code in pure_ocean_breeze/labor/process.py
Python
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
def de_cross(
    y: pd.DataFrame, xs: Union[List[pd.DataFrame], pd.DataFrame]
) -> pd.DataFrame:
    """使用若干因子对某个因子进行正交化处理

    Parameters
    ----------
    y : pd.DataFrame
        研究的目标，回归中的y
    xs : Union[List[pd.DataFrame],pd.DataFrame]
        用于正交化的若干因子，回归中的x

    Returns
    -------
    pd.DataFrame
        正交化之后的因子
    """
    if not isinstance(xs, list):
        xs = [xs]
    y = pure_fallmount(y)
    xs = [pure_fallmount(i) for i in xs]
    return (y - xs)()

deboth(df)

通过回测的方式，对月频因子做行业市值中性化

Parameters

df : pd.DataFrame 未中性化的因子

Returns

pd.DataFrame 行业市值中性化之后的因子

Source code in pure_ocean_breeze/labor/process.py
Python
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
@do_on_dfs
def deboth(df: pd.DataFrame) -> pd.DataFrame:
    """通过回测的方式，对月频因子做行业市值中性化

    Parameters
    ----------
    df : pd.DataFrame
        未中性化的因子

    Returns
    -------
    `pd.DataFrame`
        行业市值中性化之后的因子
    """
    shen = pure_moonnight(df, boxcox=1, plt_plot=0, print_comments=0)
    return shen()

decap(df, daily=0, monthly=0)

对因子做市值中性化

Parameters

df : pd.DataFrame 未中性化的因子，index是时间，columns是股票代码 daily : bool, optional 未中性化因子是日频的则为1，否则为0, by default 0 monthly : bool, optional 未中性化因子是月频的则为1，否则为0, by default 0

Returns

pd.DataFrame 市值中性化之后的因子

Raises

NotImplementedError 如果未指定日频或月频，将报错

Source code in pure_ocean_breeze/labor/process.py
Python
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
@do_on_dfs
@kk.desktop_sender(title="嘿，行业中性化做完啦～🛁")
def decap(df: pd.DataFrame, daily: bool = 0, monthly: bool = 0) -> pd.DataFrame:
    """对因子做市值中性化

    Parameters
    ----------
    df : pd.DataFrame
        未中性化的因子，index是时间，columns是股票代码
    daily : bool, optional
        未中性化因子是日频的则为1，否则为0, by default 0
    monthly : bool, optional
        未中性化因子是月频的则为1，否则为0, by default 0

    Returns
    -------
    `pd.DataFrame`
        市值中性化之后的因子

    Raises
    ------
    `NotImplementedError`
        如果未指定日频或月频，将报错
    """
    tqdm.auto.tqdm.pandas()
    share = read_daily(sharenum=1)
    undi_close = read_daily(close=1, unadjust=1)
    cap = (share * undi_close).stack().reset_index()
    cap.columns = ["date", "code", "cap"]
    cap.cap = ss.boxcox(cap.cap)[0]

    def single(x):
        x.cap = ss.boxcox(x.cap)[0]
        return x

    cap = cap.groupby(["date"]).apply(single)
    cap = cap.set_index(["date", "code"]).unstack()
    cap.columns = [i[1] for i in list(cap.columns)]
    cap_monthly = cap.resample("M").last()
    last = df.resample("M").last()
    if df.shape[0] / last.shape[0] < 2:
        monthly = True
    else:
        daily = True
    if daily:
        df = (pure_fallmount(df) - (pure_fallmount(cap),))()
    elif monthly:
        df = (pure_fallmount(df) - (pure_fallmount(cap_monthly),))()
    else:
        raise NotImplementedError("必须指定频率")
    return df

decap_industry(df, daily=0, monthly=0, swindustry=0, zxindustry=0)

对因子做行业市值中性化

Parameters

df : pd.DataFrame 未中性化的因子，index是时间，columns是股票代码 daily : bool, optional 未中性化因子是日频的则为1，否则为0, by default 0 monthly : bool, optional 未中性化因子是月频的则为1，否则为0, by default 0 swindustry : bool, optional 选择申万一级行业, by default 0 zxindustry : bool, optional 选择中信一级行业, by default 0

Returns

pd.DataFrame 行业市值中性化之后的因子

Raises

NotImplementedError 如果未指定日频或月频，将报错

Source code in pure_ocean_breeze/labor/process.py
Python
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
@do_on_dfs
@kk.desktop_sender(title="嘿，行业市值中性化做完啦～🛁")
def decap_industry(
    df: pd.DataFrame,
    daily: bool = 0,
    monthly: bool = 0,
    swindustry: bool = 0,
    zxindustry: bool = 0,
) -> pd.DataFrame:
    """对因子做行业市值中性化

    Parameters
    ----------
    df : pd.DataFrame
        未中性化的因子，index是时间，columns是股票代码
    daily : bool, optional
        未中性化因子是日频的则为1，否则为0, by default 0
    monthly : bool, optional
        未中性化因子是月频的则为1，否则为0, by default 0
    swindustry : bool, optional
        选择申万一级行业, by default 0
    zxindustry : bool, optional
        选择中信一级行业, by default 0

    Returns
    -------
    `pd.DataFrame`
        行业市值中性化之后的因子

    Raises
    ------
    `NotImplementedError`
        如果未指定日频或月频，将报错
    """
    start_date = int(datetime.datetime.strftime(df.index.min(), "%Y%m%d"))
    last = df.resample("M").last()
    homeplace = HomePlace()
    if daily == 0 and monthly == 0:
        if df.shape[0] / last.shape[0] < 2:
            monthly = True
        else:
            daily = True
    if monthly:
        cap = read_daily(flow_cap=1, start=start_date).resample("M").last()
    else:
        cap = read_daily(flow_cap=1, start=start_date)
    cap = cap.stack().reset_index()
    cap.columns = ["date", "code", "cap"]
    cap.cap = ss.boxcox(cap.cap)[0]

    def single(x):
        x.cap = ss.boxcox(x.cap)[0]
        return x

    cap = cap.groupby(["date"]).apply(single)
    df = df.stack().reset_index()
    df.columns = ["date", "code", "fac"]
    df = pd.merge(df, cap, on=["date", "code"])

    def neutralize_factors(df):
        """组内对因子进行市值中性化"""
        industry_codes = list(df.columns)
        industry_codes = [i for i in industry_codes if i.startswith("w")]
        industry_codes_str = "+".join(industry_codes)
        ols_result = smf.ols("fac~cap+" + industry_codes_str, data=df).fit()
        ols_w = ols_result.params["cap"]
        ols_b = ols_result.params["Intercept"]
        ols_bs = {}
        for ind in industry_codes:
            ols_bs[ind] = ols_result.params[ind]
        df.fac = df.fac - ols_w * df.cap - ols_b
        for k, v in ols_bs.items():
            df.fac = df.fac - v * df[k]
        df = df[["fac"]]
        return df

    if swindustry:
        file_name = "申万行业2021版哑变量.parquet"
    else:
        file_name = "中信一级行业哑变量代码版.parquet"

    if monthly:
        industry_dummy = (
            pd.read_parquet(homeplace.daily_data_file + file_name)
            .fillna(0)
            .set_index("date")
            .groupby("code")
            .resample("M")
            .last()
        )
        industry_dummy = industry_dummy.fillna(0).drop(columns=["code"]).reset_index()
        industry_ws = [f"w{i}" for i in range(1, industry_dummy.shape[1] - 1)]
        col = ["code", "date"] + industry_ws
    elif daily:
        industry_dummy = pd.read_parquet(homeplace.daily_data_file + file_name).fillna(
            0
        )
        industry_ws = [f"w{i}" for i in range(1, industry_dummy.shape[1] - 1)]
        col = ["date", "code"] + industry_ws
    else:
        raise NotImplementedError("必须指定频率")
    industry_dummy.columns = col
    df = pd.merge(df, industry_dummy, on=["date", "code"])
    df = df.set_index(["date", "code"])
    tqdm.auto.tqdm.pandas()
    df = df.groupby(["date"]).progress_apply(neutralize_factors)
    df = df.unstack()
    df.columns = [i[1] for i in list(df.columns)]
    return df

follow_tests(fac, trade_cost_double_side_list=[0.001, 0.002, 0.003, 0.004, 0.005], groups_num=10, index_member_value_weighted=0, comments_writer=None, net_values_writer=None, pos=0, neg=0, swindustry=0, zxindustry=0, nums=[3], opens_average_first_day=0, total_cap=0, without_industry=1)

因子完成全A测试后，进行的一些必要的后续测试，包括各个分组表现、相关系数与纯净化、3510的多空和多头、各个行业Rank IC、各个行业买3只超额表现

Parameters

fac : pd.DataFrame 要进行后续测试的因子值，index是时间，columns是股票代码，values是因子值 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 groups_num : int, optional 分组数量, by default 10 index_member_value_weighted : bool, optional 成分股多头采取流通市值加权 comments_writer : pd.ExcelWriter, optional 写入评价指标的excel, by default None net_values_writer : pd.ExcelWriter, optional 写入净值序列的excel, by default None pos : bool, optional 因子的方向为正, by default 0 neg : bool, optional 因子的方向为负, by default 0 swindustry : bool, optional 使用申万一级行业, by default 0 zxindustry : bool, optional 使用中信一级行业, by default 0 nums : List[int], optional 各个行业买几只股票, by default [3] opens_average_first_day : bool, optional 买入时使用第一天的平均价格, by default 0 total_cap : bool, optional 加权和行业市值中性化时使用总市值, by default 0 without_industry : bool, optional 是否不对行业做测试, by default 1

