siat 3.10.132__py3-none-any.whl → 3.11.1__py3-none-any.whl

build/lib/siat/holding_risk.py

@@ -1,867 +0,0 @@
-# -*- coding: utf-8 -*-
-"""
-本模块功能：投资组合的VaR(在险价值)和ES(预期损失)分析
-所属工具包：证券投资分析工具SIAT 
-SIAT：Security Investment Analysis Tool
-创建日期：2018年6月16日
-最新修订日期：2020年7月23日
-作者：王德宏 (WANG Dehong, Peter)
-作者单位：北京外国语大学国际商学院
-作者邮件：wdehong2000@163.com
-版权所有：王德宏
-用途限制：仅限研究与教学使用，不可商用！商用需要额外授权。
-特别声明：作者不对使用本工具进行证券投资导致的任何损益负责！
-"""
-#==============================================================================
-#统一屏蔽一般性警告
-import warnings; warnings.filterwarnings("ignore")     
-from siat.common import *
-from siat.translate import *
-from siat.security_prices import *
-from siat.security_price2 import *
-#==============================================================================
-import matplotlib.pyplot as plt
-
-#处理绘图汉字乱码问题
-import sys; czxt=sys.platform
-if czxt in ['win32','win64']:
-    plt.rcParams['font.sans-serif'] = ['SimHei']  # 设置默认字体
-    mpfrc={'font.family': 'SimHei'}
-
-if czxt in ['darwin']: #MacOSX
-    plt.rcParams['font.family']= ['Heiti TC']
-    mpfrc={'font.family': 'Heiti TC'}
-
-if czxt in ['linux']: #website Jupyter
-    plt.rcParams['font.family']= ['Heiti TC']
-    mpfrc={'font.family':'Heiti TC'}
-
-# 解决保存图像时'-'显示为方块的问题
-plt.rcParams['axes.unicode_minus'] = False 
-#==============================================================================
-#统一设定绘制的图片大小：数值为英寸，1英寸=100像素
-#plt.rcParams['figure.figsize']=(12.8,7.2)
-plt.rcParams['figure.figsize']=(12.8,6.4)
-plt.rcParams['figure.dpi']=300
-plt.rcParams['font.size'] = 13
-plt.rcParams['xtick.labelsize']=11 #横轴字体大小
-plt.rcParams['ytick.labelsize']=11 #纵轴字体大小
-
-title_txt_size=16
-ylabel_txt_size=14
-xlabel_txt_size=14
-legend_txt_size=14
-
-#设置绘图风格：网格虚线
-plt.rcParams['axes.grid']=True
-#plt.rcParams['grid.color']='steelblue'
-#plt.rcParams['grid.linestyle']='dashed'
-#plt.rcParams['grid.linewidth']=0.5
-#plt.rcParams['axes.facecolor']='whitesmoke'
-
-#==============================================================================
-import os, sys
-class HiddenPrints:
-    def __enter__(self):
-        self._original_stdout = sys.stdout
-        sys.stdout = open(os.devnull, 'w')
-
-    def __exit__(self, exc_type, exc_val, exc_tb):
-        sys.stdout.close()
-        sys.stdout = self._original_stdout
-#==============================================================================
-
-#==============================================================================
-if __name__ == '__main__':
-    portfolio={'Market':('China','000001.SS'),'300782.SZ':1}
-    fromdate='2022-1-1'
-    todate='2022-4-18'
-
-def get_portfolio_prices0(portfolio,fromdate,todate):
-    """
-    功能：抓取投资组合portfolio的每日价值
-    输入：投资组合portfolio，开始日期，结束日期
-    fromdate: 样本开始日期。格式：'YYYY-MM-DD'
-    todate: 样本结束日期。既可以是今天日期，也可以是一个历史日期    
-    
-    输出：投资组合的价格序列，按照日期升序排列
-    """
-    
-    #解构投资组合
-    _,mktidx,tickerlist,sharelist,ticker_type=decompose_portfolio(portfolio)
-    
-    #检查股票列表个数与份额列表个数是否一致
-    if len(tickerlist) != len(sharelist):
-        print("#Error(get_portfolio_prices0): numbers of stocks and shares mismatch.")
