siat 3.10.130__py3-none-any.whl → 3.10.132__py3-none-any.whl

build/lib/siat/valuation_china.py

@@ -0,0 +1,2076 @@
+# -*- coding: utf-8 -*-
+"""
+本模块功能：中国股市估值
+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！
+"""
+#==============================================================================
+#屏蔽所有警告性信息
+import warnings; warnings.filterwarnings('ignore')
+#==============================================================================
+from siat.common import *
+from siat.translate import *
+from siat.grafix import *
+from siat.security_prices import *
+from siat.security_price2 import *
+from siat.sector_china import *
+
+#==============================================================================
+
+if __name__ =="__main__":
+    start='2020-1-1'; end='2022-10-9'
+    measure='pb'; method='lyr'; value='value'; statistic='median'
+
+def get_valuation_market_china(start,end,measure='pe',method='lyr',value='value',statistic='median'):
+    """
+    功能：中国A股市场估值趋势，一段时间内
+    measure：默认市盈率'pe'，可选市净率pb
+    method：默认滚动'ttm'，可选静态lyr
+    value：默认数值'value'，可选分位数quantile
+    statistic：默认使用中位数'median'，可选等权重equal-weighted
+    """
+    
+    #检查日期的合理性
+    result,startpd,endpd=check_period(start,end)
+    if not result:
+        print("  #Error(get_valuation_market_china): invalid date period",start,end)
+        return None
+    
+    #检查选项
+    measure1=measure.lower(); measurelist=['pe','pb']
+    method1=method.lower(); methodlist=['ttm','lyr']
+    value1=value.lower(); valuelist=['value','quantile']
+    statistic1=statistic.lower(); statisticlist=['median','equal-weighted']
+    
+    if not (measure1 in measurelist):
+        print("  #Error(get_valuation_market_china): invalid measurement",measure)
+        print("  Valid measurement:")
+        return None
+    if not (method1 in methodlist):
+        print("  #Error(get_valuation_market_china): invalid method",method)
+        print("  Valid method:",methodlist)
+        return None    
+    if not (value1 in valuelist):
+        print("  #Error(get_valuation_market_china): invalid value",value)
+        print("  Valid value:",valuelist)
+        return None
+    if not (statistic1 in statisticlist):
+        print("  #Error(get_valuation_market_china): invalid statistic",statistic)
+        print("  Valid statistic:",statisticlist)
+        return None
+
+    # 构造组合矩阵   
+    import pandas as pd
+    matrix=pd.DataFrame([
+        #['pe ttm value median','middlePETTM','市盈率(滚动TTM，全A股中位数)'],
+        #['pe ttm value equal-weighted','averagePETTM','市盈率(滚动TTM，全A股等权平均)'],
+        #['pe ttm quantile median','quantileInRecent10YearsMiddlePeTtm','市盈率分位数(滚动TTM，全A股中位数，近10年)'],
+        #['pe ttm quantile equal-weighted','quantileInRecent10YearsAveragePeTtm','市盈率分位数(滚动TTM，全A股等权平均，近10年)'],
+         
+        #['pe lyr value median','middlePELYR','市盈率(静态LYR，全A股中位数)'],
+        #['pe lyr value equal-weighted','averagePELYR','市盈率(静态LYR，全A股等权平均)'],
+        #['pe lyr quantile median','quantileInRecent10YearsMiddlePeLyr','市盈率分位数(静态LYR，全A股中位数，近10年)'],
+        #['pe lyr quantile equal-weighted','quantileInRecent10YearsAveragePeLyr','市盈率分位数(静态LYR，全A股等权平均，近10年)'],
+         
+        #['pb lyr value median','middlePB','市净率(静态LYR，全A股中位数)'],
+        #['pb lyr value equal-weighted','equalWeightAveragePB','市净率(静态LYR，全A股等权平均)'],
+        #['pb lyr quantile median','quantileInRecent10YearsMiddlePB','市净率分位数(静态LYR，全A股中位数，近10年)'],
+        #['pb lyr quantile equal-weighted','quantileInRecent10YearsEqualWeightAveragePB','市净率分位数(静态LYR，全A股等权平均，近10年)'],
+        
+        ['pe ttm value median','middlePETTM','市盈率(TTM，全A股中位数)'],
+        ['pe ttm value equal-weighted','averagePETTM','市盈率(TTM，全A股等权平均)'],
+        ['pe ttm quantile median','quantileInRecent10YearsMiddlePeTtm','市盈率分位数(TTM，全A股近十年中位数)'],
+        ['pe ttm quantile equal-weighted','quantileInRecent10YearsAveragePeTtm','市盈率分位数(TTM，全A股近十年等权平均)'],
+         
+        ['pe lyr value median','middlePELYR','市盈率(全A股中位数)'],
+        ['pe lyr value equal-weighted','averagePELYR','市盈率(全A股等权平均)'],
+        ['pe lyr quantile median','quantileInRecent10YearsMiddlePeLyr','市盈率分位数(全A股近十年中位数)'],
+        ['pe lyr quantile equal-weighted','quantileInRecent10YearsAveragePeLyr','市盈率分位数(全A股近十年等权平均)'],
+         
+        ['pb lyr value median','middlePB','市净率(全A股中位数)'],
+        ['pb lyr value equal-weighted','equalWeightAveragePB','市净率(全A股等权平均)'],
+        ['pb lyr quantile median','quantileInRecent10YearsMiddlePB','市净率分位数(全A股近十年中位数)'],
+        ['pb lyr quantile equal-weighted','quantileInRecent10YearsEqualWeightAveragePB','市净率分位数(全A股近十年等权平均)'],
+        
+        ], columns=['combine','field','desc'])
+
+    #查找组合方式对应的字段名称
+    combine=measure1+' '+method1+' '+value1+' '+statistic1
+    try:
+        field=matrix[matrix['combine']==combine]['field'].values[0]
+        desc=matrix[matrix['combine']==combine]['desc'].values[0]
+    except:
+        #未查到组合
+        print("  #Error(get_valuation_market_china): parameter combination not available for",combine)
+        return None
+
+    import akshare as ak
+    
+    #获取全A股市场的市盈率
+    if measure1 == 'pe':
+        try:
+            mp = ak.stock_a_ttm_lyr()
+        except:
+            #akshare版本需要更新
+            print("  #Error(get_valuation_market_china): may need to upgrade akshare")
+            return None
+    
+        #截取选定的日期范围
+        mp['Date']=mp['date']
+        mp.sort_values(by=['Date'],ascending=True,inplace=True)
+        mp.set_index(['Date'],inplace=True) 
+        
+        try:
+            mp1=mp[(mp.index >= startpd) & (mp.index <=endpd)]
+        except:
+            startpd=startpd.date()
+            endpd=endpd.date()
+            mp1=mp[(mp.index >= startpd) & (mp.index <=endpd)]            
+        
+        mp9=mp1[['date',field,'close']]
+        mp9['field']=mp9[field]
+        mp9['index']=mp9['close']
+        mp9['index name']="沪深300指数"
+        mp9['measure']=measure1
+        mp9['method']=method1