Raises

IOError 如果未指定因子正负方向，将报错

Source code in pure_ocean_breeze/labor/process.py
Python
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
@do_on_dfs
def follow_tests(
    fac: pd.DataFrame,
    trade_cost_double_side_list: float = [0.001, 0.002, 0.003, 0.004, 0.005],
    groups_num: int = 10,
    index_member_value_weighted: bool = 0,
    comments_writer: pd.ExcelWriter = None,
    net_values_writer: pd.ExcelWriter = None,
    pos: bool = 0,
    neg: bool = 0,
    swindustry: bool = 0,
    zxindustry: bool = 0,
    nums: List[int] = [3],
    opens_average_first_day: bool = 0,
    total_cap: bool = 0,
    without_industry: bool = 1,
):
    """因子完成全A测试后，进行的一些必要的后续测试，包括各个分组表现、相关系数与纯净化、3510的多空和多头、各个行业Rank IC、各个行业买3只超额表现

    Parameters
    ----------
    fac : pd.DataFrame
        要进行后续测试的因子值，index是时间，columns是股票代码，values是因子值
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    groups_num : int, optional
        分组数量, by default 10 
    index_member_value_weighted : bool, optional
        成分股多头采取流通市值加权
    comments_writer : pd.ExcelWriter, optional
        写入评价指标的excel, by default None
    net_values_writer : pd.ExcelWriter, optional
        写入净值序列的excel, by default None
    pos : bool, optional
        因子的方向为正, by default 0
    neg : bool, optional
        因子的方向为负, by default 0
    swindustry : bool, optional
        使用申万一级行业, by default 0
    zxindustry : bool, optional
        使用中信一级行业, by default 0
    nums : List[int], optional
        各个行业买几只股票, by default [3]
    opens_average_first_day : bool, optional
        买入时使用第一天的平均价格, by default 0
    total_cap : bool, optional
        加权和行业市值中性化时使用总市值, by default 0
    without_industry : bool, optional
        是否不对行业做测试, by default 1

    Raises
    ------
    IOError
        如果未指定因子正负方向，将报错
    """
    if comments_writer is None:
        from pure_ocean_breeze.state.states import COMMENTS_WRITER

        comments_writer = COMMENTS_WRITER
    if net_values_writer is None:
        from pure_ocean_breeze.state.states import NET_VALUES_WRITER

        net_values_writer = NET_VALUES_WRITER

    shen = pure_moonnight(
        fac,
        opens_average_first_day=opens_average_first_day,
        trade_cost_double_side=0.003,
    )
    if (
        shen.shen.group_net_values.group1.iloc[-1]
        > shen.shen.group_net_values.group10.iloc[-1]
    ):
        neg = 1
    else:
        pos = 1
    if comments_writer is not None:
        shen.comments_ten().to_excel(comments_writer, sheet_name="十分组")
    print(shen.comments_ten())
    """相关系数与纯净化"""
    pure_fac = pure_snowtrain(fac)
    if comments_writer is not None:
        pure_fac.corr.to_excel(comments_writer, sheet_name="相关系数")
    print(pure_fac.corr)
    shen = pure_moonnight(
        pure_fac(),
        comments_writer=comments_writer,
        net_values_writer=net_values_writer,
        sheetname="纯净",
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
    )
    """3510多空和多头"""
    # 300
    fi300 = daily_factor_on300500(fac, hs300=1)
    shen = pure_moonnight(
        fi300,
        groups_num=groups_num,
        value_weighted=index_member_value_weighted,
        comments_writer=comments_writer,
        net_values_writer=net_values_writer,
        sheetname="300多空",
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
        trade_cost_double_side=0.003,
    )
    if pos:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], hs300=1).to_excel(
                comments_writer, sheet_name="300超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    hs300=1,
                ).to_excel(comments_writer, sheet_name=f"300超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], hs300=1)
        if net_values_writer is not None:
            make_relative_comments_plot(shen.shen.group_rets[f'group{groups_num}'], hs300=1).to_excel(
                net_values_writer, sheet_name="300超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    hs300=1,
                ).to_excel(net_values_writer, sheet_name=f"300超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets[f'group{groups_num}'], hs300=1)
    elif neg:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets.group1, hs300=1).to_excel(
                comments_writer, sheet_name="300超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    hs300=1,
                ).to_excel(comments_writer, sheet_name=f"300超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets.group1, hs300=1)
        if net_values_writer is not None:
            make_relative_comments_plot(shen.shen.group_rets.group1, hs300=1).to_excel(
                net_values_writer, sheet_name="300超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    hs300=1,
                ).to_excel(net_values_writer, sheet_name=f"300超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets.group1, hs300=1)
    else:
        raise IOError("请指定因子的方向是正是负🤒")
    # 500
    fi500 = daily_factor_on300500(fac, zz500=1)
    shen = pure_moonnight(
        fi500,
        groups_num=groups_num,
        value_weighted=index_member_value_weighted,
        comments_writer=comments_writer,
        net_values_writer=net_values_writer,
        sheetname="500多空",
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
        trade_cost_double_side=0.003,
    )
    if pos:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], zz500=1).to_excel(
                comments_writer, sheet_name="500超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    zz500=1,
                ).to_excel(comments_writer, sheet_name=f"500超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], zz500=1)
        if net_values_writer is not None:
            make_relative_comments_plot(shen.shen.group_rets[f'group{groups_num}'], zz500=1).to_excel(
                net_values_writer, sheet_name="500超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    zz500=1,
                ).to_excel(net_values_writer, sheet_name=f"500超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets[f'group{groups_num}'], zz500=1)
    else:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets.group1, zz500=1).to_excel(
                comments_writer, sheet_name="500超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    zz500=1,
                ).to_excel(comments_writer, sheet_name=f"500超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets.group1, zz500=1)
        if net_values_writer is not None:
            make_relative_comments_plot(shen.shen.group_rets.group1, zz500=1).to_excel(
                net_values_writer, sheet_name="500超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    zz500=1,
                ).to_excel(net_values_writer, sheet_name=f"500超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets.group1, zz500=1)
    # 1000
    fi1000 = daily_factor_on300500(fac, zz1000=1)
    shen = pure_moonnight(
        fi1000,
        groups_num=groups_num,
        value_weighted=index_member_value_weighted,
        comments_writer=comments_writer,
        net_values_writer=net_values_writer,
        sheetname="1000多空",
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
        trade_cost_double_side=0.003,
    )
    if pos:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], zz1000=1).to_excel(
                comments_writer, sheet_name="1000超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    zz1000=1,
                ).to_excel(comments_writer, sheet_name=f"1000超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets[f'group{groups_num}'], zz1000=1)
        if net_values_writer is not None:
            make_relative_comments_plot(
                shen.shen.group_rets[f'group{groups_num}'], zz1000=1
            ).to_excel(net_values_writer, sheet_name="1000超额")
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets[f'group{groups_num}']
                    - shen.shen.factor_turnover_rates[f'group{groups_num}'] * i,
                    zz1000=1,
                ).to_excel(net_values_writer, sheet_name=f"1000超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets[f'group{groups_num}'], zz1000=1)
    else:
        if comments_writer is not None:
            make_relative_comments(shen.shen.group_rets.group1, zz1000=1).to_excel(
                comments_writer, sheet_name="1000超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    zz1000=1,
                ).to_excel(comments_writer, sheet_name=f"1000超额双边费率{i}")
        else:
            make_relative_comments(shen.shen.group_rets.group1, zz1000=1)
        if net_values_writer is not None:
            make_relative_comments_plot(shen.shen.group_rets.group1, zz1000=1).to_excel(
                net_values_writer, sheet_name="1000超额"
            )
            for i in trade_cost_double_side_list:
                make_relative_comments_plot(
                    shen.shen.group_rets.group1
                    - shen.shen.factor_turnover_rates.group1 * i,
                    zz1000=1,
                ).to_excel(net_values_writer, sheet_name=f"1000超额双边费率{i}")
        else:
            make_relative_comments_plot(shen.shen.group_rets.group1, zz1000=1)
    if not without_industry:
        # 各行业Rank IC
        rankics = rankic_test_on_industry(fac, comments_writer)
        # 买3只超额表现
        rets = long_test_on_industry(
            fac, nums, pos=pos, neg=neg, swindustry=swindustry, zxindustry=zxindustry
        )
    logger.success("因子后续的必要测试全部完成")

get_group(df, group_num=10)

使用groupby的方法，将一组因子值改为截面上的分组值，此方法相比qcut的方法更加稳健，但速度更慢一些

Parameters

df : pd.DataFrame 因子值，index为时间，columns为股票代码，values为因子值 group_num : int, optional 分组的数量, by default 10

Returns

pd.DataFrame 转化为分组值后的df，index为时间，columns为股票代码，values为分组值

Source code in pure_ocean_breeze/labor/process.py
Python
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
@do_on_dfs
def get_group(df: pd.DataFrame, group_num: int = 10) -> pd.DataFrame:
    """使用groupby的方法，将一组因子值改为截面上的分组值，此方法相比qcut的方法更加稳健，但速度更慢一些

    Parameters
    ----------
    df : pd.DataFrame
        因子值，index为时间，columns为股票代码，values为因子值
    group_num : int, optional
        分组的数量, by default 10

