siat 2.10.13__py3-none-any.whl → 2.11.2__py3-none-any.whl

siat/security_trend.py

@@ -1,329 +1,426 @@
-# -*- coding: utf-8 -*-
+# -*- coding: utf-8 -*-
+
 """
-本模块功能：投资组合的风险调整收益率教学插件
-所属工具包：证券投资分析工具SIAT 
-SIAT：Security Investment Analysis Tool
-创建日期：2023年7月15日
-最新修订日期：2023年7月16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！
-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！
+版权：王德宏，北京外国语大学国际商学院
+功能：计算CAPM模型贝塔系数的调整值
+版本：2.1，2019-7-25
 """
 
 #==============================================================================
 #关闭所有警告
 import warnings; warnings.filterwarnings('ignore')
-#==============================================================================
 from siat.common import *
 from siat.translate import *
-from siat.stock import *
-from siat.security_prices import *
-from siat.risk_adjusted_return import *
 from siat.grafix import *
+from siat.security_prices import *
+#==============================================================================
+import matplotlib.pyplot as plt
 
-import pandas as pd
+#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'}
 
-import datetime as dt; today=str(dt.date.today())
-#==============================================================================
+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 
 #==============================================================================
+def prepare_capm(stkcd,mktidx,start,end):
+    """
+    函数功能：准备计算一只股票CAPM模型贝塔系数的数据，并标记年度
+    输入参数：
+    stkcd: 股票代码
+    mktidx: 指数代码
+    start：使用股票价格数据的开始日期，MM/DD/YYYY
+    end：使用股票价格数据的结束日期，MM/DD/YYYY
+    输出数据：
+    返回数据：带年度标记的可直接用于capm回归的股票收益率数据
+    """
+        
+    #仅用于调试，正式使用前应注释掉
+    #stkcd='002504.SZ'; mktidx='000001.SS'
+    #start="12/31/2011"; end="12/31/2018"
+
+    #抓取股价和指数
+    stock=get_price(stkcd,start,end)
+    if stock is None:
+        print("  #Error(prepare_capm): no data retrieved from server!")
+        return None
+    market=get_price(mktidx,start,end)
+    if market is None:
+        print("  #Error(prepare_capm): no index data retrieved from server!")
+        return None    
+
+    #计算日收益率
+    import pandas as pd
+    stkret=pd.DataFrame(stock['Close'].pct_change())
+    mktret=pd.DataFrame(market['Close'].pct_change())
+
+    #合并，去掉空缺
+    R=pd.merge(mktret,stkret,how='left',left_index=True,right_index=True)
+    R=R.dropna()
+
+    #标记各个年度
+    R['Year']=R.index.strftime("%Y")
+
+    #返回带年份的股票收益率序列
+    return R
+
 if __name__=='__main__':
-    #测试组1
-    ticker='JD'
-    indicator='Exp Ret%'
-    start='2022-1-1'
-    end='2022-12-31'
-    datatag=False
-    power=1
-    graph=True
-    source='auto'
-    
-    df=security_trend(ticker,indicator=indicator,power=1)
-    
-    #测试组2
-    ticker='AAPL'
-    indicator=['Close','Open']
-    start='default'
-    end='default'
-    datatag=False
-    power=0
-    graph=True
-    twinx=True
-    loc1='upper left'
-    loc2='lower right'
-    source='auto'
-    
-    #测试组3
-    ticker='AAPL'
-    indicator=['Close','Open','High','Low']
-    start='default'
-    end='default'
-    datatag=False
-    power=0
-    graph=True
-    twinx=True
-    loc1='upper left'
-    loc2='lower right'
-    source='auto'
+    R1=prepare_capm('0700.HK','^HSI','2014-01-01','2018-12-31')
+
+#==============================================================================
+#==============================================================================
+def get_beta_ML(stkcd,mktidx,yearlist,printout=True,graph=True):
+    """
+    函数功能：使用ML方法调整一只股票的CAPM模型贝塔系数
+    输入参数：
+    stkcd: 股票代码
+    mktidx: 指数代码
+    yearlist：年度列表，列出其中期间的贝塔系数
+    输出数据：
+    显示CAPM市场模型回归的beta, 以及ML调整后的beta系数
+    返回数据:年度CAPM贝塔系数和ML调整后的beta系数
+    """
+
+    #仅为测试用，完成后应立即注释掉
+    #stkcd='0700.HK'
+    #mktidx='^HSI'
+    #yearlist=['2015','2016','2017','2018']
     