+        mp9['value']=value1
+        mp9['statistic']=statistic1
+        mp9['desc']=desc
+    
+    
+    #获取全A股市场的市净率
+    if measure1 == 'pb':
+        try:
+            mp = ak.stock_a_all_pb()
+        except:
+            #akshare版本需要更新
+            print("  #Error(get_valuation_market_china): may need to upgrade akshare")
+            return None
+    
+        #截取选定的日期范围
+        mp['Date']=mp['date']
+        mp.sort_values(by=['Date'],ascending=True,inplace=True)
+        mp.set_index(['Date'],inplace=True) 
+        
+        try:
+            mp1=mp[(mp.index >= startpd) & (mp.index <=endpd)]
+        except:
+            startpd=startpd.date()
+            endpd=endpd.date()
+            mp1=mp[(mp.index >= startpd) & (mp.index <=endpd)]            
+        
+        mp9=mp1[['date',field,'close']]
+        mp9['field']=mp9[field]
+        mp9['index']=mp9['close']
+        mp9['index name']="上证综合指数"
+        mp9['measure']=measure1
+        mp9['method']=method1
+        mp9['value']=value1
+        mp9['statistic']=statistic1
+        mp9['desc']=desc
+        
+    df=mp9[['date','field','index','index name','measure','method','value','statistic','desc']]
+    
+    return df    
+
+if __name__ =="__main__":
+    start='2020-1-1'; end='2022-10-9'
+    df=get_valuation_market_china(start,end,measure='pe',method='lyr',value='value',statistic='median')
+
+#==============================================================================
+if __name__ =="__main__":
+    start='2024-1-1'; end='2025-3-31'
+    indicator='pe'
+    method='ttm'
+    value='value'
+    statistic='median'
+    
+    indicator=['pb','pe']
+    method=['lyr','ttm']
+    value=['value','quantile']
+    statistic=['median','equal-weighted']
+    
+    twinx=False; loc1='best'; loc2='best'
+    power=0; twinx=False; average_value=True
+    annotate=False; annotate_value=False; plus_sign=True
+    mark_top=False; mark_bottom=False; mark_end=False
+        
+    loc1='upper left'; loc2='lower right'
+    facecolor='whitesmoke'; maxticks=20
+    
+
+
+def valuation_china(start='MRY',end='today',indicator='pe', \
+                    method='lyr',value='value',statistic='median', \
+                    average_value=False,show_index=False, \
+                    power=0,twinx=False, \
+                               
+                    band_area='', \
+                    attention_value='',attention_value_area='', \
+                    attention_point='',attention_point_area='', \
+                               
+                    annotate=False,annotate_value=False,plus_sign=True, \
+                    mark_start=False,mark_top=False,mark_bottom=False,mark_end=False, \
+                               
+                    loc1='upper left',loc2='lower right', \
+                    facecolor='whitesmoke',maxticks=20):
+    """
+    ===========================================================================
+    功能：比较中国全A股市场的估值指标变化趋势
+    start: 开始日期，格式YYYY-MM-DD。
+    或者简写版，例如'MRY'近1年、'L3Y'表示近3年等。
+    end: 结束日期，默认今天。
+    
+    indicator: 估值指标市盈率'pe'或市净率'pb'，或其组合。不支持股息率
+    method: 滚动'ttm或静态取样'lyr'，或其组合
+    value: 直接采用估值指标数值'value'或分位数'quantile'，或其组合
+    statistic: 采用中位数'median'或等权均值'equal-weighted'，或其组合
+    twinx：是否使用双轴绘图法，默认否False。
+        如果同时选择了市盈率'pe'或市净率'pb'，建议打开该选项True。
+    loc1/loc2：图例1/2的位置，默认自动决定'best'。
+        如果自动位置不理想，可以手动设置位置。
+    """
+    print("  Working on valuating China stock market ... ...")
+    
+    import pandas as pd
+    #处理日期
+    start,end=start_end_preprocess(start,end)
+    
+    # 情形1：双indicator：双指标，PE+PB
+    if isinstance(indicator,list) and len(indicator) >= 2:
+        indicator=indicator[:2]
+        if isinstance(method,list): method=method[0]
+            
+        if isinstance(value,list): value=value[0]
+        if isinstance(statistic,list): statistic=statistic[0]
+        
+        df=None
+        for i in indicator:
+            if 'pb' != i.lower():
+                dftmp=get_valuation_market_china(start,end,measure=i.lower(), \
+                                                 method=method, \
+                                                 value=value,statistic=statistic)
+            else:
+                # pb不支持ttm
+                dftmp=get_valuation_market_china(start,end,measure=i.lower(), \
+                                                 method='lyr', \
+                                                 value=value,statistic=statistic)
+                
+            val_desc=dftmp['desc'].values[0]
+            method_desc=dftmp['method'].values[0]
+            """
+            if method_desc == 'lyr': val_desc=val_desc+'静态'
+            elif method_desc == 'ttm': 
+                #val_desc=val_desc+'动态'
+                val_desc=val_desc
+            if value == 'value': val_desc=val_desc+'数值'
+            elif value == 'quantile': val_desc=val_desc+'分位数'
+            """
+            dftmp[val_desc]=dftmp['field']
+            dftmp2=dftmp[[val_desc]]
+            
+            if df is None:
+                df=dftmp2
+            else:
+                df=pd.merge(df,dftmp2,left_index=True,right_index=True)
+                
+    # 情形2：单indicator+双method：ttm+lyr（仅适用于pe；pb仅适用lyr）；
+    elif isinstance(method,list) and len(method) >= 2:
+        method=method[:2]
+        if isinstance(indicator,list): indicator=indicator[0]
+            
+        if isinstance(value,list): value=value[0]
+        if isinstance(statistic,list): statistic=statistic[0]
+        
+        df=None
+        for i in method:
+            if (indicator.lower() == 'pe') or \
+                (indicator.lower() == 'pb' and i.lower() == 'lyr'):
+                dftmp=get_valuation_market_china(start,end,measure=indicator.lower(), \
+                                                 method=i.lower(), \
+                                                 value=value,statistic=statistic)
+                
+            val_desc=dftmp['desc'].values[0]