    Returns
    -------
    pd.DataFrame
        转化为分组值后的df，index为时间，columns为股票代码，values为分组值
    """
    a = pure_moon(no_read_indu=1)
    df = df.stack().reset_index()
    df.columns = ["date", "code", "fac"]
    df = a.get_groups(df, group_num).pivot(index="date", columns="code", values="group")
    return df

group_test_on_industry(df, group_num=10, trade_cost_double_side=0, net_values_writer=None, swindustry=0, zxindustry=0)

在申万一级行业上测试每个行业的分组回测

Parameters

df : pd.DataFrame 全市场的因子值，index是时间，columns是股票代码 group_num : int, optional 分组数量, by default 10 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 net_values_writer : pd.ExcelWriter, optional 用于存储各个行业分组及多空对冲净值序列的excel文件, by default None swindustry : bool, optional 选择使用申万一级行业, by default 0 zxindustry : bool, optional 选择使用中信一级行业, by default 0

Returns

pd.DataFrame 各个行业的绩效评价汇总

Source code in pure_ocean_breeze/labor/process.py
Python
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
@do_on_dfs
def group_test_on_industry(
    df: pd.DataFrame,
    group_num: int = 10,
    trade_cost_double_side: float = 0,
    net_values_writer: pd.ExcelWriter = None,
    swindustry: bool = 0,
    zxindustry: bool = 0,
) -> pd.DataFrame:
    """在申万一级行业上测试每个行业的分组回测

    Parameters
    ----------
    df : pd.DataFrame
        全市场的因子值，index是时间，columns是股票代码
    group_num : int, optional
        分组数量, by default 10
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    net_values_writer : pd.ExcelWriter, optional
        用于存储各个行业分组及多空对冲净值序列的excel文件, by default None
    swindustry : bool, optional
        选择使用申万一级行业, by default 0
    zxindustry : bool, optional
        选择使用中信一级行业, by default 0

    Returns
    -------
    pd.DataFrame
        各个行业的绩效评价汇总
    """
    dfs = daily_factor_on_industry(df, swindustry=swindustry, zxindustry=zxindustry)

    ks = []
    vs = []
    if swindustry:
        for k, v in dfs.items():
            shen = pure_moonnight(
                v,
                groups_num=group_num,
                trade_cost_double_side=trade_cost_double_side,
                net_values_writer=net_values_writer,
                sheetname=INDUS_DICT[k],
                plt_plot=0,
            )
            ks.append(k)
            vs.append(shen.shen.total_comments.T)
        vs = pd.concat(vs)
        vs.index = [INDUS_DICT[i] for i in ks]
    else:
        for k, v in dfs.items():
            shen = pure_moonnight(
                v,
                groups_num=group_num,
                trade_cost_double_side=trade_cost_double_side,
                net_values_writer=net_values_writer,
                sheetname=k,
                plt_plot=0,
            )
            ks.append(k)
            vs.append(shen.shen.total_comments.T)
        vs = pd.concat(vs)
        vs.index = ks
    return vs

icir_weight(facs, backsee=6, boxcox=0, rank_corr=0, only_ic=0)

使用icir滚动加权的方式，加权合成几个因子

Parameters

facs : list[pd.DataFrame] 要合成的若干因子，每个df都是index为时间，columns为股票代码，values为因子值的df backsee : int, optional 用来计算icir的过去期数, by default 6 boxcox : bool, optional 是否对因子进行行业市值中性化, by default 0 rank_corr : bool, optional 是否计算rankicir, by default 0 only_ic : bool, optional 是否只计算IC或Rank IC, by default 0

Returns

pd.DataFrame 合成后的因子

Raises

ValueError 因子期数少于回看期数时将报错

Source code in pure_ocean_breeze/labor/process.py
Python
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
def icir_weight(
    facs: list[pd.DataFrame],
    backsee: int = 6,
    boxcox: bool = 0,
    rank_corr: bool = 0,
    only_ic: bool = 0,
) -> pd.DataFrame:
    """使用icir滚动加权的方式，加权合成几个因子

    Parameters
    ----------
    facs : list[pd.DataFrame]
        要合成的若干因子，每个df都是index为时间，columns为股票代码，values为因子值的df
    backsee : int, optional
        用来计算icir的过去期数, by default 6
    boxcox : bool, optional
        是否对因子进行行业市值中性化, by default 0
    rank_corr : bool, optional
        是否计算rankicir, by default 0
    only_ic : bool, optional
        是否只计算IC或Rank IC, by default 0

    Returns
    -------
    pd.DataFrame
        合成后的因子

    Raises
    ------
    ValueError
        因子期数少于回看期数时将报错
    """
    date_first_max = max([i.index[0] for i in facs])
    facs = [i[i.index >= date_first_max] for i in facs]
    date_last_min = min([i.index[-1] for i in facs])
    facs = [i[i.index <= date_last_min] for i in facs]
    facs = [i.shift(1) for i in facs]
    ret = read_daily(
        close=1, start=datetime.datetime.strftime(date_first_max, "%Y%m%d")
    )
    ret = ret / ret.shift(20) - 1
    if boxcox:
        facs = [decap_industry(i) for i in facs]
    facs = [((i.T - i.T.mean()) / i.T.std()).T for i in facs]
    dates = list(facs[0].index)
    fis = []
    for num, date in tqdm.auto.tqdm(list(enumerate(dates))):
        if num < backsee:
            ...
        else:
            nears = [i.iloc[num - backsee : num, :] for i in facs]
            targets = [i[i.index == date] for i in facs]
            if rank_corr:
                weights = [
                    show_corr(
                        i, ret[ret.index.isin(i.index)], plt_plot=0, show_series=1
                    )
                    for i in nears
                ]
            else:
                weights = [
                    show_corr(
                        i,
                        ret[ret.index.isin(i.index)],
                        plt_plot=0,
                        show_series=1,
                        method="pearson",
                    )
                    for i in nears
                ]
            if only_ic:
                weights = [i.mean() for i in weights]
            else:
                weights = [i.mean() / i.std() for i in weights]
            fi = sum([i * j for i, j in zip(weights, targets)])
            fis.append(fi)
    if len(fis) > 0:
        return pd.concat(fis).shift(-1)
    else:
        raise ValueError("输入的因子值长度不太够吧？")

long_test_on_industry(df, nums, pos=0, neg=0, save_stock_list=0, swindustry=0, zxindustry=0)

对每个申万/中信一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

Parameters

df : pd.DataFrame 使用的因子，index为时间，columns为股票代码 nums : list 多头想选取的股票的数量，例如[3,4,5] pos : bool, optional 因子方向为正，即Rank IC为正，则指定此处为True, by default 0 neg : bool, optional 因子方向为负，即Rank IC为负，则指定此处为False, by default 0 save_stock_list : bool, optional 是否保存每月每个行业的多头名单，会降低运行速度, by default 0 swindustry : bool, optional 在申万一级行业上测试, by default 0 zxindusrty : bool, optional 在中信一级行业上测试, by default 0 Returns

List[dict] 超额收益绩效、每月超额收益、每月每个行业的多头名单

Raises

IOError pos和neg必须有一个为1，否则将报错

Source code in pure_ocean_breeze/labor/process.py
Python
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
@do_on_dfs
def long_test_on_industry(
    df: pd.DataFrame,
    nums: list,
    pos: bool = 0,
    neg: bool = 0,
    save_stock_list: bool = 0,
    swindustry: bool = 0,
    zxindustry: bool = 0,
) -> List[dict]:
    """对每个申万/中信一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

    Parameters
    ----------
    df : pd.DataFrame
        使用的因子，index为时间，columns为股票代码
    nums : list
        多头想选取的股票的数量，例如[3,4,5]
    pos : bool, optional
        因子方向为正，即Rank IC为正，则指定此处为True, by default 0
    neg : bool, optional
        因子方向为负，即Rank IC为负，则指定此处为False, by default 0
    save_stock_list : bool, optional
        是否保存每月每个行业的多头名单，会降低运行速度, by default 0
    swindustry : bool, optional
        在申万一级行业上测试, by default 0
    zxindusrty : bool, optional
        在中信一级行业上测试, by default 0
    Returns
    -------
    List[dict]
        超额收益绩效、每月超额收益、每月每个行业的多头名单

    Raises
    ------
    IOError
        pos和neg必须有一个为1，否则将报错
    """
    fac = decap_industry(df, monthly=True)

    if swindustry:
        industry_dummy = pd.read_parquet(
            homeplace.daily_data_file + "申万行业2021版哑变量.parquet"
        ).fillna(0)
        indus = read_swindustry_prices()
    else:
        industry_dummy = pd.read_parquet(
            homeplace.daily_data_file + "中信一级行业哑变量名称版.parquet"
        ).fillna(0)
        indus = read_zxindustry_prices()
    inds = list(industry_dummy.columns)
    ret_next = (
        read_daily(close=1).resample("M").last()
        / read_daily(open=1).resample("M").first()
        - 1
    )
    ages = read_daily(age=1).resample("M").last()
    ages = (ages >= 60) + 0
    ages = ages.replace(0, np.nan)
    ret_next = ret_next * ages
    ret_next_dummy = 1 - ret_next.isna()

    def save_ind(code, num):
        ind = industry_dummy[["date", "code", code]]
        ind = ind.pivot(index="date", columns="code", values=code)
        ind = ind.resample("M").last()
        ind = ind.replace(0, np.nan)
        fi = ind * fac
        fi = fi.dropna(how="all")
        fi = fi.shift(1)
        fi = fi * ret_next_dummy
        fi = fi.dropna(how="all")

        def sing(x):
            if neg:
                thr = x.nsmallest(num).iloc[-1]
            elif pos:
                thr = x.nlargest(num).iloc[-1]
            else:
                raise IOError("您需要指定一下因子方向🤒")
            x = (x <= thr) + 0
            return x