-    df=security_trend(ticker,indicator=indicator)
+    Y4=str(int(yearlist[0])-1)
+    start=Y4+'-01-01'
+    end=yearlist[-1]+'-12-31'    
+        
+    #读取股价并准备好收益率数据
+    try:
+        R=prepare_capm(stkcd,mktidx,start,end)
+    except:
+        print("  #Error(get_beta_ML): Preparing CAPM data failed!")
+        return None
+
+    if (R is None):
+        print("  #Error(get_beta_ML): server time out")
+        return None
+    if (len(R) == 0):
+        print("  #Error(get_beta_ML): server returned empty data")
+        return None
     
-    #测试组4
-    ticker=["GCZ25.CMX","GCZ24.CMX"]
-    indicator='Close'
-    start="2020-1-1"
-    end="2020-6-30"
+    #用于保存beta(CAPM)和beta(ML)
+    import pandas as pd
+    betas=pd.DataFrame(columns=('Year','Beta(CAPM)','Beta(ML)'))
+
+    #计算Merrill-Lynch方法贝塔系数调整
+    from scipy import stats
+    for year in yearlist:
+        r=R[R['Year']==year]
+        if len(r) != 0:
+            output=stats.linregress(r['Close_x'],r['Close_y'])
+            (beta,alpha,r_value,p_value,std_err)=output
+            beta_ML=beta*2.0/3.0+1.0/3.0
+            #整齐输出 
+            #print(year,"%6.4f "%(beta),"%6.4f "%(beta_ML))
+
+            row=pd.Series({'Year':year,'Beta(CAPM)':beta,'Beta(ML)':beta_ML})
+            try:
+                betas=betas.append(row,ignore_index=True)
+            except:
+                betas=betas._append(row,ignore_index=True)
+
+    betas.set_index(["Year"], inplace=True)
     
+    if printout == True: printdf_betas(betas,2)
+    if graph == True:
+        model="贝塔系数的简单调整法"
+        draw2_betas(model,mktidx,stkcd,betas)
+
+    return betas
+
+#==============================================================================
+def printdf_betas(df,decimal=2):
+    """
+    功能：整齐地显示数据框的内容，自动调整各列宽度
+    """
+    #打印时保留的小数点位数
+    dec="%."+str(decimal)+"f"
+    format=lambda x: dec % x
+    df1=df.applymap(format)    
     
+    import pandas as pd
+    #调整最佳列宽
+    old_width = pd.get_option('display.max_colwidth')
+    pd.set_option('display.max_colwidth', -1)
+    print(df1)
+    pd.set_option('display.max_colwidth', old_width)
+
+    return
     
-def security_trend(ticker,indicator='Close', \
-                   start='default',end='default', \
-                   kline=False,kline_demo=False,mav=[5,10,20], \
-                   stock_dividend=False,stock_split=False, \
-                   market="China",market_index="000300.SS",RF=False,window=252, \
-                   sortby='tpw_mean',trailing=10,trend_threshhold=0.001, \
-                   graph=True,twinx=False,loc1='best',loc2='best', \
-                   datatag=False,power=0, \
-                   smooth=True,date_range=False,date_freq=False,annotate=False, \
-                   preprocess='none',scaling_option='start', \
-                   printout=False, \
-                   source='auto'):
+if __name__=='__main__':
+    yearlist=gen_yearlist['2010','2019']
+    betas=get_beta_ML('AAPL','^GSPC',yearlist)    
+    betas2=get_beta_ML('BILI','^GSPC',yearlist)
+    betas3=get_beta_ML('0700.HK','^HSI',yearlist)
+    yearlist1=['2015','2016','2017','2018']
+    betas3=get_beta_ML('0700.HK','^HSI',yearlist1)
+
+#==============================================================================
+def draw2_betas(model,scope,ticker,betas):    
+    """
+    功能：绘制双曲线的贝塔因子变化图
+    输入参数：
+    model: 模型类型, 任意字符串(例如Merrill-Lynch Beta Adjustment)
+    scope: 市场指数, 任意字符串(例如Standard & Poor 500)
+    ticker：股票代码
+    输出：图形
+    """
+    #仅用作测试，完成后应注释掉
+    #model="Merrill-Lynch Beta Adjustment"
+    #scope="Standard & Poor 500"
+    #ticker="AAPL"
 