+            method_desc=dftmp['method'].values[0]
+            """
+            if method_desc == 'lyr': val_desc=val_desc+'静态'
+            elif method_desc == 'ttm': 
+                #val_desc=val_desc+'动态'
+                val_desc=val_desc
+            if value == 'value': val_desc=val_desc+'数值'
+            elif value == 'quantile': val_desc=val_desc+'分位数'
+            """
+            dftmp[val_desc]=dftmp['field']
+            dftmp2=dftmp[[val_desc]]
+            
+            if df is None:
+                df=dftmp2
+            else:
+                df=pd.merge(df,dftmp2,left_index=True,right_index=True)
+                
+    # 情形3：单indicator+单method+双value；
+    elif isinstance(value,list) and len(value) >= 2:
+        value=value[:2]
+        if isinstance(indicator,list): indicator=indicator[0]
+        if isinstance(method,list): method=method[0]
+        if indicator == 'pb': method='lyr'
+            
+        if isinstance(statistic,list): statistic=statistic[0]
+        
+        df=None
+        for i in value:
+            dftmp=get_valuation_market_china(start,end,measure=indicator.lower(), \
+                                             method=method.lower(), \
+                                             value=i.lower(),statistic=statistic)
+                
+            val_desc=dftmp['desc'].values[0]
+            method_desc=dftmp['method'].values[0]
+            """
+            if method_desc == 'lyr': val_desc=val_desc+'静态'
+            elif method_desc == 'ttm': 
+                #val_desc=val_desc+'动态'
+                val_desc=val_desc
+            if value == 'value': val_desc=val_desc+'数值'
+            elif value == 'quantile': val_desc=val_desc+'分位数'
+            """
+            dftmp[val_desc]=dftmp['field']
+            dftmp2=dftmp[[val_desc]]
+            
+            if df is None:
+                df=dftmp2
+            else:
+                df=pd.merge(df,dftmp2,left_index=True,right_index=True)
+                
+        # 数值与分位数差距较大，使用twinx=True
+        twinx=True
+                
+    # 情形4：单indicator+单method+单value+双statistic；
+    elif isinstance(statistic,list) and len(statistic) >= 2:
+        statistic=statistic[:2]
+        if isinstance(indicator,list): indicator=indicator[0]
+        if isinstance(method,list): method=method[0]
+        if indicator == 'pb': method='lyr'
+        
+        df=None
+        for i in statistic:
+            dftmp=get_valuation_market_china(start,end,measure=indicator.lower(), \
+                            method=method.lower(), \
+                            value=value.lower(),statistic=i.lower())
+                
+            val_desc=dftmp['desc'].values[0]
+            method_desc=dftmp['method'].values[0]
+            """
+            if method_desc == 'lyr': val_desc=val_desc+'静态'
+            elif method_desc == 'ttm': 
+                #val_desc=val_desc+'动态'
+                val_desc=val_desc
+            if value == 'value': val_desc=val_desc+'数值'
+            elif value == 'quantile': val_desc=val_desc+'分位数'
+            """
+            dftmp[val_desc]=dftmp['field']
+            dftmp2=dftmp[[val_desc]]
+            
+            if df is None:
+                df=dftmp2
+            else:
+                df=pd.merge(df,dftmp2,left_index=True,right_index=True)
+                
+    # 情形5：单indicator+单method+单value+单statistic；
+    else:
+        if isinstance(indicator,list): indicator=indicator[0]
+        if isinstance(method,list): method=method[0]
+        if indicator == 'pb': method='lyr'
+        
+        if isinstance(value,list): value=value[0]
+        if isinstance(statistic,list): statistic=statistic[0]
+        
+        dftmp=get_valuation_market_china(start,end,measure=indicator.lower(), \
+                                         method=method.lower(), \
+                            value=value.lower(),statistic=statistic.lower())
+            
+        val_desc=dftmp['desc'].values[0]
+        method_desc=dftmp['method'].values[0]
+        """
+        if method_desc == 'lyr': val_desc=val_desc+'静态'
+        elif method_desc == 'ttm': 
+            #val_desc=val_desc+'动态'
+            val_desc=val_desc
+        if value == 'value': val_desc=val_desc+'数值'
+        elif value == 'quantile': val_desc=val_desc+'分位数'
+        """
+        dftmp[val_desc]=dftmp['field']
+        if show_index:
+            index_desc=dftmp['index name'].values[0]
+            dftmp[index_desc]=dftmp['index']
+            df=dftmp[[val_desc,index_desc]]
+            
+            # 指数与指标差异大，设定twinx=True
+            twinx=True
+        else:
+            df=dftmp[[val_desc]]
+        
+    # 绘图
+    import datetime; todaydt = datetime.date.today()
+    sourcetxt=text_lang("数据来源：legulegu","Data source: legulegu")
+    footnote=sourcetxt+', '+str(todaydt)
+    
+    if show_index:
+        # 显示股票市场指数
+        twinx=True
+        
+    titletxt=text_lang("中国A股市场估值趋势","China A-share Market Valuation")
+    ylabeltxt=text_lang("估值指标","Valuation Indicator")
+    
+    if len(list(df)) == 1:
+        colname=collabel=list(df)[0]
+        #titletxt=titletxt+': '+colname
+        ylabeltxt=colname
+        plot_line(df,colname,collabel,ylabeltxt,titletxt,footnote, \
+                      power=power, \
+            attention_value=attention_value,attention_value_area=attention_value_area, \
+            attention_point=attention_point,attention_point_area=attention_point_area, \
+                      average_value=average_value, \
+                          
+                      loc=loc1, \
+                      mark_start=True,mark_top=True,mark_bottom=True,mark_end=mark_end, \
+                      facecolor=facecolor,maxticks=maxticks)
+    
+    #elif twinx and len(list(df)) >= 2:
+    elif twinx != False and len(list(df)) >= 2:
+        ticker1=ticker2=''
+        colname1=label1=list(df)[0]; df1=df[[colname1]]
+        colname2=label2=list(df)[1]; df2=df[[colname2]]