        fi = fi.T.apply(sing).T
        fi = fi.replace(0, np.nan)
        fi = fi * ret_next
        ret_long = fi.mean(axis=1)
        return ret_long

    ret_longs = {k: [] for k in nums}
    for num in tqdm.auto.tqdm(nums):
        for code in inds[2:]:
            df = save_ind(code, num).to_frame(code)
            ret_longs[num] = ret_longs[num] + [df]

    indus = indus.resample("M").last().pct_change()

    if swindustry:
        coms = {
            k: indus_name(pd.concat(v, axis=1).dropna(how="all").T).T
            for k, v in ret_longs.items()
        }
        rets = {
            k: (v - indus_name(indus.T).T).dropna(how="all") for k, v in coms.items()
        }
    else:
        coms = {k: pd.concat(v, axis=1).dropna(how="all") for k, v in ret_longs.items()}
        rets = {k: (v - indus).dropna(how="all") for k, v in coms.items()}

    nets = {k: (v + 1).cumprod() for k, v in rets.items()}
    nets = {
        k: v.apply(lambda x: x.dropna() / x.dropna().iloc[0]) for k, v in nets.items()
    }

    def comments_on_twins(nets: pd.Series, rets: pd.Series) -> pd.Series:
        series = nets.copy()
        series1 = rets.copy()
        ret = (series.iloc[-1] - series.iloc[0]) / series.iloc[0]
        duration = (series.index[-1] - series.index[0]).days
        year = duration / 365
        ret_yearly = (series.iloc[-1] / series.iloc[0]) ** (1 / year) - 1
        max_draw = -(series / series.expanding(1).max() - 1).min()
        vol = np.std(series1) * (12**0.5)
        sharpe = ret_yearly / vol
        wins = series1[series1 > 0]
        win_rate = len(wins) / len(series1)
        return pd.Series(
            [ret, ret_yearly, vol, sharpe, win_rate, max_draw],
            index=["总收益率", "年化收益率", "年化波动率", "信息比率", "胜率", "最大回撤率"],
        )

    if swindustry:
        name = "申万"
    else:
        name = "中信"
    w = pd.ExcelWriter(f"各个{name}一级行业多头超额绩效.xlsx")

    def com_all(df1, df2, num):
        cs = []
        for ind in list(df1.columns):
            c = comments_on_twins(df2[ind].dropna(), df1[ind].dropna()).to_frame(ind)
            cs.append(c)
        res = pd.concat(cs, axis=1).T
        res.to_excel(w, sheet_name=str(num))
        return res

    coms_finals = {k: com_all(rets[k], nets[k], k) for k in rets.keys()}
    w.save()
    w.close()

    rets_save = {k: v.dropna() for k, v in rets.items() if k in nums}
    u = pd.ExcelWriter(f"各个{name}一级行业每月超额收益率.xlsx")
    for k, v in rets_save.items():
        v.to_excel(u, sheet_name=str(k))
    u.save()
    u.close()

    if save_stock_list:

        def save_ind_stocks(code, num):
            ind = industry_dummy[["date", "code", code]]
            ind = ind.pivot(index="date", columns="code", values=code)
            ind = ind.replace(0, np.nan)
            fi = ind * fac
            fi = fi.dropna(how="all")
            fi = fi.shift(1)
            fi = fi * ret_next_dummy
            fi = fi.dropna(how="all")

            def sing(x):
                if neg:
                    thr = x.nsmallest(num)
                elif pos:
                    thr = x.nlargest(num)
                else:
                    raise IOError("您需要指定一下因子方向🤒")
                return tuple(thr.index)

            fi = fi.T.apply(sing)
            return fi

        stocks_longs = {k: {} for k in nums}

        for num in tqdm.auto.tqdm(nums):
            for code in inds[2:]:
                stocks_longs[num][code] = save_ind_stocks(code, num)

        for num in nums:
            w1 = pd.ExcelWriter(f"各个{name}一级行业买{num}只的股票名单.xlsx")
            for k, v in stocks_longs[num].items():
                v = v.T
                v.index = v.index.strftime("%Y/%m/%d")
                v.to_excel(w1, sheet_name=INDUS_DICT[k])
            w1.save()
            w1.close()

        return [coms_finals, rets_save, stocks_longs]
    else:
        return [coms_finals, rets_save]

long_test_on_swindustry(df, nums, pos=0, neg=0, save_stock_list=0)

对每个申万一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

Parameters

df : pd.DataFrame 使用的因子，index为时间，columns为股票代码 nums : list 多头想选取的股票的数量，例如[3,4,5] pos : bool, optional 因子方向为正，即Rank IC为正，则指定此处为True, by default 0 neg : bool, optional 因子方向为负，即Rank IC为负，则指定此处为False, by default 0 save_stock_list : bool, optional 是否保存每月每个行业的多头名单，会降低运行速度, by default 0 Returns

List[dict] 超额收益绩效、每月超额收益、每月每个行业的多头名单

Raises

IOError pos和neg必须有一个为1，否则将报错

Source code in pure_ocean_breeze/labor/process.py
Python
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
@do_on_dfs
def long_test_on_swindustry(
    df: pd.DataFrame,
    nums: list,
    pos: bool = 0,
    neg: bool = 0,
    save_stock_list: bool = 0,
) -> List[dict]:
    """对每个申万一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

    Parameters
    ----------
    df : pd.DataFrame
        使用的因子，index为时间，columns为股票代码
    nums : list
        多头想选取的股票的数量，例如[3,4,5]
    pos : bool, optional
        因子方向为正，即Rank IC为正，则指定此处为True, by default 0
    neg : bool, optional
        因子方向为负，即Rank IC为负，则指定此处为False, by default 0
    save_stock_list : bool, optional
        是否保存每月每个行业的多头名单，会降低运行速度, by default 0
    Returns
    -------
    List[dict]
        超额收益绩效、每月超额收益、每月每个行业的多头名单

    Raises
    ------
    IOError
        pos和neg必须有一个为1，否则将报错
    """
    res = long_test_on_industry(
        df=df,
        nums=nums,
        pos=pos,
        neg=neg,
        save_stock_list=save_stock_list,
        swindustry=1,
    )
    return res

long_test_on_zxindustry(df, nums, pos=0, neg=0, save_stock_list=0)

对每个中信一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

Parameters

df : pd.DataFrame 使用的因子，index为时间，columns为股票代码 nums : list 多头想选取的股票的数量，例如[3,4,5] pos : bool, optional 因子方向为正，即Rank IC为正，则指定此处为True, by default 0 neg : bool, optional 因子方向为负，即Rank IC为负，则指定此处为False, by default 0 save_stock_list : bool, optional 是否保存每月每个行业的多头名单，会降低运行速度, by default 0 Returns

List[dict] 超额收益绩效、每月超额收益、每月每个行业的多头名单

Raises

IOError pos和neg必须有一个为1，否则将报错

Source code in pure_ocean_breeze/labor/process.py
Python
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
@do_on_dfs
def long_test_on_zxindustry(
    df: pd.DataFrame,
    nums: list,
    pos: bool = 0,
    neg: bool = 0,
    save_stock_list: bool = 0,
) -> List[dict]:
    """对每个中信一级行业成分股，使用某因子挑选出最多头的n值股票，考察其超额收益绩效、每月超额收益、每月每个行业的多头名单

    Parameters
    ----------
    df : pd.DataFrame
        使用的因子，index为时间，columns为股票代码
    nums : list
        多头想选取的股票的数量，例如[3,4,5]
    pos : bool, optional
        因子方向为正，即Rank IC为正，则指定此处为True, by default 0
    neg : bool, optional
        因子方向为负，即Rank IC为负，则指定此处为False, by default 0
    save_stock_list : bool, optional
        是否保存每月每个行业的多头名单，会降低运行速度, by default 0
    Returns
    -------
    List[dict]
        超额收益绩效、每月超额收益、每月每个行业的多头名单

    Raises
    ------
    IOError
        pos和neg必须有一个为1，否则将报错
    """
    res = long_test_on_industry(
        df=df,
        nums=nums,
        pos=pos,
        neg=neg,
        save_stock_list=save_stock_list,
        zxindustry=1,
    )
    return res

market_kind(df, zhuban=0, chuangye=0, kechuang=0, beijing=0)

与宽基指数成分股的函数类似，限定股票在某个具体板块上

Parameters

df : pd.DataFrame 原始全部股票的因子值 zhuban : bool, optional 限定在主板范围内, by default 0 chuangye : bool, optional 限定在创业板范围内, by default 0 kechuang : bool, optional 限定在科创板范围内, by default 0 beijing : bool, optional 限定在北交所范围内, by default 0

Returns

pd.DataFrame 限制范围后的因子值，其余为空

Raises

ValueError 如果未指定任何股票池，将报错

Source code in pure_ocean_breeze/labor/process.py
Python
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
@do_on_dfs
def market_kind(
    df: pd.DataFrame,
    zhuban: bool = 0,
    chuangye: bool = 0,
    kechuang: bool = 0,
    beijing: bool = 0,
) -> pd.DataFrame:
    """与宽基指数成分股的函数类似，限定股票在某个具体板块上

    Parameters
    ----------
    df : pd.DataFrame
        原始全部股票的因子值
    zhuban : bool, optional
        限定在主板范围内, by default 0
    chuangye : bool, optional
        限定在创业板范围内, by default 0
    kechuang : bool, optional
        限定在科创板范围内, by default 0
    beijing : bool, optional
        限定在北交所范围内, by default 0