-        return None        
-    
-    #抓取股票价格
-    p=get_prices(tickerlist,fromdate,todate)
-    if p is None: return None
-    
-    import pandas as pd
-    #计算投资组合的开盘价
-    op=p[['Open']]
-    #计算投资组合的价值
-    oprice=pd.DataFrame(op.dot(sharelist))
-    oprice.rename(columns={0: 'Open'}, inplace=True)    
-
-    #计算投资组合的收盘价
-    cp=p[['Close']]
-    #计算投资组合的价值
-    cprice=pd.DataFrame(cp.dot(sharelist))
-    cprice.rename(columns={0: 'Close'}, inplace=True) 
-    
-    #计算投资组合的调整收盘价
-    acp=p[['Adj Close']]
-    #计算投资组合的价值
-    acprice=pd.DataFrame(acp.dot(sharelist))
-    acprice.rename(columns={0: 'Adj Close'}, inplace=True) 
-    
-    #计算投资组合的交易量
-    vol=p[['Volume']]
-    #计算投资组合的价值
-    pfvol=pd.DataFrame(vol.dot(sharelist))
-    pfvol.rename(columns={0: 'Volume'}, inplace=True) 
-    
-    #计算投资组合的交易金额
-    if len(tickerlist)==1:
-        p['Amount']=p['Close']*p['Volume']
-    else:
-        for t in tickerlist:
-            p['Amount',t]=p['Close',t]*p['Volume',t]
-    amt=p[['Amount']]
-    #计算投资组合的价值
-    pfamt=pd.DataFrame(amt.dot(sharelist))
-    pfamt.rename(columns={0: 'Amount'}, inplace=True) 
-
-    #合成开盘价、收盘价、调整收盘价、交易量和交易金额
-    pf1=pd.merge(oprice,cprice,how='inner',left_index=True,right_index=True)    
-    pf2=pd.merge(pf1,acprice,how='inner',left_index=True,right_index=True)
-    pf3=pd.merge(pf2,pfvol,how='inner',left_index=True,right_index=True)
-    pf4=pd.merge(pf3,pfamt,how='inner',left_index=True,right_index=True)
-    pf4['Ret']=pf4['Close'].pct_change()
-
-    #获得期间的市场收益率：假设无风险收益率非常小，可以忽略
-    m=get_prices(mktidx,fromdate,todate)
-    m['Mkt']=m['Close'].pct_change()
-    rf_df=m[['Mkt']]
-    
-    #合并pf4与rf_df
-    prices=pd.merge(pf4,rf_df,how='left',left_index=True,right_index=True)
-
-    #提取日期和星期几
-    prices['Date']=prices.index.strftime("%Y-%m-%d")
-    prices['Weekday']=prices.index.weekday+1
-
-    prices['Portfolio']=str(tickerlist)
-    prices['Shares']=str(sharelist)
-    prices['Adjustment']=prices.apply(lambda x: \
-          False if x['Close']==x['Adj Close'] else True, axis=1)
-
-    pfdf=prices[['Portfolio','Shares','Date','Weekday', \
-                 'Open','Close','Adj Close','Adjustment', \
-                'Volume','Amount','Ret','Mkt']]  
-
-    #判断空值，控制空值可能引起的程序崩溃
-    if pfdf is None:
-        print("#Error(get_portfolio_prices0): failed to retrieve portfolio data")
-        return None
-    pfdf.dropna(inplace=True)
-    if (pfdf is None) or (len(pfdf)==0):
-        print("#Error(get_portfolio_prices0): failed to retrieve portfolio data")
-        return None   
-
-    return pfdf      
-
-    
-#==============================================================================
-def calc_VaR_normal_standard(ret_series,future_days=1,alpha=0.99):
-    
-    """
-    功能：VaR算法之标准正态法
-    输入参数：收益率序列(非百分比)，未来持有时间(天)，置信度
-    输出参数：VaR比率，正数
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    import numpy as np
-    from scipy import stats
-    
-    z=stats.norm.ppf(1-alpha)
-    
-    miu_daily=np.mean(r)
-    miu_days=np.power(miu_daily+1,future_days)-1
-    
-    sigma_daily=np.std(r)    
-    sigma_days=np.sqrt(future_days)*sigma_daily
-    
-    VaR_ratio=abs(miu_days+z*sigma_days)
-
-    return VaR_ratio
-
-#==============================================================================
-def calc_ES_normal_standard(ret_series,future_days=1,alpha=0.99):