+        
+        plot_line2(df1,ticker1,colname1,label1, \
+                   df2,ticker2,colname2,label2, \
+                   ylabeltxt,titletxt,footnote, \
+                   power=power,datatag1=False,datatag2=False,yscalemax=5, \
+                   zeroline=False, \
+                       twinx=twinx, \
+                       yline=999,attention_value_area='', \
+                       xline=999,attention_point_area='', \
+                   resample_freq='H',loc1=loc1,loc2=loc2, \
+                   color1='red',color2='blue',facecolor='whitesmoke', \
+                       maxticks=maxticks)
+    else:
+        x_label=footnote
+        axhline_value=0; axhline_label=''
+        draw_lines(df,ylabeltxt,x_label,axhline_value,axhline_label,titletxt, \
+                   band_area=band_area,loc=loc1, \
+            attention_value=attention_value,attention_value_area=attention_value_area, \
+            attention_point=attention_point,attention_point_area=attention_point_area, \
+                   annotate=annotate,annotate_value=annotate_value,plus_sign=False, \
+                   mark_top=mark_top,mark_bottom=mark_bottom,mark_end=mark_end, \
+                   facecolor=facecolor,maxticks_enable=True,maxticks=maxticks)
+
+    return df
+
+
+if __name__ =="__main__":
+    start='2020-1-1'; end='2022-10-9'
+    measures=['pb','pe']; methods='lyr'; values='value'; statistics='median'
+    
+    measures='pe'; methods=['lyr','ttm']; values='value'; statistics='median'
+
+
+def valuation_market_china(start='MRY',end='today',measures=['pe','pb'], \
+                           methods='lyr',values='value',statistics='median', \
+                           twinx=False,loc1='best',loc2='best'):
+    """
+    ===========================================================================
+    功能：比较中国全A股市场的估值指标变化趋势
+    start: 开始日期，格式YYYY-MM-DD。
+    或者简写版，例如'MRY'近1年、'L3Y'表示近3年等。
+    end: 结束日期，默认今天。
+    
+    measures: 估值指标市盈率'pe'或市净率'pb'，不支持股息率
+    methods: 滚动'ttm/静态取样'lyr'.
+    values: 直接采用估值指标数值'value'或分位数'quantile'
+    statistics: 采用中位数'median'或等权均值'equal-weighted'
+    twinx：是否使用双轴绘图法，默认否False。
+        如果同时选择了市盈率'pe'或市净率'pb'，建议打开该选项True。
+    loc1/loc2：图例1/2的位置，默认自动决定'best'。
+        如果自动位置不理想，可以手动设置位置。
+    """
+    #处理日期
+    start,end=start_end_preprocess(start,end)
+    
+    #解析比较的指标，以第一个双指标为准
+    found2=False
+    parmlist=['measures','methods','values','statistics']
+    for v in parmlist:
+        
+        #如果是一个字符串
+        if isinstance(eval(v),str): 
+            globals()[v+'1']=eval(v)
+            globals()[v+'2']=eval(v)
+
+        #如果是一个列表
+        if isinstance(eval(v),list): 
+            num = len(eval(v))
+            #print("num=",num)
+            
+            if num == 0:
+                print("  #Error(valuation_market_china)：need at least 1 parameter for",eval(v))
+                return None,None
+            
+            if num == 1:
+                globals()[v+'1']=eval(v)[0]
+                globals()[v+'2']=eval(v)[0]
+            
+            
+            if num >= 2:
+                globals()[v+'1']=eval(v)[0]
+                globals()[v+'2']=eval(v)[1]
+                found2=True
+                
+    if not found2:
+        print("  #Warning(valuation_market_china)：parameters mismatch among",parmlist)
+        #return None,None
+    """            
+    print("measures1=",measures1,"measures2=",measures2)
+    print("methods1=",methods1,"methods2=",methods2)
+    print("values1=",values1,"values2=",values2)
+    print("statistics1=",statistics1,"statistics2=",statistics2)
+    """
+    
+    if values=='value':
+        ylabeltxt='估值比率'
+    else:
+        ylabeltxt='分位数'
+        
+    print("  Working on valuating China stock market ... ...")
+        
+    titletxt='中国全A股市场估值的变化趋势'
+    
+    import datetime
+    today = datetime.date.today()
+    footnote="数据来源: legulegu，"+str(today)
+    
+    #获取指标1
+    df1=get_valuation_market_china(start,end,measure=measures1,method=methods1,value=values1,statistic=statistics1)
+    if df1 is None:
+        print("  #Error(valuation_market_china)：no data available for the combine of",measures1,methods1,values1,statistics1)
+        return None,None
+    
+    ticker1=df1['desc'].values[0]
+    colname1='field'
+    label1=''
+    
+    if not found2:  
+        plot_line(df1,colname1,ticker1,ylabeltxt,titletxt,footnote, \
+                      power=0,loc=loc1, \
+                      date_fmt='%Y-%m-%d')
+        return df1,None
+
+    #获取指标2
+    df2=get_valuation_market_china(start,end,measure=measures2,method=methods2,value=values2,statistic=statistics2)
+    if df2 is None:
+        print("  #Error(valuation_market_china)：data unavailable for the combine of",measures2,methods2,values2,statistics2)
+        return None,None
+    
+    ticker2=df2['desc'].values[0]
+    colname2='field'
+    label2=''
+    
+    if twinx == 'auto':
+        twinx=False
+        
+        max1=df1[colname1].max()
+        max2=df2[colname2].max()
+        bili=max1/max2
+        if (bili > 2) or (bili < 0.5):
+            twinx=True
+    
+    plot2_line2(df1,ticker1,colname1,label1, \
+               df2,ticker2,colname2,label2, \
+               ylabeltxt,titletxt,footnote, \
+               twinx=twinx,loc1=loc1,loc2=loc2)
+    #清除变量
+    #"""    
+    for v in parmlist:
+        del globals()[v+'1'],globals()[v+'2']
+    #"""
+    
+    return df1,df2
+
+if __name__ =="__main__":
+    start='2020-1-1'; end='2022-10-9'
+    measures=['pe','pb']; methods='lyr'; values='value'; statistics='median'
+    df1,df2=valuation_market_china(start,end,measures=['pe','pb'],methods='lyr',values='value',statistics='median')
+
+#==============================================================================
+# 行业估值：申万宏远行业，韭圈儿
+#==============================================================================
+if __name__=='__main__':
+    top=5
+    vtype='PE'
+    vsorting='quantile'