    Returns
    -------
    `pd.DataFrame`
        限制范围后的因子值，其余为空

    Raises
    ------
    `ValueError`
        如果未指定任何股票池，将报错
    """
    trs = read_daily(tr=1)
    codes = list(trs.columns)
    dates = list(trs.index)
    if chuangye and kechuang:
        dummys = [1 if code[:2] in ["30", "68"] else np.nan for code in codes]
    else:
        if zhuban:
            dummys = [1 if code[:2] in ["00", "60"] else np.nan for code in codes]
        elif chuangye:
            dummys = [1 if code.startswith("3") else np.nan for code in codes]
        elif kechuang:
            dummys = [1 if code.startswith("68") else np.nan for code in codes]
        elif beijing:
            dummys = [1 if code.startswith("8") else np.nan for code in codes]
        else:
            raise ValueError("你总得选一个股票池吧？🤒")
    dummy_dict = {k: v for k, v in zip(codes, dummys)}
    dummy_df = pd.DataFrame(dummy_dict, index=dates)
    df = df * dummy_df
    return df

rankic_test_on_industry(df, excel_name='行业rankic.xlsx', png_name='行业rankic图.png', swindustry=0, zxindustry=0)

专门计算因子值在各个申万一级行业上的Rank IC值，并绘制柱状图

Parameters

df : pd.DataFrame 全市场的因子值，index是时间，columns是股票代码 excel_name : str, optional 用于保存各个行业Rank IC值的excel文件的名字, by default '行业rankic.xlsx' png_name : str, optional 用于保存各个行业Rank IC值的柱状图的名字, by default '行业rankic图.png' swindustry : bool, optional 选择使用申万一级行业, by default 0 zxindustry : bool, optional 选择使用中信一级行业, by default 0

Returns

pd.DataFrame 行业名称与对应的Rank IC

Source code in pure_ocean_breeze/labor/process.py
Python
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
@do_on_dfs
def rankic_test_on_industry(
    df: pd.DataFrame,
    excel_name: str = "行业rankic.xlsx",
    png_name: str = "行业rankic图.png",
    swindustry: bool = 0,
    zxindustry: bool = 0,
) -> pd.DataFrame:
    """专门计算因子值在各个申万一级行业上的Rank IC值，并绘制柱状图

    Parameters
    ----------
    df : pd.DataFrame
        全市场的因子值，index是时间，columns是股票代码
    excel_name : str, optional
        用于保存各个行业Rank IC值的excel文件的名字, by default '行业rankic.xlsx'
    png_name : str, optional
        用于保存各个行业Rank IC值的柱状图的名字, by default '行业rankic图.png'
    swindustry : bool, optional
        选择使用申万一级行业, by default 0
    zxindustry : bool, optional
        选择使用中信一级行业, by default 0

    Returns
    -------
    pd.DataFrame
        行业名称与对应的Rank IC
    """
    vs = group_test_on_industry(df, swindustry=swindustry, zxindustry=zxindustry)
    rankics = vs[["RankIC"]].T
    if excel_name is not None:
        rankics.to_excel(excel_name)
    rankics.plot(kind="bar")
    plt.show()
    plt.savefig(png_name)
    return rankics

remove_unavailable(df)

对日频或月频因子值，剔除st股、不正常交易的股票和上市不足60天的股票

Parameters

df : pd.DataFrame 因子值，index是时间，columns是股票代码，values是因子值

Returns

pd.DataFrame 剔除后的因子值

Source code in pure_ocean_breeze/labor/process.py
Python
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
@do_on_dfs
def remove_unavailable(df: pd.DataFrame) -> pd.DataFrame:
    """对日频或月频因子值，剔除st股、不正常交易的股票和上市不足60天的股票

    Parameters
    ----------
    df : pd.DataFrame
        因子值，index是时间，columns是股票代码，values是因子值

    Returns
    -------
    pd.DataFrame
        剔除后的因子值
    """
    df0 = df.resample("M").last()
    if df.shape[0] / df0.shape[0] > 2:
        daily = 1
    else:
        daily = 0
    if daily:
        state = read_daily(state=1).replace(0, np.nan)
        st = read_daily(st=1)
        age = read_daily(age=1)
        st = (1 - st).replace(0, np.nan)
        age = ((age >= 60) + 0).replace(0, np.nan)
        df = df * age * st * state
    else:
        moon = pure_moon(no_read_indu=1)
        moon.set_basic_data()
        moon.judge_month()
        df = moon.tris_monthly * df
    return df

scipy_weight(facs, backsee=6, boxcox=0, rank_corr=0, only_ic=0, upper_bound=None, lower_bound=0)

使用scipy的minimize优化求解的方式，寻找最优的因子合成权重，默认优化条件为最大ICIR

Parameters

facs : list[pd.DataFrame] 要合成的因子，每个df都是index为时间，columns为股票代码，values为因子值的df backsee : int, optional 用来计算icir的过去期数, by default 6 boxcox : bool, optional 是否对因子进行行业市值中性化, by default 0 rank_corr : bool, optional 是否计算rankicir, by default 0 only_ic : bool, optional 是否只计算IC或Rank IC, by default 0 upper_bound : float, optional 每个因子的权重上限，如果不指定，则为每个因子平均权重的2倍，即2除以因子数量, by default None lower_bound : float, optional 每个因子的权重下限, by default 0

Returns

pd.DataFrame 合成后的因子

Source code in pure_ocean_breeze/labor/process.py
Python
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
def scipy_weight(
    facs: list[pd.DataFrame],
    backsee: int = 6,
    boxcox: bool = 0,
    rank_corr: bool = 0,
    only_ic: bool = 0,
    upper_bound: float = None,
    lower_bound: float = 0,
) -> pd.DataFrame:
    """使用scipy的minimize优化求解的方式，寻找最优的因子合成权重，默认优化条件为最大ICIR

    Parameters
    ----------
    facs : list[pd.DataFrame]
        要合成的因子，每个df都是index为时间，columns为股票代码，values为因子值的df
    backsee : int, optional
        用来计算icir的过去期数, by default 6
    boxcox : bool, optional
        是否对因子进行行业市值中性化, by default 0
    rank_corr : bool, optional
        是否计算rankicir, by default 0
    only_ic : bool, optional
        是否只计算IC或Rank IC, by default 0
    upper_bound : float, optional
        每个因子的权重上限，如果不指定，则为每个因子平均权重的2倍，即2除以因子数量, by default None
    lower_bound : float, optional
        每个因子的权重下限, by default 0

    Returns
    -------
    pd.DataFrame
        合成后的因子
    """
    date_first_max = max([i.index[0] for i in facs])
    facs = [i[i.index >= date_first_max] for i in facs]
    date_last_min = min([i.index[-1] for i in facs])
    facs = [i[i.index <= date_last_min] for i in facs]
    facs = [i.shift(1) for i in facs]
    ret = read_daily(
        close=1, start=datetime.datetime.strftime(date_first_max, "%Y%m%d")
    )
    ret = ret / ret.shift(20) - 1
    if boxcox:
        facs = [decap_industry(i) for i in facs]
    facs = [((i.T - i.T.mean()) / i.T.std()).T for i in facs]
    if upper_bound is None:
        upper_bound = 2 / len(facs)
    dates = list(facs[0].index)
    fis = []
    for num, date in tqdm.auto.tqdm(list(enumerate(dates))):
        if num <= backsee:
            ...
        else:
            nears = [i.iloc[num - backsee : num, :] for i in facs]
            targets = [i[i.index == date] for i in facs]
            if rank_corr:
                weights = [
                    show_corr(
                        i, ret[ret.index.isin(i.index)], plt_plot=0, show_series=1
                    )
                    for i in nears
                ]
            else:
                weights = [
                    show_corr(
                        i,
                        ret[ret.index.isin(i.index)],
                        plt_plot=0,
                        show_series=1,
                        method="pearson",
                    )
                    for i in nears
                ]
            if only_ic:
                weights = [i.mean() for i in weights]
            else:
                weights = [i.mean() / i.std() for i in weights]
            weights = pd.concat(weights, axis=1)

            def func(x):
                w = np.array(x).reshape((-1, 1))
                y = weights @ w
                return np.mean(y) / np.std(y)

            cons = {"type": "eq", "fun": lambda x: np.sum(x) - 1}
            res = minimize(
                func,
                np.random.rand(weights.shape[1], 1),
                constraints=cons,
                bounds=[(lower_bound, upper_bound)] * weights.shape[1],
            )
            xs = res.x.tolist()
            fac = sum([i * j for i, j in zip(xs, targets)])
            fis.append(fac)
    return pd.concat(fis).shift(-1)

show_corr(fac1, fac2, method='pearson', plt_plot=1, show_series=0, old_way=0)

展示两个因子的截面相关性

Parameters

fac1 : pd.DataFrame 因子1 fac2 : pd.DataFrame 因子2 method : str, optional 计算相关系数的方法, by default "pearson" plt_plot : bool, optional 是否画出相关系数的时序变化图, by default 1 show_series : bool, optional 返回相关性的序列，而非均值 old_way : bool, optional 使用3.x版本的方式求相关系数

Returns

float 平均截面相关系数

Source code in pure_ocean_breeze/labor/process.py
Python
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
def show_corr(
    fac1: pd.DataFrame,
    fac2: pd.DataFrame,
    method: str = "pearson",
    plt_plot: bool = 1,
    show_series: bool = 0,
    old_way: bool = 0,
) -> float:
    """展示两个因子的截面相关性