+    #取得股票和指数名字，对于非美股可能耗时较长
     """
-    国内：描述证券指标走势
+    import yfinance as yf
+    mktidx= yf.Ticker(scope)
+    idxinfo=mktidx.info
+    idxname=idxinfo['shortName']
+    stkcd=yf.Ticker(ticker)
+    stkinfo=stkcd.info
+    stkname=stkinfo['shortName']   
+    title1="\n"+stkname+"\n"+model+"\n(Benchmark on "+idxname+")"
+    """
+    title1=codetranslate(ticker)+": "+model+"\n(基于"+codetranslate(scope)+")"
+   
+    #转换索引类型为DatetimeIndex，便于后续处理
+    """
+    import pandas as pd
+    betas['Date']=betas.index
+    betas['Date']=pd.to_datetime(betas['Date'])
+    betas.set_index('Date',inplace=True)
+    """
+
+    #获得列明
+    betalist=betas.columns.values.tolist()
+    beta1=betalist[0]
+    beta2=betalist[1]
+
+    try:
+        plt.plot(betas[beta1],label=beta1,marker='o',color='red')
+        plt.plot(betas[beta2],label=beta2,marker='*',linewidth=2,ls='-.',color='blue')
+    except:
+        print("  #Error(draw2_betas): no available data for drawing!")
+        return
+    plt.axhline(y=1.0,color='b',linestyle=':',label='市场线')  
+    plt.title(title1,fontsize=12,fontweight='bold')
+    plt.ylabel("贝塔系数",fontsize=12,fontweight='bold')
     
-    """    
+    plt.gcf().autofmt_xdate() # 优化标注（自动倾斜）
+    #plt.xticks(rotation=30)
+    plt.legend(loc='best')    
     
-    # 检查证券代码
-    if isinstance(ticker,str):
-        ticker_num=1
-        tickers=[ticker]
-    elif isinstance(ticker,list):
-        ticker_num=len(ticker)
-        tickers=ticker
-    else:
-        print("  #Error(security_trend): unrecognizable security codes",ticker)
-        return None
+    import datetime; today = datetime.date.today()
+    plt.xlabel("数据来源：新浪，"+str(today))    
     
-    # 检查日期：截至日期
-    import datetime as dt; today=dt.date.today()
-    end=end.lower()
-    if end in ['default','today']:
-        todate=today
-    else:
-        validdate,todate=check_date2(end)
-        if not validdate:
-            print("  #Warning(security_trend): invalid date for",end)
-            todate=today
-
-    # 检查日期：开始日期
-    start=start.lower()
-    if start in ['default','mrm']:  # 默认近一个月
-        fromdate=date_adjust(todate,adjust=-31)
-    elif start in ['mrq']:  # 近三个月
-        fromdate=date_adjust(todate,adjust=-63)   
-    elif start in ['mry']:  # 近一年
-        fromdate=date_adjust(todate,adjust=-366)  
-    elif start in ['lty']:  # 近三年以来
-        fromdate=date_adjust(todate,adjust=-366*3)  
-    elif start in ['lfy']:  # 近五年以来
-        fromdate=date_adjust(todate,adjust=-366*5)          
-    elif start in ['ytd']:  # 今年以来
-        fromdate=str(today.year)+'-1-1'        
-    else:
-        validdate,fromdate=check_date2(start)
-        if not validdate:
-            print("  #Warning(security_trend): invalid date for",start,"/b, set to MRM")
-            fromdate=date_adjust(todate,adjust=-31)    
-
-    # 处理K线图
-    if kline and not kline_demo:
-        # 跟踪
-        #print(tickers[0],fromdate,todate)
-        if start in ['default']:
-            fromdate=date_adjust(todate,adjust=-60)
-        if not isinstance(mav,list):
-            mav=[mav]
-        df=candlestick(stkcd=tickers[0],fromdate=fromdate,todate=todate,mav=mav)
-        return df
-
-    if kline and kline_demo:
-        if start in ['default']:
-            fromdate=date_adjust(todate,adjust=-7)
-        
-        df=candlestick_demo(tickers[0],fromdate=fromdate,todate=todate)
-        return df
+    plt.show()       
+    
+    return
 