+    printout=True
+    graph=True
+    axisamp=3
+    px=False
+    
+def industry_valuation_sw(top=10,vtype='PE',vsorting='quantile', \
+                          graph=True,axisamp=1.2,px=False):
+    """
+    功能：列示申万行业指数估值最高和最低的行业
+    vtype: PE, PB, dividend
+    vsorting: 分位数绝对值
+    """
+    
+    import akshare as ak
+    # 如果出错，升级一下akshare
+    df = ak.index_value_name_funddb()    
+    
+    # 筛选申万行业指数
+    substr='(申万)'
+    df['申万标志']=df['指数名称'].apply(lambda x: substr in x)
+    df1=df[df['申万标志']]
+    df1['行业代码']=df1['指数代码'].apply(lambda x: x[:6])
+    df1['行业名称']=df1['指数名称'].apply(lambda x: x[:x.index(substr)])
+    
+    #检查估值类型
+    typelist=['pe','pb','dividend']
+    vtypeu=vtype.lower()
+    if not (vtypeu in typelist):
+        print("  #Warning(industry_valuation_sw): unsupported valuation type",vtype)
+        print("  Supported types:",typelist)
+        return None
+    
+    #检查排序类型
+    sortlist=['quantile','value']
+    vsortingu=vsorting.lower()
+    if not (vsortingu in sortlist):
+        print("  #Warning(industry_valuation_sw): unsupported sorting type",vsorting)
+        print("  Supported types:",sortlist)
+        return None
+    
+    #排序：高位优先
+    if vtypeu == 'pe':
+        if vsortingu == 'value':
+            df2=df1.sort_values(by='最新PE',ascending=False)
+            collist=['行业名称','最新PE','PE分位','最新PB','PB分位','股息率','股息率分位','行业代码']
+            colname='最新PE'
+        else:
+            df2=df1.sort_values(by='PE分位',ascending=False)
+            collist=['行业名称','PE分位','最新PE','PB分位','最新PB','股息率','股息率分位','行业代码']
+            colname='PE分位'
+        
+    if vtypeu == 'pb':
+        if vsortingu == 'value':
+            df2=df1.sort_values(by='最新PB',ascending=False)
+            collist=['行业名称','最新PB','PB分位','最新PE','PE分位','股息率','股息率分位','行业代码']
+            colname='最新PB'
+        else:
+            df2=df1.sort_values(by='PB分位',ascending=False)
+            collist=['行业名称','PB分位','最新PB','最新PE','PE分位','股息率','股息率分位','行业代码']
+            colname='PB分位'
+        
+    if vtypeu == 'dividend':
+        if vsortingu == 'value':
+            df2=df1.sort_values(by='股息率',ascending=False)
+            collist=['行业名称','股息率','股息率分位','最新PB','PB分位','最新PE','PE分位','行业代码']
+            colname='股息率'
+        else:
+            df2=df1.sort_values(by='股息率分位',ascending=False)
+            collist=['行业名称','股息率分位','股息率','最新PB','PB分位','最新PE','PE分位','行业代码']
+            colname='股息率分位'
+            
+    df2.reset_index(drop=True,inplace=True)
+    df2.index=df2.index+1
+    df3=df2[collist]
+    
+    if top > 0:
+        df4=df3.head(top)
+    elif top < 0:
+        df4=df3.tail(-top)
+    else:
+        df4=df3
+    df5=df4.set_index('行业名称')
+    df5.sort_values(by=colname,ascending=True,inplace=True)
+
+    #绘图
+    if graph:
+        if top > 0:
+            prefix="最高的"
+        else:
+            prefix="最低的"
+
+        if vsortingu=='quantile':
+            suffix="数百分比"
+            footnote1="分位数表示历史上比当前便宜的百分比，"
+        else:
+            suffix="数值"
+            footnote1=''
+            
+        titletxt="估值分析：基于"+colname+suffix+"，"+prefix+str(abs(top))+"个行业"
+        
+        import datetime
+        today = datetime.date.today()
+        footnote0="注：申万宏源行业分类，"
+        footnote2="数据来源: 申万宏源/韭圈儿，"+str(today)
+        footnote=footnote0+footnote1+footnote2
+        
+        if not px:
+            fig=plot_barh(df5,colname,titletxt,footnote,axisamp=axisamp)
+        else:
+            fig=plot_barh2(df5,colname,titletxt,footnote)
+        
+    return df2
+
+if __name__=='__main__':
+    df=industry_valuation_sw(top=10,vtype='PE',vsorting='quantile',axisamp=2.3)
+    df=industry_valuation_sw(top=-10,vtype='PE',vsorting='quantile',axisamp=1.2)
+    df=industry_valuation_sw(top=10,vtype='PE',vsorting='value',axisamp=1.5)
+    df=industry_valuation_sw(top=-10,vtype='PE',vsorting='value',axisamp=1.6)
+    
+    df=industry_valuation_sw(top=10,vtype='PB',vsorting='quantile',axisamp=2.1)
+    df=industry_valuation_sw(top=-10,vtype='PB',vsorting='quantile',axisamp=1.2)
+    df=industry_valuation_sw(top=10,vtype='PB',vsorting='value',axisamp=2)
+    df=industry_valuation_sw(top=-10,vtype='PB',vsorting='value',axisamp=1.6)
+    
+    df=industry_valuation_sw(top=10,vtype='dividend',vsorting='quantile',axisamp=32)
+    df=industry_valuation_sw(top=-10,vtype='dividend',vsorting='quantile',axisamp=1.2) 
+    df=industry_valuation_sw(top=10,vtype='dividend',vsorting='value',axisamp=2)
+    df=industry_valuation_sw(top=-10,vtype='dividend',vsorting='value',axisamp=1.3)    
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    industry='食品饮料'
+    industry='生猪养殖'
+    industry='国防军工'
+    
+    industry='801853.SW'
+    
+    start='2021-1-1'
+    end='2022-11-15'
+    vtype='PE'
+    
+    graph=True
+    loc='best'
+    
+    df=industry_valuation_history_sw_daily(industry,start,end,vtype)
+    
+def industry_valuation_history_sw_daily(industry,start,end,vtype='PE', \
+                                  graph=True,loc='best', \
+                                  error_message=True):
+    """
+    功能：绘制一个申万行业的日历史估值趋势，不支持二级三级行业分类
+    vtype: PE, PB, dividend
+    
+    *** 注意：必须安装插件ipywidgets
+    如果出现下列错误信息：Error displaying widget: model not found
+    先卸载再重新安装插件ipywidgets
+    """
+    #检查日期期间
+    result,start1,end1=check_period(start,end)
+    if not result:
+        print("  #Warning(industry_valuation_history_sw_daily): invalid date period",start,end)
+        return None
+    
+    #检查估值类型
+    typelist=['pe','pb','dividend']
+    vtypeu=vtype.lower()
+    if not (vtypeu in typelist):
+        print("  #Warning(industry_valuation_history_sw_daily): unsupported valuation type",vtype)
+        print("  Supported types:",typelist)
+        return None
+
+    vtypelist=['pe','pb','dividend']
+    typelist=['市盈率','市净率','股息率']
+    pos=vtypelist.index(vtypeu)
+    vtypes=typelist[pos]
+
+    # 适配industry名称/代码
+    industry_split=industry.split('.')