    Parameters
    ----------
    fac1 : pd.DataFrame
        因子1
    fac2 : pd.DataFrame
        因子2
    method : str, optional
        计算相关系数的方法, by default "pearson"
    plt_plot : bool, optional
        是否画出相关系数的时序变化图, by default 1
    show_series : bool, optional
        返回相关性的序列，而非均值
    old_way : bool, optional
        使用3.x版本的方式求相关系数

    Returns
    -------
    `float`
        平均截面相关系数
    """
    if old_way:
        if method == "spearman":
            corr = show_x_with_func(fac1, fac2, lambda x: x.rank().corr().iloc[0, 1])
        else:
            corr = show_x_with_func(
                fac1, fac2, lambda x: x.corr(method=method).iloc[0, 1]
            )
    else:
        corr = fac1.corrwith(fac2, axis=1, method=method)
    if show_series:
        return corr
    else:
        if plt_plot:
            corr.plot(rot=60)
            plt.show()
        return corr.mean()

show_corrs(factors, factor_names=None, print_bool=True, show_percent=True, method='pearson')

展示很多因子两两之间的截面相关性

Parameters

factors : List[pd.DataFrame] 所有因子构成的列表, by default None factor_names : List[str], optional 上述因子依次的名字, by default None print_bool : bool, optional 是否打印出两两之间相关系数的表格, by default True show_percent : bool, optional 是否以百分数的形式展示, by default True method : str, optional 计算相关系数的方法, by default "pearson"

Returns

pd.DataFrame 两两之间相关系数的表格

Source code in pure_ocean_breeze/labor/process.py
Python
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
def show_corrs(
    factors: List[pd.DataFrame],
    factor_names: List[str] = None,
    print_bool: bool = True,
    show_percent: bool = True,
    method: str = "pearson",
) -> pd.DataFrame:
    """展示很多因子两两之间的截面相关性

    Parameters
    ----------
    factors : List[pd.DataFrame]
        所有因子构成的列表, by default None
    factor_names : List[str], optional
        上述因子依次的名字, by default None
    print_bool : bool, optional
        是否打印出两两之间相关系数的表格, by default True
    show_percent : bool, optional
        是否以百分数的形式展示, by default True
    method : str, optional
        计算相关系数的方法, by default "pearson"

    Returns
    -------
    `pd.DataFrame`
        两两之间相关系数的表格
    """
    corrs = []
    for i in range(len(factors)):
        main_i = factors[i]
        follows = factors[i + 1 :]
        corr = [show_corr(main_i, i, plt_plot=False, method=method) for i in follows]
        corr = [np.nan] * (i + 1) + corr
        corrs.append(corr)
    if factor_names is None:
        factor_names = [f"fac{i}" for i in list(range(1, len(factors) + 1))]
    corrs = pd.DataFrame(corrs, columns=factor_names, index=factor_names)
    np.fill_diagonal(corrs.to_numpy(), 1)
    if show_percent:
        pcorrs = corrs.applymap(to_percent)
    else:
        pcorrs = corrs.copy()
    if print_bool:
        return pcorrs
    else:
        return corrs

show_corrs_with_old(df=None, method='pearson', only_new=1, with_son_factors=1, freq='M', old_database=0)

计算新因子和已有因子的相关系数

Parameters

df : pd.DataFrame, optional 新因子, by default None method : str, optional 求相关系数的方法, by default 'pearson' only_new : bool, optional 仅计算新因子与旧因子之间的相关系数, by default 1 with_son_factors : bool, optional 计算新因子与数据库中各个细分因子的相关系数, by default 1 freq : str, optional 读取因子数据的频率, by default 'M' old_database : bool, optional 使用3.x版本的数据库, by default 0

Returns

pd.DataFrame 相关系数矩阵

Source code in pure_ocean_breeze/labor/process.py
Python
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
@do_on_dfs
def show_corrs_with_old(
    df: pd.DataFrame = None,
    method: str = "pearson",
    only_new: bool = 1,
    with_son_factors: bool = 1,
    freq: str = "M",
    old_database: bool = 0,
) -> pd.DataFrame:
    """计算新因子和已有因子的相关系数

    Parameters
    ----------
    df : pd.DataFrame, optional
        新因子, by default None
    method : str, optional
        求相关系数的方法, by default 'pearson'
    only_new : bool, optional
        仅计算新因子与旧因子之间的相关系数, by default 1
    with_son_factors : bool, optional
        计算新因子与数据库中各个细分因子的相关系数, by default 1
    freq : str, optional
        读取因子数据的频率, by default 'M'
    old_database : bool, optional
        使用3.x版本的数据库, by default 0


    Returns
    -------
    pd.DataFrame
        相关系数矩阵
    """
    if df is not None:
        df0 = df.resample(freq).last()
        if df.shape[0] / df0.shape[0] > 2:
            daily = 1
        else:
            daily = 0
    if old_database:
        nums = os.listdir(homeplace.final_factor_file)
        nums = sorted(
            set(
                [
                    int(i.split("多因子")[1].split("_月")[0])
                    for i in nums
                    if i.endswith("月.parquet")
                ]
            )
        )
        olds = []
        for i in nums:
            try:
                if daily:
                    old = database_read_final_factors(order=i)[0]
                else:
                    old = database_read_final_factors(order=i)[0].resample("M").last()
                olds.append(old)
            except Exception:
                break
        if df is not None:
            if only_new:
                corrs = [
                    to_percent(show_corr(df, i, plt_plot=0, method=method))
                    for i in olds
                ]
                corrs = pd.Series(corrs, index=[f"old{i}" for i in nums])
                corrs = corrs.to_frame(f"{method}相关系数").T
            else:
                olds = [df] + olds
                corrs = show_corrs(
                    olds, ["new"] + [f"old{i}" for i in nums], method=method
                )
        else:
            corrs = show_corrs(olds, [f"old{i}" for i in nums], method=method)
    else:
        qdb = Questdb()
        if freq == "M":
            factor_infos = qdb.get_data("select * from factor_infos where freq='月'")
        else:
            factor_infos = qdb.get_data("select * from factor_infos where freq='周'")
        if not with_son_factors:
            old_orders = list(set(factor_infos.order))
            if daily:
                olds = [FactorDone(order=i)() for i in old_orders]
            else:
                olds = [FactorDone(order=i)().resample(freq).last() for i in old_orders]
        else:
            old_orders = [
                i.order + i.son_name.replace("因子", "")
                for i in factor_infos.dropna().itertuples()
            ]
            if daily:
                olds = [
                    FactorDone(order=i.order)(i.son_name)
                    for i in factor_infos.dropna().itertuples()
                ]
            else:
                olds = [
                    FactorDone(order=i.order)(i.son_name).resample(freq).last()
                    for i in factor_infos.dropna().itertuples()
                ]
        if df is not None:
            if only_new:
                corrs = [
                    to_percent(show_corr(df, i, plt_plot=0, method=method))
                    for i in olds
                ]
                corrs = pd.Series(corrs, index=old_orders)
                corrs = corrs.to_frame(f"{method}相关系数")
                if corrs.shape[0] <= 30:
                    ...
                elif corrs.shape[0] <= 60:
                    corrs = corrs.reset_index()
                    corrs.columns = ["因子名称", "相关系数"]
                    corrs1 = corrs.iloc[:30, :]
                    corrs2 = corrs.iloc[30:, :].reset_index(drop=True)
                    corrs = pd.concat([corrs1, corrs2], axis=1).fillna("")
                elif corrs.shape[0] <= 90:
                    corrs = corrs.reset_index()
                    corrs.columns = ["因子名称", "相关系数"]
                    corrs1 = corrs.iloc[:30, :]
                    corrs2 = corrs.iloc[30:60, :].reset_index(drop=True)
                    corrs3 = corrs.iloc[60:90, :].reset_index(drop=True)
                    corrs = pd.concat([corrs1, corrs2, corrs3], axis=1).fillna("")
            else:
                olds = [df] + olds
                corrs = show_corrs(olds, old_orders, method=method)
        else:
            corrs = show_corrs(olds, old_orders, method=method)
    return corrs.sort_index()

show_cov(fac1, fac2, plt_plot=1, show_series=0)

展示两个因子的截面相关性

Parameters

fac1 : pd.DataFrame 因子1 fac2 : pd.DataFrame 因子2 method : str, optional 计算相关系数的方法, by default "spearman" plt_plot : bool, optional 是否画出相关系数的时序变化图, by default 1 show_series : bool, optional 返回相关性的序列，而非均值

Returns

float 平均截面相关系数

Source code in pure_ocean_breeze/labor/process.py
Python
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
def show_cov(
    fac1: pd.DataFrame,
    fac2: pd.DataFrame,
    plt_plot: bool = 1,
    show_series: bool = 0,
) -> float:
    """展示两个因子的截面相关性

    Parameters
    ----------
    fac1 : pd.DataFrame
        因子1
    fac2 : pd.DataFrame
        因子2
    method : str, optional
        计算相关系数的方法, by default "spearman"
    plt_plot : bool, optional
        是否画出相关系数的时序变化图, by default 1
    show_series : bool, optional
        返回相关性的序列，而非均值

    Returns
    -------
    `float`
        平均截面相关系数
    """
    cov = show_x_with_func(fac1, fac2, lambda x: x.cov().iloc[0, 1])
    if show_series:
        return cov
    else:
        if plt_plot:
            cov.plot(rot=60)
            plt.show()
        return cov.mean()

show_covs(factors, factor_names=None, print_bool=True, show_percent=True)

展示很多因子两两之间的截面相关性

Parameters

factors : List[pd.DataFrame] 所有因子构成的列表, by default None factor_names : List[str], optional 上述因子依次的名字, by default None print_bool : bool, optional 是否打印出两两之间相关系数的表格, by default True show_percent : bool, optional 是否以百分数的形式展示, by default True