-    # 处理股票分红和股票分拆：需要访问雅虎财经
-    if stock_dividend:
-        if start in ['default']:
-            fromdate=date_adjust(todate,adjust=-365*5)  
-            
-        df=stock_dividend(ticker=tickers[0],fromdate=fromdate,todate=todate)
-        return df
+if __name__=='__main__':
+    model="ML Beta Adjustment"
+    scope="SP500"
+    ticker="AAPL"
+    draw2_betas(model,scope,ticker,betas)
 
-    if stock_split:
-        if start in ['default']:
-            fromdate=date_adjust(todate,adjust=-365*5)  
-        
-        df=stock_split(ticker=tickers[0],fromdate=fromdate,todate=todate)
-        return df
+
+#==============================================================================
+def get_beta_SW(stkcd,mktidx,yearlist,printout=True,graph=True):
+    """
+    函数功能：使用SW方法调整一只股票的CAPM模型贝塔系数
+    输入参数：
+    stkcd: 股票代码
+    mktidx: 指数代码
+    yearlist：年度列表，列出其中期间的贝塔系数
+    输出数据：显示CAPM市场模型回归的beta, 以及调整后的beta系数
+    返回数据：CAPM市场模型回归的beta, 以及调整后的beta系数
+    """
+
+    #仅为测试用，完成后应立即注释掉
+    #stkcd='0700.HK'
+    #mktidx='^HSI'
+    #yearlist=['2015','2016','2017','2018']
     
+    #生成开始结束日期
+    Y4=str(int(yearlist[0])-1)
+    start=Y4+'-01-01'
+    end=yearlist[-1]+'-12-31'   
+        
+    #读取股价并准备好收益率数据
+    try:
+        R=prepare_capm(stkcd,mktidx,start,end)
+    except:
+        print("  #Error(get_beta_SW): preparing CAPM data failed!")
+        return None
 