-    """
-    功能：计算ES，标准正态法
-    输入参数：收益率序列，未来持有日期，置信度
-    输出：ES比率，正数
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    import numpy as np
-    from scipy import stats
-    
-    z=stats.norm.ppf(1-alpha)
-    miu_daily=np.mean(r)
-    miu_days=np.power(miu_daily+1,future_days)-1
-    
-    sigma_daily=np.std(r)    
-    sigma_days=np.sqrt(future_days)*sigma_daily
-    
-    zES=-stats.norm.pdf(z)/(1-alpha)
-    ratio=abs(miu_days+zES*sigma_days)
-    
-    return ratio
-
-#==============================================================================
-def calc_VaR_normal_modified(ret_series,future_days=1,alpha=0.99):
-    """
-    功能：VaR算法之修正正态法
-    #输入参数：日收益率序列(非百分比)，未来持有日期，置信度
-    #输出参数：VaR比率，正数
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    from scipy import stats
-    import numpy as np
-
-    z=np.abs(stats.norm.ppf(1-alpha))
-    S=stats.skew(r)
-    K=stats.kurtosis(r)
-    
-    t1=1/6*(np.power(z,2)-1)*S
-    t2=1/24*(np.power(z,3)-3*z)*K
-    t3=1/36*(2*np.power(z,3)-5*z)*np.power(S,2)
-    t=z+t1+t2-t3
-    
-    miu_daily=np.mean(r)
-    miu_days=np.power(miu_daily+1,future_days)-1
-    sigma_daily=np.std(r)    
-    sigma_days=np.sqrt(future_days)*sigma_daily
-    
-    ratio=abs(miu_days+t*sigma_days)
-    
-    return ratio
-
-#==============================================================================
-def calc_ES_normal_modified(ret_series,future_days=1,alpha=0.99):
-    """
-    功能：ES算法之修正正态法
-    #输入参数：日收益率序列(非百分比)，未来持有日期，置信度
-    #输出参数：ES比率，正数
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    from scipy import stats
-    import numpy as np
-    
-    #计算替代z的t值
-    z=np.abs(stats.norm.ppf(1-alpha))
-    S=stats.skew(r)
-    K=stats.kurtosis(r)
-    
-    t1=1/6*(np.power(z,2)-1)*S
-    t2=1/24*(np.power(z,3)-3*z)*K
-    t3=1/36*(2*np.power(z,3)-5*z)*np.power(S,2)
-    t=z+t1+t2-t3
-    
-    miu_daily=np.mean(r)
-    miu_days=np.power(miu_daily+1,future_days)-1
-    sigma_daily=np.std(r)    
-    sigma_days=np.sqrt(future_days)*sigma_daily
-    
-    #使用t替代原来的z    
-    zES=-stats.norm.pdf(t)/(1-alpha)
-    ratio=abs(miu_days+zES*sigma_days)
-    
-    return ratio
-
-#==============================================================================
-def get_grouped_rets(ret_series,groupsize=1):
-    """
-    功能：给定收益率序列pfdf，按照组的大小，在pfdf内创建滚动分组，
-    计算每组组内的累计收益率，消除空值，返回组收益率序列
-    """
-    #测试用数据，测试后应注释掉
-    """
-    portfolio={'Market':('US','^GSPC'),'AAPL':0.5,'MSFT':0.3,'IBM':0.2}
-    fromdate='2019-7-31'
-    todate  ='2019-8-31'
-    groupsize=2
-    #获得投资组合的收益率
-    pfdf=get_portfolio_prices0(portfolio,fromdate,todate)    
-    import pandas as pd
-    ret_series=pd.Series(pfdf['Ret'])
-    """    
-
-    #定义组内累计收益率计算方法
-    cumret=lambda x:(x+1.0).prod()-1.0
-    #使用滚动窗口分组，计算各组收益率gret    
-    gret_series=ret_series.rolling(groupsize).apply(cumret)
-    gret_series=gret_series[~gret_series.isnull()]
-    
-    #返回组收益率序列
-    return gret_series
-
-#==============================================================================
-def calc_VaR_historical(ret_series,future_days=1,alpha=0.99):
-    """
-    功能：VaR算法之历史模拟法
-    输入参数：历史日收益率序列，未来持有日期，置信度
-    输出：VaR比率，正数
-    """
-    #去掉空值
-    r0=ret_series[~ret_series.isnull()]
-    #按未来持有期间数计算分组收益率
-    r=get_grouped_rets(r0,future_days)
-    