Returns

pd.DataFrame 两两之间相关系数的表格

Source code in pure_ocean_breeze/labor/process.py
Python
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
def show_covs(
    factors: List[pd.DataFrame],
    factor_names: List[str] = None,
    print_bool: bool = True,
    show_percent: bool = True,
) -> pd.DataFrame:
    """展示很多因子两两之间的截面相关性

    Parameters
    ----------
    factors : List[pd.DataFrame]
        所有因子构成的列表, by default None
    factor_names : List[str], optional
        上述因子依次的名字, by default None
    print_bool : bool, optional
        是否打印出两两之间相关系数的表格, by default True
    show_percent : bool, optional
        是否以百分数的形式展示, by default True

    Returns
    -------
    `pd.DataFrame`
        两两之间相关系数的表格
    """
    corrs = []
    for i in range(len(factors)):
        main_i = factors[i]
        follows = factors[i + 1 :]
        corr = [show_cov(main_i, i, plt_plot=False) for i in follows]
        corr = [np.nan] * (i + 1) + corr
        corrs.append(corr)
    if factor_names is None:
        factor_names = [f"fac{i}" for i in list(range(1, len(factors) + 1))]
    corrs = pd.DataFrame(corrs, columns=factor_names, index=factor_names)
    np.fill_diagonal(corrs.to_numpy(), 1)
    if show_percent:
        pcorrs = corrs.applymap(to_percent)
    else:
        pcorrs = corrs.copy()
    if print_bool:
        print(pcorrs)
    return corrs

show_x_with_func(fac1, fac2, func)

展示两个因子的某种截面关系

Parameters

fac1 : pd.DataFrame 因子1 fac2 : pd.DataFrame 因子2 func : Callable 要对两个因子在截面上的进行的操作

Returns

pd.Series 截面关系

Source code in pure_ocean_breeze/labor/process.py
Python
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
def show_x_with_func(
    fac1: pd.DataFrame,
    fac2: pd.DataFrame,
    func: Callable,
) -> pd.Series:
    """展示两个因子的某种截面关系

    Parameters
    ----------
    fac1 : pd.DataFrame
        因子1
    fac2 : pd.DataFrame
        因子2
    func : Callable
        要对两个因子在截面上的进行的操作

    Returns
    -------
    `pd.Series`
        截面关系
    """
    the_func = partial(func)
    both1 = fac1.stack().reset_index()
    befo1 = fac2.stack().reset_index()
    both1.columns = ["date", "code", "both"]
    befo1.columns = ["date", "code", "befo"]
    twins = pd.merge(both1, befo1, on=["date", "code"]).set_index(["date", "code"])
    corr = twins.groupby("date").apply(the_func)
    return corr

symmetrically_orthogonalize(dfs)

对多个因子做对称正交，每个因子得到正交其他因子后的结果

Parameters

dfs : list[pd.DataFrame] 多个要做正交的因子，每个df都是index为时间，columns为股票代码，values为因子值的df

Returns

list[pd.DataFrame] 对称正交后的各个因子

Source code in pure_ocean_breeze/labor/process.py
Python
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
def symmetrically_orthogonalize(dfs: list[pd.DataFrame]) -> list[pd.DataFrame]:
    """对多个因子做对称正交，每个因子得到正交其他因子后的结果

    Parameters
    ----------
    dfs : list[pd.DataFrame]
        多个要做正交的因子，每个df都是index为时间，columns为股票代码，values为因子值的df

    Returns
    -------
    list[pd.DataFrame]
        对称正交后的各个因子
    """

    def sing(dfs: list[pd.DataFrame], date: pd.Timestamp):
        dds = []
        for num, i in enumerate(dfs):
            i = i[i.index == date]
            i.index = [f"fac{num}"]
            i = i.T
            dds.append(i)
        dds = pd.concat(dds, axis=1)
        cov = dds.cov()
        d, u = np.linalg.eig(cov)
        d = np.diag(d ** (-0.5))
        new_facs = pd.DataFrame(
            np.dot(dds, np.dot(np.dot(u, d), u.T)), columns=dds.columns, index=dds.index
        )
        new_facs = new_facs.stack().reset_index()
        new_facs.columns = ["code", "fac_number", "fac"]
        new_facs = new_facs.assign(date=date)
        dds = []
        for num, i in enumerate(dfs):
            i = new_facs[new_facs.fac_number == f"fac{num}"]
            i = i.pivot(index="date", columns="code", values="fac")
            dds.append(i)
        return dds

    dfs = [standardlize(i) for i in dfs]
    date_first = max([i.index.min() for i in dfs])
    date_last = min([i.index.max() for i in dfs])
    dfs = [i[(i.index >= date_first) & (i.index <= date_last)] for i in dfs]
    fac_num = len(dfs)
    ddss = [[] for i in range(fac_num)]
    for date in tqdm.auto.tqdm(dfs[0].index):
        dds = sing(dfs, date)
        for num, i in enumerate(dds):
            ddss[num].append(i)
    ds = []
    for i in tqdm.auto.tqdm(ddss):
        ds.append(pd.concat(i))
    return ds

test_on_300500(df, trade_cost_double_side=0, group_num=10, value_weighted=1, boxcox=0, hs300=0, zz500=0, zz1000=0, gz2000=0, iplot=1, opens_average_first_day=0, total_cap=0)

对因子在指数成分股内进行多空和多头测试

Parameters

df : pd.DataFrame 因子值，index为时间，columns为股票代码 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 group_num : int 分组数量, by default 10 value_weighted : bool 是否进行流通市值加权, by default 0 hs300 : bool, optional 在沪深300成分股内测试, by default 0 zz500 : bool, optional 在中证500成分股内测试, by default 0 zz1000 : bool, optional 在中证1000成分股内测试, by default 0 gz1000 : bool, optional 在国证2000成分股内测试, by default 0 iplot : bo0l,optional 多空回测的时候，是否使用cufflinks绘画 opens_average_first_day : bool, optional 买入时使用第一天的平均价格, by default 0 total_cap : bool, optional 加权和行业市值中性化时使用总市值, by default 0

Returns

pd.Series 多头组在该指数上的超额收益序列

Source code in pure_ocean_breeze/labor/process.py
Python
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
@do_on_dfs
def test_on_300500(
    df: pd.DataFrame,
    trade_cost_double_side: float = 0,
    group_num: int = 10,
    value_weighted: bool = 1,
    boxcox: bool = 0,
    hs300: bool = 0,
    zz500: bool = 0,
    zz1000: bool = 0,
    gz2000: bool = 0,
    iplot: bool = 1,
    opens_average_first_day: bool = 0,
    total_cap: bool = 0,
) -> pd.Series:
    """对因子在指数成分股内进行多空和多头测试

    Parameters
    ----------
    df : pd.DataFrame
        因子值，index为时间，columns为股票代码
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    group_num : int
        分组数量, by default 10
    value_weighted : bool
        是否进行流通市值加权, by default 0
    hs300 : bool, optional
        在沪深300成分股内测试, by default 0
    zz500 : bool, optional
        在中证500成分股内测试, by default 0
    zz1000 : bool, optional
        在中证1000成分股内测试, by default 0
    gz1000 : bool, optional
        在国证2000成分股内测试, by default 0
    iplot : bo0l,optional
        多空回测的时候，是否使用cufflinks绘画
    opens_average_first_day : bool, optional
        买入时使用第一天的平均价格, by default 0
    total_cap : bool, optional
        加权和行业市值中性化时使用总市值, by default 0

    Returns
    -------
    pd.Series
        多头组在该指数上的超额收益序列
    """
    fi300 = daily_factor_on300500(
        df, hs300=hs300, zz500=zz500, zz1000=zz1000, gz2000=gz2000
    )
    shen = pure_moonnight(
        fi300,
        value_weighted=value_weighted,
        groups_num=group_num,
        trade_cost_double_side=trade_cost_double_side,
        boxcox=boxcox,
        iplot=iplot,
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
    )
    if (
        shen.shen.group_net_values.group1.iloc[-1]
        > shen.shen.group_net_values[f"group{group_num}"].iloc[-1]
    ):
        print(
            make_relative_comments(
                shen.shen.group_rets.group1,
                hs300=hs300,
                zz500=zz500,
                zz1000=zz1000,
                gz2000=gz2000,
            )
        )
        abrets = make_relative_comments_plot(
            shen.shen.group_rets.group1,
            hs300=hs300,
            zz500=zz500,
            zz1000=zz1000,
            gz2000=gz2000,
        )
        return abrets
    else:
        print(
            make_relative_comments(
                shen.shen.group_rets[f"group{group_num}"],
                hs300=hs300,
                zz500=zz500,
                zz1000=zz1000,
                gz2000=gz2000,
            )
        )
        abrets = make_relative_comments_plot(
            shen.shen.group_rets[f"group{group_num}"],
            hs300=hs300,
            zz500=zz500,
            zz1000=zz1000,
            gz2000=gz2000,
        )
        return abrets

test_on_index_four(df, value_weighted=1, group_num=10, trade_cost_double_side=0, iplot=1, gz2000=0, boxcox=1, opens_average_first_day=0, total_cap=0)