-    # 检查指标：是否字符串或列表
-    if isinstance(indicator,str):
-        measures=[indicator]
-        indicator_num=1
-    elif isinstance(indicator,list):
-        measures=indicator
-        indicator_num=len(indicator)
-    else:
-        print("  #Error(security_trend): invalid indicator for",indicator)
+    if (R is None):
+        print("  #Error(get_beta_SW): server time out")
         return None
-            
-    # 检查指标
-    indicator_list1=['Open','Close','Adj Close','High','Low',
-             'Daily Ret','Daily Ret%','Daily Adj Ret','Daily Adj Ret%',
-             'log(Daily Ret)','log(Daily Adj Ret)','Weekly Ret','Weekly Ret%',
-             'Weekly Adj Ret','Weekly Adj Ret%','Monthly Ret','Monthly Ret%',
-             'Monthly Adj Ret','Monthly Adj Ret%','Quarterly Ret','Quarterly Ret%',
-             'Quarterly Adj Ret','Quarterly Adj Ret%','Annual Ret','Annual Ret%',
-             'Annual Adj Ret','Annual Adj Ret%','Exp Ret','Exp Ret%','Exp Adj Ret',
-             'Exp Adj Ret%','Weekly Price Volatility','Weekly Adj Price Volatility',
-             'Monthly Price Volatility','Monthly Adj Price Volatility',
-             'Quarterly Price Volatility','Quarterly Adj Price Volatility',
-             'Annual Price Volatility','Annual Adj Price Volatility',
-             'Exp Price Volatility','Exp Adj Price Volatility',
-             'Weekly Ret Volatility','Weekly Ret Volatility%',
-             'Weekly Adj Ret Volatility','Weekly Adj Ret Volatility%',
-             'Monthly Ret Volatility', 'Monthly Ret Volatility%',
-             'Monthly Adj Ret Volatility', 'Monthly Adj Ret Volatility%',
-             'Quarterly Ret Volatility', 'Quarterly Ret Volatility%',
-             'Quarterly Adj Ret Volatility', 'Quarterly Adj Ret Volatility%',
-             'Annual Ret Volatility', 'Annual Ret Volatility%',
-             'Annual Adj Ret Volatility', 'Annual Adj Ret Volatility%',
-             'Exp Ret Volatility', 'Exp Ret Volatility%', 'Exp Adj Ret Volatility',
-             'Exp Adj Ret Volatility%', 'Weekly Ret LPSD', 'Weekly Ret LPSD%',
-             'Weekly Adj Ret LPSD', 'Weekly Adj Ret LPSD%', 'Monthly Ret LPSD',
-             'Monthly Ret LPSD%', 'Monthly Adj Ret LPSD', 'Monthly Adj Ret LPSD%',
-             'Quarterly Ret LPSD', 'Quarterly Ret LPSD%', 'Quarterly Adj Ret LPSD',
-             'Quarterly Adj Ret LPSD%', 'Annual Ret LPSD', 'Annual Ret LPSD%',
-             'Annual Adj Ret LPSD', 'Annual Adj Ret LPSD%', 'Exp Ret LPSD',
-             'Exp Ret LPSD%', 'Exp Adj Ret LPSD', 'Exp Adj Ret LPSD%',
-             ]
-
-    indicator_list2=['treynor','sharpe','sortino','alpha']
-    
-    # 是否属于支持的指标
-    for m in measures:
-        if not (m in indicator_list1 + indicator_list2):
-            print("  #Error(security_trend): unsupported indicator for",m)
-            print("  Supported indicators:")
-            printlist(indicator_list1,numperline=4,beforehand='  ',separator='   ')
-            printlist(indicator_list2,numperline=5,beforehand='  ',separator='   ')
-            return None
-    # 不能同时支持indicator_list1和indicator_list2的指标
-    indicator_group1=False
-    indicator_group2=False
-    
-    for m in measures:
-        if m in indicator_list2:
-            indicator_group2=True
-        else:
-            indicator_group1=True
-    if indicator_group1 and indicator_group2:
-        print("  #Error(security_trend): cannot support these indicators together",measure)
+    if (len(R) == 0):
+        print("  #Error(get_beta_SW): server returned empty data")
         return None
 
+    #用于保存beta(CAPM)和beta(SW)
+    import pandas as pd
+    betas=pd.DataFrame(columns=('Year','Beta(CAPM)','Beta(SW)'))
+
+    #计算Scholes-William调整
+    R['Close_x+1']=R['Close_x'].shift(1)    
+    R['Close_x-1']=R['Close_x'].shift(-1)
+    R=R.dropna()    #stats.linregress不接受空缺值
+
+    from scipy import stats    
+    for year in yearlist:
+        r=R[R['Year']==year]
+        if len(r) != 0:
+            output=stats.linregress(r['Close_x'],r['Close_y'])
+            (beta0,alpha,r_value,p_value,std_err)=output
+
+            output=stats.linregress(r['Close_x+1'],r['Close_y'])
+            (beta1,alpha,r_value,p_value,std_err)=output 
+
+            output=stats.linregress(r['Close_x-1'],r['Close_y'])
+            (beta_1,alpha,r_value,p_value,std_err)=output    
+
+            output=stats.linregress(r['Close_x-1'],r['Close_x'])
+            (rou,alpha,r_value,p_value,std_err)=output    
+
+            beta_SW=(beta_1+beta0+beta1)/(1.0+2.0*rou)
+            row=pd.Series({'Year':year,'Beta(CAPM)':beta0,'Beta(SW)':beta_SW})
+            try:
+                betas=betas.append(row,ignore_index=True)
+            except:
+                betas=betas._append(row,ignore_index=True)
     