-    import numpy as np
-    n=len(r)
-    t=int(n*(1-alpha))
-    SR=np.sort(r)
-    
-    if t>=1:
-        A=SR[t-1]   #SR的第一个元素的序号是0
-    else:
-        A=SR[0]
-    
-    VaR_ratio=abs(A)
-    
-    return VaR_ratio
-
-#==============================================================================
-def calc_ES_historical(ret_series,future_days=1,alpha=0.99):
-    """
-    功能：ES算法之历史模拟法
-    输入参数：历史日收益率序列，未来持有日期，置信度
-    输出：ES比率，正数
-    要求：足够多的历史数据
-    """
-    #去掉空值
-    r0=ret_series[~ret_series.isnull()]
-    #按未来持有期间数计算分组收益率
-    r=get_grouped_rets(r0,future_days)
-    
-    import numpy as np
-    n=len(r)
-    t=int(n*(1-alpha))
-    SR=np.sort(r)
-    
-    if t>2:
-        #SR中第t个元素是VaR，第0~(t-1)个元素的均值是ES
-        #SR的第一个元素的序号是0
-        A=np.mean(SR[0:(t-2)])  
-    else:
-        A=SR[0]
-    
-    ratio=abs(A)
-    
-    return ratio
-
-if __name__ == '__main__':
-    Market={'Market':('China','000001.SS')}
-    Stocks={'300782.SZ':2,'300661.SZ':3,'688019.SS':4}
-    portfolio=dict(Market,**Stocks)
-    prices=get_portfolio_prices0(portfolio,'2019-7-20','2020-7-20')
-    ret_series=prices['Ret']
-    future_days=3
-    alpha=0.99
-    
-#==============================================================================
-def calc_VaR_montecarlo(ret_series,future_days=1,alpha=0.99, \
-                        random=10000,mctype='random'):
-    """
-    功能：VaR算法之蒙特卡洛模拟法，持有多日
-    输入参数：历史日收益率序列，未来持有天数，置信度，重复模拟次数，模拟类型
-    注：重复模拟次数越多，准确率就越高，但耗时也越多
-    输出：持有多天的VaR(金额)
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    import pandas as pd
-    import numpy as np
-    #蒙特卡洛模拟类型：随机数or超采样
-    if mctype=='random':    #随机数产生新的序列
-        #取得历史日收益率的均值和标准差
-        miu=np.mean(r)
-        sigma=np.std(r)
-        
-        #指定随机数种子
-        np.random.seed(12345)
-        #按照历史日收益率的均值和标准差重复模拟一定次数，生成新的日收益率序列
-        RR=pd.Series(np.random.normal(miu,sigma,random))
-    else:   #超采样产生新的序列
-        #将收益率序列转变为df
-        r1=pd.DataFrame(r)
-        r2=r1.sample(n=random,replace=True)
-        r2.sort_index(inplace=True)
-        RR=pd.Series(r2.iloc[:,0])
-        """
-        r1=np.random.choice(r,size=random,replace=True)
-        RR=pd.Series(r1)
-        """
-        
-    #基于新的日收益率序列，使用标准正态法计算VaR
-    ratio=calc_VaR_normal_standard(RR,future_days,alpha)        
-    #ratio=calc_VaR_historical(RR,future_days,alpha)
-    
-    return ratio
-
-if __name__ == '__main__':
-    random=1000
-    mctype='oversampling'
-
-#==============================================================================
-def calc_ES_montecarlo(ret_series,future_days=1,alpha=0.99, \
-                        random=10000,mctype='random'):
-    """
-    功能：ES算法之蒙特卡洛模拟法，持有多日
-    输入参数：历史日收益率序列，未来持有天数，置信度，重复模拟次数，模拟类型
-    注：重复模拟次数越多，准确率就越高，但耗时也越多
-    输出：持有多天的ES比率
-    """
-    #去掉空值
-    r=ret_series[~ret_series.isnull()]
-    
-    import pandas as pd
-    import numpy as np
-    #蒙特卡洛模拟类型：随机数or超采样
-    if mctype=='random':    #随机数产生新的序列
-        #取得历史日收益率的均值和标准差
-        miu=np.mean(r)
-        sigma=np.std(r)
-        
-        #指定随机数种子
-        np.random.seed(12345)
-        #按照历史日收益率的均值和标准差重复模拟一定次数，生成新的日收益率序列
-        RR=pd.Series(np.random.normal(miu,sigma,random))
-    else:   #超采样产生新的序列
-        #将收益率序列转变为df
-        r1=pd.DataFrame(r)
-        r2=r1.sample(n=random,replace=True)
-        r2.sort_index(inplace=True)