对因子同时在沪深300、中证500、中证1000、国证2000这4个指数成分股内进行多空和多头超额测试

Parameters

df : pd.DataFrame 因子值，index为时间，columns为股票代码 value_weighted : bool 是否进行流通市值加权, by default 0 group_num : int 分组数量, by default 10 trade_cost_double_side : float, optional 交易的双边手续费率, by default 0 iplot : bol,optional 多空回测的时候，是否使用cufflinks绘画 gz2000 : bool, optional 是否进行国证2000上的测试, by default 0 boxcox : bool, optional 是否进行行业市值中性化处理, by default 1 opens_average_first_day : bool, optional 买入时使用第一天的平均价格, by default 0 total_cap : bool, optional 加权和行业市值中性化时使用总市值, by default 0

Returns

pd.DataFrame 多头组在各个指数上的超额收益序列

Source code in pure_ocean_breeze/labor/process.py
Python
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
@do_on_dfs
def test_on_index_four(
    df: pd.DataFrame,
    value_weighted: bool = 1,
    group_num: int = 10,
    trade_cost_double_side: float = 0,
    iplot: bool = 1,
    gz2000: bool = 0,
    boxcox: bool = 1,
    opens_average_first_day: bool = 0,
    total_cap: bool = 0,
) -> pd.DataFrame:
    """对因子同时在沪深300、中证500、中证1000、国证2000这4个指数成分股内进行多空和多头超额测试

    Parameters
    ----------
    df : pd.DataFrame
        因子值，index为时间，columns为股票代码
    value_weighted : bool
        是否进行流通市值加权, by default 0
    group_num : int
        分组数量, by default 10
    trade_cost_double_side : float, optional
        交易的双边手续费率, by default 0
    iplot : bol,optional
        多空回测的时候，是否使用cufflinks绘画
    gz2000 : bool, optional
        是否进行国证2000上的测试, by default 0
    boxcox : bool, optional
        是否进行行业市值中性化处理, by default 1
    opens_average_first_day : bool, optional
        买入时使用第一天的平均价格, by default 0
    total_cap : bool, optional
        加权和行业市值中性化时使用总市值, by default 0

    Returns
    -------
    pd.DataFrame
        多头组在各个指数上的超额收益序列
    """
    fi300 = daily_factor_on300500(df, hs300=1)
    shen = pure_moonnight(
        fi300,
        groups_num=group_num,
        value_weighted=value_weighted,
        trade_cost_double_side=trade_cost_double_side,
        iplot=iplot,
        boxcox=boxcox,
        opens_average_first_day=opens_average_first_day,
        total_cap=total_cap,
    )
    if (
        shen.shen.group_net_values.group1.iloc[-1]
        > shen.shen.group_net_values[f"group{group_num}"].iloc[-1]
    ):
        com300, net300 = make_relative_comments(
            shen.shen.group_rets.group1, hs300=1, show_nets=1
        )
        fi500 = daily_factor_on300500(df, zz500=1)
        shen = pure_moonnight(
            fi500,
            groups_num=group_num,
            value_weighted=value_weighted,
            trade_cost_double_side=trade_cost_double_side,
            iplot=iplot,
            boxcox=boxcox,
            opens_average_first_day=opens_average_first_day,
            total_cap=total_cap,
        )
        com500, net500 = make_relative_comments(
            shen.shen.group_rets.group1, zz500=1, show_nets=1
        )
        fi1000 = daily_factor_on300500(df, zz1000=1)
        shen = pure_moonnight(
            fi1000,
            groups_num=group_num,
            value_weighted=value_weighted,
            trade_cost_double_side=trade_cost_double_side,
            iplot=iplot,
            boxcox=boxcox,
            opens_average_first_day=opens_average_first_day,
            total_cap=total_cap,
        )
        com1000, net1000 = make_relative_comments(
            shen.shen.group_rets.group1, zz1000=1, show_nets=1
        )
        if gz2000:
            fi2000 = daily_factor_on300500(df, gz2000=1)
            shen = pure_moonnight(
                fi2000,
                groups_num=group_num,
                trade_cost_double_side=trade_cost_double_side,
                iplot=iplot,
                boxcox=boxcox,
                opens_average_first_day=opens_average_first_day,
                total_cap=total_cap,
            )
            com2000, net2000 = make_relative_comments(
                shen.shen.group_rets.group1, gz2000=1, show_nets=1
            )
    else:
        com300, net300 = make_relative_comments(
            shen.shen.group_rets[f"group{group_num}"], hs300=1, show_nets=1
        )
        fi500 = daily_factor_on300500(df, zz500=1)
        shen = pure_moonnight(
            fi500,
            groups_num=group_num,
            value_weighted=value_weighted,
            trade_cost_double_side=trade_cost_double_side,
            iplot=iplot,
            boxcox=boxcox,
            opens_average_first_day=opens_average_first_day,
            total_cap=total_cap,
        )
        com500, net500 = make_relative_comments(
            shen.shen.group_rets[f"group{group_num}"], zz500=1, show_nets=1
        )
        fi1000 = daily_factor_on300500(df, zz1000=1)
        shen = pure_moonnight(
            fi1000,
            groups_num=group_num,
            value_weighted=value_weighted,
            trade_cost_double_side=trade_cost_double_side,
            iplot=iplot,
            boxcox=boxcox,
            opens_average_first_day=opens_average_first_day,
            total_cap=total_cap,
        )
        com1000, net1000 = make_relative_comments(
            shen.shen.group_rets[f"group{group_num}"], zz1000=1, show_nets=1
        )
        if gz2000:
            fi2000 = daily_factor_on300500(df, gz2000=1)
            shen = pure_moonnight(
                fi2000,
                groups_num=group_num,
                value_weighted=value_weighted,
                trade_cost_double_side=trade_cost_double_side,
                iplot=iplot,
                boxcox=boxcox,
                opens_average_first_day=opens_average_first_day,
                total_cap=total_cap,
            )
            com2000, net2000 = make_relative_comments(
                shen.shen.group_rets[f"group{group_num}"], gz2000=1, show_nets=1
            )
    com300 = com300.to_frame("300超额")
    com500 = com500.to_frame("500超额")
    com1000 = com1000.to_frame("1000超额")
    if gz2000:
        com2000 = com2000.to_frame("2000超额")
        coms = pd.concat([com300, com500, com1000, com2000], axis=1)
    else:
        coms = pd.concat([com300, com500, com1000], axis=1)
    coms = np.around(coms, 3)
    if gz2000:
        nets = pd.concat([net300, net500, net1000, net2000], axis=1)
        nets.columns = ["300超额", "500超额", "1000超额", "2000超额"]
    else:
        nets = pd.concat([net300, net500, net1000], axis=1)
        nets.columns = ["300超额", "500超额", "1000超额"]
    coms = coms.reset_index()
    if iplot:
        figs = cf.figures(
            nets,
            [dict(kind="line", y=list(nets.columns))],
            asList=True,
        )
        coms = coms.rename(columns={list(coms)[0]: "绩效指标"})
        table = FF.create_table(coms.iloc[::-1])
        table.update_yaxes(matches=None)
        figs.append(table)
        figs = [figs[-1]] + figs[:-1]
        figs[1].update_layout(
            legend=dict(yanchor="top", y=0.99, xanchor="left", x=0.01)
        )
        base_layout = cf.tools.get_base_layout(figs)
        if gz2000:
            sp = cf.subplots(
                figs,
                shape=(2, 10),
                base_layout=base_layout,
                vertical_spacing=0.15,
                horizontal_spacing=0.03,
                shared_yaxes=False,
                specs=[
                    [
                        None,
                        {"rowspan": 2, "colspan": 4},
                        None,
                        None,
                        None,
                        {"rowspan": 2, "colspan": 3},
                        None,
                        None,
                        None,
                        None,
                    ],
                    [
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                    ],
                ],
            )
        else:
            sp = cf.subplots(
                figs,
                shape=(2, 10),
                base_layout=base_layout,
                vertical_spacing=0.15,
                horizontal_spacing=0.03,
                shared_yaxes=False,
                specs=[
                    [
                        None,
                        {"rowspan": 2, "colspan": 3},
                        None,
                        None,
                        {"rowspan": 2, "colspan": 3},
                        None,
                        None,
                        None,
                        None,
                        None,
                    ],
                    [
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                        None,
                    ],
                ],
            )
        sp["layout"].update(showlegend=True)
        cf.iplot(sp)
    else:
        nets.plot()
        plt.show()
        tb = Texttable()
        tb.set_cols_width([8] + [7] + [8] * 2 + [7] * 2 + [8])
        tb.set_cols_dtype(["f"] * 7)
        tb.header(list(coms.T.reset_index().columns))
        tb.add_rows(coms.T.reset_index().to_numpy(), header=True)
        print(tb.draw())

to_tradeends(df)

将最后一个自然日改变为最后一个交易日

Parameters

df : pd.DataFrame index为时间，为每个月的最后一天

Returns

pd.DataFrame 修改为交易日标注后的pd.DataFrame

Source code in pure_ocean_breeze/labor/process.py
Python
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
@do_on_dfs
def to_tradeends(df: pd.DataFrame) -> pd.DataFrame:
    """将最后一个自然日改变为最后一个交易日

    Parameters
    ----------
    df : pd.DataFrame
        index为时间，为每个月的最后一天

    Returns
    -------
    `pd.DataFrame`
        修改为交易日标注后的pd.DataFrame
    """
    """"""
    start = df.index.min()
    start = start - pd.tseries.offsets.MonthBegin()
    start = datetime.datetime.strftime(start, "%Y%m%d")
    trs = read_daily(tr=1, start=start)
    trs = trs.assign(tradeends=list(trs.index))
    trs = trs[["tradeends"]]
    trs = trs.resample("M").last()
    df = pd.concat([trs, df], axis=1)
    df = df.set_index(["tradeends"])
    return df