+    split1=industry.split('.')[0]
+    split1_name=industry_sw_code(split1)
+    
+    #industry情形1：无后缀，名称
+    if len(industry_split)==1 and not split1.isdigit():
+        
+        if not split1_name is None: #是申万名称
+            sindustry=industry+'(申万)'
+        else: #不是申万名称
+            sindustry=industry
+    #industry情形2：数字
+    else: 
+        if not split1_name is None: #是申万代码
+            sindustry=industry_sw_name(split1)+'(申万)'
+        else: #不是申万代码
+            index_val=ak.index_value_name_funddb()
+            sindustry=index_val[index_val['指数代码']==industry]['指数名称'].values[0]
+    
+    # 获取行业估值历史数据
+    import akshare as ak
+    try:
+        # symbol：指数名称，申万行业名称需要加后缀(申万)，仅支持申万一级行业和部分二级行业，不支持申万三级行业
+        # 支持的指数查看：ak.index_value_name_funddb()
+        df = ak.index_value_hist_funddb(symbol=sindustry, indicator=vtypes)  
+    except:
+        if error_message:
+            print("  #Warning(industry_valuation_history_sw_daily): failed to access index",sindustry)
+        return None
+
+    import pandas as pd
+    df['date']=pd.to_datetime(df['日期'])
+    df.set_index('date',inplace=True)
+    df1=df[[vtypes]]
+    
+    #筛选期间
+    df2=df1[(df1.index >= start1) & (df1.index <= end1)]
+
+    #绘图
+    if graph:
+        df2['平均值']=df2[vtypes].mean()
+        df2['中位数']=df2[vtypes].median()
+        
+        titletxt="行业估值趋势："+industry_sw_name(industry)+'，'+vtypes   
+        
+        footnote0="注：申万宏源行业指数，"
+        footnote1=''
+        import datetime
+        today = datetime.date.today()
+        footnote2="数据来源: 申万宏源/韭圈儿，"+str(today)
+        footnote=footnote0+footnote1+footnote2
+        
+        colname=vtypes
+        collabel=vtypes
+        ylabeltxt=vtypes
+        
+        draw_lines(df2,y_label=ylabeltxt,x_label=footnote, \
+                   axhline_value=0,axhline_label='', \
+                   title_txt=titletxt,data_label=False,resample_freq='D')
+        
+    return df2
+
+if __name__=='__main__':
+    df=industry_valuation_history_sw_daily(industry,start,end,vtype='PE')
+    df=industry_valuation_history_sw_daily(industry,start,end,vtype='PB')
+    df=industry_valuation_history_sw_daily(industry,start,end,vtype='dividend')
+    
+    df=industry_valuation_history_sw_daily(industry='纺织服饰',start=start,end=end,vtype='PE')
+    df=industry_valuation_history_sw_daily(industry='纺织服饰',start=start,end=end,vtype='PB')
+    df=industry_valuation_history_sw_daily(industry='纺织服饰',start=start,end=end,vtype='dividend')
+
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    adate='2023-12-14'
+
+def get_last_friday(adate):
+    """
+    功能：给定日期，找出上一个周五的日期，配合申万指数估值函数使用
+    """
+    
+    result,fdate=check_date2(adate)
+    if not result: 
+        return None
+    
+    import pendulum
+    wrk=pendulum.parse(fdate).day_of_week
+    
+    import datetime
+    todaydt = datetime.date.today().strftime('%Y-%m-%d')
+    if fdate > todaydt:
+        fdate=todaydt
+
+    if wrk==5:
+        if fdate != todaydt:
+            adj=-1
+        else:
+            adj=-(2+wrk)
+    elif wrk==6:
+        adj=-1
+    else:
+        adj=-(2+wrk)
+    last_fri=date_adjust(fdate,adjust=adj)
+    
+    return last_fri
+
+if __name__=='__main__':
+    start='2023-1-1'
+    end='2023-12-14'
+    get_all_friday(start,end)
+    
+def get_all_friday(start,end):
+    """
+    功能：获取start和end之间所有的周五日期，配合申万指数估值函数使用
+    """
+    #import pandas as pd
+    start_fri=get_last_friday(start)
+    end_fri=get_last_friday(end)
+    
+    import akshare as ak
+    wrk_df=ak.index_analysis_week_month_sw("week")
+    wrk_df['Date']=wrk_df['date'].apply(lambda x:x.strftime('%Y-%m-%d'))
+    frilist=list(wrk_df['Date'])    
+    
+    period_frilist=[]
+    for f in frilist:
+        if (f >= start_fri) and (f <= end_fri):
+            period_frilist=period_frilist+[f]
+        
+    return period_frilist
+    
+#==============================================================================
+
+if __name__=='__main__':
+    industry='食品饮料'
+    industry='白酒Ⅱ'
+    start='2023-10-1'
+    end='2023-12-15'
+    vtype='PE'
+    
+    graph=True
+    loc='best'
+    df=industry_valuation_history_sw_weekly(industry,start,end,vtype)
+    
+def industry_valuation_history_sw_weekly(industry,start,end,vtype='PE', \
+                                       graph=True,loc='best'):
+    """
+    功能：绘制一个申万行业的周历史估值趋势，支持申万"市场表征", "一级行业", "二级行业", "风格指数"
+    不支持三级行业，若为非二级行业，转为industry_valuation_history_sw_daily函数处理日数据，专注处理二级行业周数据
+    vtype: PE, PB, dividend
+    
+    """
+    #检查日期期间
+    result,start1,end1=check_period(start,end)
+    if not result:
+        print("  #Warning(industry_valuation_history_sw_weekly): invalid date period",start,end)
+        return None
+    fridays=get_all_friday(start,end)
+    
+    #检查估值类型
+    typelist=['pe','pb','dividend']
+    vtypeu=vtype.lower()
+    if not (vtypeu in typelist):
+        print("  #Warning(industry_valuation_history_sw_weekly): unsupported valuation type",vtype)
+        print("  Supported types:",typelist)
+        return None
+
+    vtypelist=['pe','pb','dividend']
+    typelist=['市盈率','市净率','股息率']
+    pos=vtypelist.index(vtypeu)
+    vtypes=typelist[pos]
+    
+    #分辨申万行业代码类别
+    sw_codes=industry_sw_list()
+    sw_codes['type_name']=''
+    sw_codes['type_name']=sw_codes.apply(lambda x: '市场表征' if x['type']=='F' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '一级行业' if x['type']=='I' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '二级行业' if x['type']=='T' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '风格指数' if x['type']=='S' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '三级行业' if x['type']=='3' else x['type_name'],axis=1)