-        RR=pd.Series(r2.iloc[:,0])
-        """
-        r1=np.random.choice(r,size=random,replace=True)
-        RR=pd.Series(r1)
-        """
-        
-    #基于新的日收益率序列，使用标准正态法计算ES
-    ratio=calc_ES_normal_standard(RR,future_days,alpha)        
-    
-    return ratio
-
-#==============================================================================
-if __name__ == '__main__':
-    portfolio={'Market':('China','000001.SS'),'000661.SZ':2,'603392.SS':3,'300601.SZ':4}
-    today='2020-7-20'
-    future_days=1
-    alpha=0.99
-    pastyears=1
-    model='all'
-    printout=True
-
-
-def get_VaR_portfolio(portfolio,today,future_days=1,alpha=0.99, \
-                      pastyears=1,model='normal_standard',printout=True):
-    """
-    功能：基于指定模型model，计算投资组合portfolio的VaR金额和比率
-    """
-    print("  Start working on portfolio VaR, please wait ...")
-
-    #检查model类型
-    modellist=['normal_standard','normal_modified','historical','montecarlo', \
-               'mc_oversampling','all']
-    modeltyp=model.lower()
-    if modeltyp not in modellist:
-        print("  #Error(get_VaR_portfolio): Unsupported type of model,",model)
-        print("  Supported models:",modellist)
-        return None,None
-    
-    #获得样本起始日期
-    start=get_start_date(today,pastyears)
-    if start is None: return None,None
-    
-    #抓取投资组合股价和收益率
-    print("  Searching for portfolio prices from",start,'to',today)
-    prices=get_portfolio_prices0(portfolio,start,today)
-    #判断空值，控制空值可能引起的程序崩溃
-    if prices is None:
-        print("  #Error(get_VaR_portfolio): failed to retrieve portfolio prices")
-        return None,None
-
-    prices.dropna(inplace=True)
-    if len(prices)==0:
-        print("  #Error(get_VaR_portfolio): zero records found for the portfolio")
-        return None,None  
-    
-    num=len(prices)
-    ret_series=prices['Ret']
-    
-    print("  Calculating portfolio VaR ...")
-    #计算当日头寸
-    position=round(prices['Close'][-1],2)
-    VaRlist=[]; ratiolist=[]
-    #标准正态法
-    if modeltyp in ['normal_standard','all']:
-        ratio=calc_VaR_normal_standard(ret_series,future_days,alpha)
-        VaR=-round(position*ratio,2)
-        VaRlist=VaRlist+[VaR]
-        VaR_ratio=round(ratio,4)
-        ratiolist=ratiolist+[VaR_ratio]
-    
-    #修正正态法
-    if modeltyp in ['normal_modified','all']:
-        ratio=calc_VaR_normal_modified(ret_series,future_days,alpha)
-        VaR=-round(position*ratio,2)
-        VaRlist=VaRlist+[VaR]
-        VaR_ratio=round(ratio,4)
-        ratiolist=ratiolist+[VaR_ratio]
-    
-    #历史模拟法
-    if modeltyp in ['historical','all']:
-        ratio=calc_VaR_historical(ret_series,future_days,alpha)
-        VaR=-round(position*ratio,2)
-        VaRlist=VaRlist+[VaR]
-        VaR_ratio=round(ratio,4)
-        ratiolist=ratiolist+[VaR_ratio]    
-    
-    #蒙特卡洛模拟法，随机数，默认
-    if modeltyp in ['montecarlo','all']:
-        ratio=calc_VaR_montecarlo(ret_series,future_days,alpha,mctype='random')
-        VaR=-round(position*ratio,2)
-        VaRlist=VaRlist+[VaR]
-        VaR_ratio=round(ratio,4)
-        ratiolist=ratiolist+[VaR_ratio]
-    
-    #蒙特卡洛模拟法，超采样
-    if modeltyp in ['mc_oversampling','all']:
-        ratio=calc_VaR_montecarlo(ret_series,future_days,alpha, \
-                                  mctype='mc_oversampling')
-        VaR=-round(position*ratio,2)