-    # 情形1：单个证券，单个普通指标
-    if ticker_num==1 and indicator_num==1 and indicator_group1:
-        df=security_indicator(ticker=tickers[0],indicator=measures[0], \
-                              fromdate=fromdate,todate=todate, \
-                              datatag=datatag,power=power,graph=graph, \
-                              source=source)
-        return df
+    betas.set_index(["Year"], inplace=True)
     
-    # 情形2：单个证券，两个普通指标，twinx==True
-    if ticker_num==1 and indicator_num == 2 and indicator_group1 and twinx:
-        df=compare_security(tickers=tickers[0],measures=measures[:2], \
-                            fromdate=fromdate,todate=todate,twinx=twinx, \
-                            loc1=loc1,loc2=loc2,graph=graph,source=source)
-        return df
+    if printout == True: printdf_betas(betas,2)
+    if graph == True:
+        model="贝塔系数的Scholes-Williams调整法"
+        draw2_betas(model,mktidx,stkcd,betas)
     
-    # 情形3：单个证券，两个及以上普通指标
-    if ticker_num==1 and indicator_num >= 2 and indicator_group1 and not twinx:
-        df=security_mindicators(ticker=tickers[0],measures=measures, \
-                         fromdate=fromdate,todate=todate, \
-                         graph=graph,smooth=smooth,loc=loc1, \
-                         date_range=date_range,date_freq=date_freq, \
-                         annotate=annotate, \
-                         source=source)
-        return df
+    return betas
+
     
-    # 情形4：两个证券，取第一个普通指标，twinx==True
-    if ticker_num==2 and indicator_group1 and twinx:
-        df=compare_security(tickers=tickers,measures=measures[0], \
-                            fromdate=fromdate,todate=todate,twinx=twinx, \
-                            loc1=loc1,loc2=loc2,graph=graph,source=source)
-        return df
-
-    # 情形5：两个及以上证券，取第一个普通指标
-    if ticker_num==2:
-        linewidth=2.5
-    elif ticker_num==3:
-        linewidth=2.0
-    else:
-        linewidth=1.5
-        
-    if ((ticker_num == 2 and not twinx) or ticker_num > 2) and indicator_group1:
-        df=compare_msecurity(tickers=tickers,measure=measures[0], \
-                             start=fromdate,end=todate, \
-                      axhline_value=0,axhline_label='', \
-                      preprocess=preprocess,linewidth=linewidth, \
-                      scaling_option=scaling_option, \
-                      graph=graph,loc=loc1, \
-                      annotate=annotate,smooth=smooth, \
-                      source=source)
-        return df
-
-    # 情形6：单个证券，单个或多个RAR指标
-    if indicator_group2 and ticker_num==1 and indicator_num >= 1:
-        df=compare_1security_mrar(ticker=tickers[0],rar_names=measures, \
-                                  start=fromdate,end=todate, \
-                        market=market,market_index=market_index,RF=RF,window=window, \
-                        axhline_value=0,axhline_label='零线',graph=graph,printout=printout, \
-                        sortby=sortby,source=source,trailing=trailing,trend_threshhold=trend_threshhold, \
-                        annotate=annotate)
-        return df
-
-    # 情形7：多个证券，取第一个RAR指标
-    if indicator_group2 and ticker_num > 1:
-        df=compare_mrar(tickers=tickers,rar_name=measures[0], \
-                        start=fromdate,end=todate, \
-                        market=market,market_index=market_index,RF=RF,window=window, \
-                        axhline_value=0,axhline_label='零线',graph=graph,printout=printout, \
-                        sortby=sortby,source=source,trailing=trailing,trend_threshhold=trend_threshhold, \
-                        annotate=annotate)
-        return df
+if __name__=='__main__':
+    yearlist=gen_yearlist('2010','2019')
+    betas_AAPL=get_beta_SW('AAPL','^GSPC',yearlist)
     