+    
+    industry1=industry.split('.')[0]
+    industry_name_flag=industry_code_flag=False
+    type_name=''
+    try:
+        type_name=sw_codes[sw_codes['name']==industry1]['type_name'].values[0]
+        industry_name_flag=True
+    except:
+        try:
+            type_name=sw_codes[sw_codes['code']==industry1]['type_name'].values[0]
+            industry_code_flag=True
+        except:
+            print("  #Warning(industry_valuation_history_sw_weekly): not a Shenwan index for",industry)
+            #return None
+
+    if type_name=='三级行业':
+        print("  #Warning(industry_valuation_history_sw_weekly): currently does not support Shenwan 3rd_level industry",industry)
+        return None
+
+    if not (type_name in ['二级行业','风格指数']):
+        df=industry_valuation_history_sw_daily(industry=industry,start=start,end=end,vtype=vtype, \
+                                               graph=graph,loc=loc,error_message=False)
+        if not (df is None):
+            return df
+
+    # 获取行业估值历史周数据：啰嗦方法，需要反复下载，容易出ipywidgets引起的错误，需要安装之或卸载后重新安装
+    import pandas as pd
+    import akshare as ak
+    df=None
+    for f in fridays:
+        f1=f[:4]+f[5:7]+f[8:]
+        try:
+            dft=ak.index_analysis_weekly_sw(symbol=type_name, date=f1)
+        except:
+            continue
+        
+        """
+        dft的结构：
+        ['指数代码','指数名称','发布日期','收盘指数','成交量','涨跌幅','换手率',
+         '市盈率','市净率','均价','成交额占比','流通市值','平均流通市值','股息率']
+        """
+        
+        if not (dft is None):
+            if industry_name_flag:
+                dft2=dft[dft['指数名称']==industry1]
+            if industry_code_flag:
+                dft2=dft[dft['指数代码']==industry1]                
+        
+        if df is None:
+            df=dft2
+        else:
+            df=pd.concat([df,dft2])
+
+    df['date']=pd.to_datetime(df['发布日期'])
+    df.set_index('date',inplace=True)
+    df1=df[[vtypes]]
+    
+    #筛选期间
+    #df2=df1[(df1.index >= start1) & (df1.index <= end1)]
+    df2=df1.dropna()
+
+    #绘图
+    if graph:
+        df2['平均值']=df2[vtypes].mean()
+        df2['中位数']=df2[vtypes].median()
+        
+        titletxt="行业估值趋势："+industry+'，'+vtypes   
+        
+        footnote0="注：申万宏源行业指数，"
+        footnote1=''
+        import datetime
+        today = datetime.date.today()
+        footnote2="数据来源: 申万宏源，"+str(today)
+        footnote=footnote0+footnote1+footnote2
+        
+        colname=vtypes
+        collabel=vtypes
+        ylabeltxt=vtypes
+        
+        draw_lines(df2,y_label=ylabeltxt,x_label=footnote, \
+                   axhline_value=0,axhline_label='', \
+                   title_txt=titletxt,data_label=False,resample_freq='D')
+        
+    return df2
+
+#==============================================================================
+
+if __name__=='__main__':
+    industry='食品饮料'
+    industry='白酒Ⅱ'
+    industry='绩优股指数'
+    industry='801853.SW'
+    
+    start='2023-10-1'
+    end='2023-12-15'
+    vtype='PE'
+    
+    graph=True
+    loc='best'
+    df=industry_valuation_history_sw(industry,start,end,vtype)
+    
+def industry_valuation_history_sw(industry,start,end,vtype='PE', \
+                                       graph=True,loc='best'):
+    """
+    功能：绘制一个申万行业的日历史估值趋势，支持申万"市场表征", "一级行业", "二级行业", "风格指数"
+    不支持三级行业，若为非二级行业，转为industry_valuation_history_sw_daily函数处理日数据，专注处理二级行业日数据
+    vtype: PE, PB, dividend
+    
+    """
+    #检查日期期间
+    result,start1,end1=check_period(start,end)
+    if not result:
+        print("  #Warning(industry_valuation_history_sw): invalid date period",start,end)
+        return None
+    
+    #检查估值类型
+    typelist=['pe','pb','dividend']
+    vtypeu=vtype.lower()
+    if not (vtypeu in typelist):
+        print("  #Warning(industry_valuation_history_sw): unsupported valuation type",vtype)
+        print("  Supported types:",typelist)
+        return None
+
+    vtypelist=['pe','pb','dividend']
+    typelist=['市盈率','市净率','股息率']
+    pos=vtypelist.index(vtypeu)
+    vtypes=typelist[pos]
+    
+    #分辨申万行业代码类别
+    sw_codes=industry_sw_list()
+    sw_codes['type_name']=''
+    sw_codes['type_name']=sw_codes.apply(lambda x: '市场表征' if x['type']=='F' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '一级行业' if x['type']=='I' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '二级行业' if x['type']=='T' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '风格指数' if x['type']=='S' else x['type_name'],axis=1)
+    sw_codes['type_name']=sw_codes.apply(lambda x: '三级行业' if x['type']=='3' else x['type_name'],axis=1)
+    
+    industry1=industry.split('.')[0]
+    industry_name_flag=industry_code_flag=False
+    try:
+        type_name=sw_codes[sw_codes['name']==industry1]['type_name'].values[0]
+        industry_name_flag=True; industry_name=industry1
+    except:
+        try:
+            type_name=sw_codes[sw_codes['code']==industry1]['type_name'].values[0]
+            industry_code_flag=True
+            industry_name=sw_codes[sw_codes['code']==industry1]['name'].values[0]
+        except:
+            print("  #Error(industry_valuation_history_sw): Shenwan industry not found for",industry)
+            return None
+
+    if type_name=='三级行业':
+        print("  #Error(industry_valuation_history_sw): currently does not support Shenwan 3rd_level industry",industry)