-        VaRlist=VaRlist+[VaR]
-        VaR_ratio=round(ratio,4)
-        ratiolist=ratiolist+[VaR_ratio]    
-    
-    if not printout: return VaRlist,ratiolist
-    
-    #输出VaR金额和比率
-    print("\n===== 在险价值VaR：投资组合 =====")
-    #_,_,tickerlist,sharelist,ticker_type=decompose_portfolio(portfolio)
-    #print("成分股列表  :",tickerlist)
-    #print("成分股的配置:",sharelist)
-    print("投资组合：",portfolio_name(portfolio))
-    print("当前日期：",today)
-    print("当前头寸(1份投资组合)：",format(position,','))
-    print("预计持有天数：",future_days)
-    print("置信度水平：",str(int(alpha*100))+'%')
-    print("历史样本年数：",pastyears)
-    print("样本数量：",num)
-    
-    import datetime as dt; today=dt.date.today()
-    footnote="\n数据来源：新浪/stooq，"+str(today)
-    if not (modeltyp == 'all'):
-        print("使用的模型：",model)
-        print("VaR金额/比率：",format(VaR,','), \
-              '\b，',str(round(VaR_ratio*100,2))+'%')
-        
-        print(footnote)
-        return VaR,VaR_ratio
-    
-    print("\n*** VaR金额/比率 ***")
-    modellist.pop()
-    for m in modellist:
-        pos=modellist.index(m)
-        v=VaRlist[pos]
-        r=ratiolist[pos]
-        print("  ",format(v,','),'\b,', \
-              str(round(r*100,2))+'%'+' ('+m+')')
-
-    print(footnote)
-
-    return VaRlist,ratiolist
-
-if __name__ == '__main__':
-    portfolio={'Market':('China','000001.SS'),'000661.SZ':1,'603392.SS':2, \
-               '300601.SZ':3}
-    today='2020-7-20'
-    future_days=1
-    alpha=0.99
-    pastyears=1
-    model='all'
-    printout=True
-
-#==============================================================================
-def get_ES_portfolio(portfolio,today,future_days=1,alpha=0.99, \
-                      pastyears=1,model='normal_standard',printout=True):
-    """
-    功能：基于指定模型model，计算投资组合portfolio的ES金额和比率
-    """
-
-    #检查model类型
-    modellist=['normal_standard','normal_modified','historical','montecarlo', \
-               'mc_oversampling','all']
-    modeltyp=model.lower()
-    if modeltyp not in modellist:
-        print("#Error(get_ES_portfolio): Unsupported type of model,",model)
-        print("Supported models:",modellist)
-        return None,None
-    
-    #获得样本起始日期
-    start=get_start_date(today,pastyears)
-    if start is None: return None,None
-    #抓取投资组合股价和收益率
-    prices=get_portfolio_prices0(portfolio,start,today)
-    #判断空值，控制空值可能引起的程序崩溃
-    if prices is None:
-        print("#Error(get_ES_portfolio): failed to retrieve portfolio prices")
-        return None,None
-    prices.dropna(inplace=True)
-    if (prices is None) or (len(prices)==0):
-        print("#Error(get_ES_portfolio): failed to retrieve portfolio prices")
-        return None,None        
-    
-    num=len(prices)
-    ret_series=prices['Ret']
-    
-    #计算当日头寸
-    position=round(prices['Close'][-1],2)
-    ESlist=[]; ratiolist=[]
-    #标准正态法
-    if modeltyp in ['normal_standard','all']:
-        ratio=calc_ES_normal_standard(ret_series,future_days,alpha)
-        ES=-round(position*ratio,2)
-        ESlist=ESlist+[ES]
-        ES_ratio=round(ratio,4)
-        ratiolist=ratiolist+[ES_ratio]
-    
-    #修正正态法
-    if modeltyp in ['normal_modified','all']:
-        ratio=calc_ES_normal_modified(ret_series,future_days,alpha)
-        ES=-round(position*ratio,2)
-        ESlist=ESlist+[ES]
-        ES_ratio=round(ratio,4)