-    # 其他未预料情形
-    print("  Sorry, unexpected security and indicator combining situation:-(")
-    return None
+    model="SW Beta Adjustment"
+    scope="SP500"
+    ticker="AAPL"
+    draw2_betas(model,scope,ticker,betas_AAPL)
 
 #==============================================================================
-#==============================================================================
-#==============================================================================
+def get_beta_dimson(stkcd,mktidx,yearlist,printout=True,graph=True):
+    """
+    函数功能：使用Dimson(1979)方法调整一只股票的CAPM模型贝塔系数
+    输入参数：
+    stkcd: 股票代码
+    mktidx: 指数代码
+    yearlist：年度列表，用于计算年度贝塔系数
+    输出数据：显示CAPM市场模型回归的beta, 以及调整后的beta系数
+    返回数据：CAPM的beta, 以及调整后的beta系数
+    """
 
+    #仅为测试用，完成后应立即注释掉
+    #stkcd='0700.HK'
+    #mktidx='^HSI'
+    #yearlist=['2015','2016','2017','2018']
+    
+    #生成开始结束日期
+    Y4=str(int(yearlist[0])-1)
+    start=Y4+'-01-01'
+    end=yearlist[-1]+'-12-31'  
+        
+    #读取股价并准备好收益率数据
+    try:
+        R=prepare_capm(stkcd,mktidx,start,end)
+    except:
+        print("  #Error(get_beta_dimson): preparing CAPM data failed!")
+        return None
 
+    if (R is None):
+        print("  #Error(get_beta_dimson): server did not respond")
+        return None
+    if (len(R) == 0):
+        print("  #Error(get_beta_dimson): server returned empty data")
+        return None
 
+    #用于保存beta(CAPM)和beta(Dimson)
+    import pandas as pd
+    betas=pd.DataFrame(columns=('Year','Beta(CAPM)','Beta(Dimson)'))
 
+    #计算Dimson(1979)调整
+    R['Close_x+1']=R['Close_x'].shift(1)    
+    R['Close_x-1']=R['Close_x'].shift(-1)   
+    R=R.dropna()
 
+    from scipy import stats    
+    import statsmodels.api as sm
+    for year in yearlist:
+        r=R[R['Year']==year]
+        if len(r) != 0:   
+            output=stats.linregress(r['Close_x'],r['Close_y'])
+            (beta_capm,alpha,r_value,p_value,std_err)=output               
+            
+            #三个解释变量
+            RX=r[['Close_x-1','Close_x','Close_x+1']]
+            X1=sm.add_constant(RX)	#要求回归具有截距项
+            Y=r['Close_y']
+            model = sm.OLS(Y,X1)	#定义回归模型，X1为多元矩阵
+            results = model.fit()	#进行OLS回归
 
+            (alpha,beta_1,beta0,beta1)=results.params	#提取回归系数
+            beta_dimson=beta_1+beta0+beta1            
 
+            row=pd.Series({'Year':year,'Beta(CAPM)':beta_capm, \
+                           'Beta(Dimson)':beta_dimson})
+            try:
+                betas=betas.append(row,ignore_index=True)
+            except:
+                betas=betas._append(row,ignore_index=True)
 
+    betas.set_index(["Year"], inplace=True)
 
+    if printout == True: printdf_betas(betas,2)
+    if graph == True:
+        model="贝塔系数的Dimson调整法"
+        draw2_betas(model,mktidx,stkcd,betas)
 
+    return betas
+    
+if __name__=='__main__':
+    yearlist=gen_yearlist('2010','2019')
+    betas_MSFT=get_beta_dimson('MSFT','^GSPC',yearlist)
+    
+    model="Dimson Beta Adjustment"
+    scope="SP500"
+    ticker="MSFT"
+    draw2_betas(model,scope,ticker,betas_MSFT)
 
+    betas_MSFT2=get_beta_dimson('MSFT','^DJI',yearlist)
+    
+    model="Dimson Beta Adjustment"
+    scope="DJIA"
+    ticker="MSFT"
+    draw2_betas(model,scope,ticker,betas_MSFT2)
+
+#==============================================================================
+#============