-        ratiolist=ratiolist+[ES_ratio]
-    
-    #历史模拟法
-    if modeltyp in ['historical','all']:
-        ratio=calc_ES_historical(ret_series,future_days,alpha)
-        ES=-round(position*ratio,2)
-        ESlist=ESlist+[ES]
-        ES_ratio=round(ratio,4)
-        ratiolist=ratiolist+[ES_ratio]    
-    
-    #蒙特卡洛模拟法，随机数，默认
-    if modeltyp in ['montecarlo','all']:
-        ratio=calc_ES_montecarlo(ret_series,future_days,alpha,mctype='random')
-        ES=-round(position*ratio,2)
-        ESlist=ESlist+[ES]
-        ES_ratio=round(ratio,4)
-        ratiolist=ratiolist+[ES_ratio]
-    
-    #蒙特卡洛模拟法，超采样
-    if modeltyp in ['mc_oversampling','all']:
-        ratio=calc_ES_montecarlo(ret_series,future_days,alpha, \
-                                  mctype='mc_oversampling')
-        ES=-round(position*ratio,2)
-        ESlist=ESlist+[ES]
-        ES_ratio=round(ratio,4)
-        ratiolist=ratiolist+[ES_ratio]    
-    
-    if not printout: return ESlist,ratiolist
-    
-    #输出ES金额和比率
-    print("\n===== 投资组合的预期不足ES =====")
-    #_,_,tickerlist,sharelist,ticker_type=decompose_portfolio(portfolio)
-    #print("成分股列表  :",tickerlist)
-    #print("成分股配置  :",sharelist)
-    print("投资组合：",portfolio_name(portfolio))
-    print("当前日期：",today)
-    print("当前头寸(1份投资组合)：",format(position,','))
-    print("预计持有天数：",future_days)
-    print("置信度水平：",str(int(alpha*100))+'%')
-    print("历史样本年数：",pastyears)
-    print("样本个数：",num)
-
-    import datetime as dt; today=dt.date.today()
-    footnote="\n数据来源：新浪/stooq，"+str(today)
-    
-    if not (modeltyp == 'all'):
-        print("使用的模型：",model)
-        print("ES金额/比率：",format(ES,','), \
-              '\b,',str(round(ES_ratio*100,2))+'%')
-        print(footnote)
-        return ES,ES_ratio
-    
-    print("ES金额/比率：")
-    modellist.pop()
-    for m in modellist:
-        pos=modellist.index(m)
-        v=ESlist[pos]
-        r=ratiolist[pos]
-        print("  ",format(v,','),'\b,', \
-              str(round(r*100,2))+'%'+' ('+m+')')
-
-    print(footnote)
-
-    return ESlist,ratiolist
-
-if __name__ == '__main__':
-    portfolio={'Market':('China','000001.SS'),'000661.SZ':1,'603392.SS':2, \
-               '300601.SZ':3}
-    today='2020-7-20'
-    future_days=1
-    alpha=0.99
-    pastyears=1
-    model='all'
-    printout=True
-
-#===========================================================================
-def ret_Normality_SW(ret_series,siglevel=0.05):
-    """
-    功能：测试一个投资组合portfolio在给定期间内（fromdate,todate）的收益率序列
-    是否符合正态分布
-    输入参数：投资组合，开始日期，结束日期，显著性要求水平
-    输出：收益率序列正态性检验的W, p-value, Skewness, Kurtosis
-    【Shapiro-Wilk正态性检验】原假设：服从正态分布
-    """
-    #去掉空值
-    ret_series=ret_series[~ret_series.isnull()]
-
-    from scipy import stats
-    (W,p_value)=stats.shapiro(ret_series)
-    
-    S=stats.skew(ret_series)
-    K=stats.kurtosis(ret_series)
-    
-    return round(W,4),round(p_value,4),round(S,2),round(K,2)
-
-#===========================================================================
-def portfolio_ret_Normality_SW(portfolio,fromdate,todate, \
-                               siglevel=0.05,printout=True):
-    """
-    功能：测试一个投资组合portfolio在给定期间内（fromdate,todate）的收益率序列
-    是否符合正态分布
-    输入参数：投资组合，开始日期，结束日期，显著性要求水平
-    输出：收益率序列正态性检验的p-value
-    【Shapiro-Wilk正态性检验】原假设：服从正态分布
-    """
-    #抓取投资组合股价和收益率
-    prices=get_portfolio_prices0(portfolio,fromdate,todate)
-    if prices is None: return None,None
-    ret_series=prices['Ret']
-    num=len(prices)-1
-    
-    #检验正态分布
-    W,p_value,S,K=ret_Normality_SW(ret_series,siglevel)    
-    
-    if not printout: return W,p_value,S,K
-    
-    _,_,tickerlist,sharelist,ticker_type=decompose_portfolio(portfolio)
-    print("*** Shapiro-Wilk Normality Test ***")
-    print("    Stock(s):",tickerlist)
-    print("    Holding proportion:",sharelist)    
-    print("    Sampling period   :",fromdate,'to',todate)
-    print("    Observations      :",num)
-    print("    Null hypothesis   : Normal")
-    print("    W statistic       :",round(W,4))
-    print("    p-value           :",round(p_value,4))
-    print("    Skewness          :",round(S,2))
-    print("    Kurtosis          :",round(K,2))
-    if p_value >= siglevel:
-        print("Result: Accept null hypothesis, normal")
-    else:
-        print("Result: Reject null hypothesis, not normal")
-    
-    footnote="... Data source: Yahoo Finance"
-    print(footnote)
-    
-    return W,p_value,S,K
-
-#==============================================================================
-if __name__ == '__main__':
-    portfolio={'Market':('China','000001.SS'),'000661.SZ':0.1,'603392.SS':0.2, \
-               '300601.SZ':0.7}
-    fromdate='2022-1-1'
-    todate='2022-4-18'
-
-def portfolio_rets_curve(portfolio,start='MRY',end='today'):
-    """
-    功能：绘制投资组合portfolio在给定期间（fromdate,todate）收益率分布的曲线，
-    并于相应的正态分布图对照
-    显示：收益率分布的直方图(实线)，相应的正态分布图(虚线)
-    x轴为收益率(非百分比)，y轴为频度(Frequency)
-    """
-    fromdate,todate=start_end_preprocess(start,end)
-    
-    #抓取投资组合股价和收益率
-    prices=get_portfolio_prices0(portfolio,fromdate,todate)
-    if prices is None: return None,None
-    
-    num=len(prices)
-    rets=prices['Ret']
-    W,p_value,S,K=ret_Normality_SW(rets)
-    _,_,tickerlist,sharelist,ticker_type=decompose_portfolio(portfolio)
-    #计算收益率的均值和标准差
-    mu=rets.mean()
-    sigma=rets.std()
-
-    #生成符合正态分布的随机数，符合股票收益率的均值和标准差
-    import numpy as np
-    x=mu+sigma*np.random.randn(10000)    
-            
-    #plt.figure(figsize=(8,4))
-    plt.figure(figsize=(12.8,6.4))
-    import seaborn as sns
-    #绘制曲线：股票收益率
-    sns.kdeplot(data=rets,shade=True,color='blue',legend=True,label='投资组合',lw=4) 
-    #绘制曲线：对应的正态分布
-    sns.kdeplot(data=x,shade=True,color='r',legend=True,label='正态分布线',ls='--')
-    #设置标题、图例、坐标轴标签
-    #plt.ylabel('Frequency')
-    
-    footnote1='日收益率%'
-    footnote2='\n【注】Shapiro-Wilk检验：W值'+str(W)+'，p值'+str(p_value)
-    footnote3="。偏度"+str(S)+"，峰度"+str(K)
-    
-    import datetime as dt; today=dt.date.today()
-    footnote4="。数据来源：新浪/stooq，"+str(today)
-    
-    footnote=footnote1+footnote2+footnote3+footnote4
-    plt.xlabel(footnote)
-    
-    plt.legend(loc='best')
-    """
-    titletxt1="投资组合收益率的正态分布检验"
-    titletxt2="\n成分股"+str(tickerlist)
-    titletxt3="\n持仓配置"+str(sharelist)
-    titletxt4="\n样本期间："+str(fromdate)+'至'+str(todate)+ \
-                '，样本数量'+str(num)+'个'
-    titletxt=titletxt1+titletxt2+titletxt3+titletxt4
-    """
-    titletxt="投资组合收益率的正态分布检验："+portfolio_name(portfolio)
-    plt.title(titletxt)
-    
-    plt.gca().set_facecolor('whitesmoke')
-    plt.show()
-
-    return
-    
-
-#==============================================================================