+        return None
+
+    #if not (type_name=='二级行业'):
+    if not (type_name in ['二级行业','风格指数']):
+        df=industry_valuation_history_sw_daily(industry=industry,start=start,end=end,vtype=vtype, \
+                                          graph=graph,loc=loc)
+        if not (df is None):
+            return df
+
+    # 获取行业估值历史周数据：笨方法，反复下载。易出ipywidgets引起的错误，可卸载后重装
+    import pandas as pd
+    import akshare as ak
+    start2=start1.strftime('%Y-%m-%d')
+    end2  =end1.strftime('%Y-%m-%d')
+    pdate=end2
+    dstep=7
+    df=None
+    while (pdate >= start2) or (abs(date_delta(pdate,start2)) < dstep):
+        if pdate >= start2:
+            enddate=pdate
+            fromdate=date_adjust(pdate,adjust=-dstep)
+        else:
+            enddate=start2
+            fromdate=pdate
+            
+        try:
+            fromdate1=fromdate[:4]+fromdate[5:7]+fromdate[8:10]
+            enddate1=enddate[:4]+enddate[5:7]+enddate[8:10]
+            dft=ak.index_analysis_daily_sw(symbol=type_name,start_date=fromdate1,end_date=enddate1)
+        except:
+            dft=None
+            """
+            try:
+                fromdate1=fromdate[:4]+fromdate[5:7]+fromdate[8:10]
+                enddate1=enddate[:4]+enddate[5:7]+enddate[8:10]
+                dft=ak.index_analysis_daily_sw(symbol=type_name,start_date=fromdate1,end_date=enddate1)
+            except:
+                continue
+            """
+        """
+        dft的结构：
+        ['指数代码','指数名称','发布日期','收盘指数','成交量','涨跌幅','换手率',
+         '市盈率','市净率','均价','成交额占比','流通市值','平均流通市值','股息率']
+        """
+        
+        if not (dft is None):
+            if industry_name_flag:
+                dft2=dft[dft['指数名称']==industry1]
+            if industry_code_flag:
+                dft2=dft[dft['指数代码']==industry1]                
+        
+            if df is None:
+                df=dft2
+            else:
+                df=pd.concat([df,dft2])
+        
+        # 开始下一轮循环    
+        pdate=date_adjust(fromdate,adjust=-1)
+
+    df.sort_values('发布日期',ascending=True,inplace=True)
+    df.drop_duplicates(inplace=True)
+    
+    #df=df[df.index >= start1]
+    #df.dropna(inplace=True)
+    
+    df['date']=pd.to_datetime(df['发布日期'])
+    df.set_index('date',inplace=True)
+    df1=df[[vtypes]]
+    
+    #筛选期间
+    #df2=df1[(df1.index >= start1) & (df1.index <= end1)]
+    df2=df1
+
+    #绘图
+    if graph:
+        df2['平均值']=df2[vtypes].mean()
+        df2['中位数']=df2[vtypes].median()
+        
+        titletxt="行业估值趋势："+industry_name+'，'+vtypes   
+        
+        footnote0="注：申万行业分类指数，"
+        footnote1=''
+        import datetime
+        today = datetime.date.today()
+        footnote2="数据来源: 申万宏源，"+str(today)
+        footnote=footnote0+footnote1+footnote2
+        
+        colname=vtypes
+        collabel=vtypes
+        ylabeltxt=vtypes
+        
+        draw_lines(df2,y_label=ylabeltxt,x_label=footnote, \
+                   axhline_value=0,axhline_label='', \
+                   title_txt=titletxt,data_label=False,resample_freq='D')
+        
+    return df2
+
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    industries=['食品饮料','纺织服饰']
+    industries=['银行','国有大型银行Ⅱ','股份制银行Ⅱ','城商行Ⅱ','农商行Ⅱ']
+    start='2023-12-1'
+    end='2023-12-15'
+    vtypes='PE'
+    
+    industries='纺织服饰'
+    vtypes=['PE','PB']
+    
+    graph=True
+    loc1='lower left'
+    loc2='upper right'
+    
+    df5=compare_industry_valuation_sw(industries,start,end,vtypes)
+
+def compare_industry_valuation_sw(industries,start,end,vtypes='PE', \
+                                  graph=True,loc1='best',loc2='best'):
+    """
+    功能：比较多个申万行业或者两个指标的历史估值趋势
+    条件：若industries为列表且多个，则取vtype的第一个值；
+    若industries为字符串或者列表但只有一个，则取vtype的前两个值比较，双轴
+    vtypes: PE, PB, dividend
+    
+    """
+    
+    #检查日期期间的合理性
+    
+    
+    vtypelist=['pe','pb','dividend']
+    typelist=['市盈率','市净率','股息率']
+
+    #检查行业个数：多个行业+单指标
+    if isinstance(industries,list) & (len(industries) >= 2):
+        
+        if isinstance(vtypes,str): 
+            vtype=vtypes
+        elif isinstance(vtypes,list): 
+            vtype=vtypes[0]
+
+        vtypeu=vtype.lower()
+        pos=vtypelist.index(vtypeu)
+        vtypec=typelist[pos]
+        
+        import pandas as pd
+        df=pd.DataFrame()        
+        for i in industries:
+            # debug
+            print("  Searching valuation info for",i,'\b, which may take time ...')
+            dft=industry_valuation_history_sw(i,start=start,end=end,vtype=vtype,graph=False)
+            if not (dft is None):
+                dft.rename(columns={vtypec:i},inplace=True)
+                if len(df) == 0:
+                    df=dft
+                else:
+                    df=pd.merge(df,dft,how='outer',left_index=True,right_index=True)
+            else:
+                continue
+            
+        #绘图
+        if graph:
+            titletxt="行业估值趋势对比："+vtypec   
+            
+            footnote0="注：申万宏源行业指数，"
+            footnote1=''
+            import datetime
+            today = datetime.date.today()
+            footnote2="数据来源: 申万宏源/韭圈儿，"+str(today)
+            footnote=footnote0+footnote1+footnote2
+            
+            ylabeltxt=vtypec
+            
+            draw_lines(df,y_label=ylabeltxt,x_label=footnote, \
+                       axhline_value=0,axhline_label='', \
+                       title_txt=titletxt,data_label=False,resample_freq='D')
+
+        return df
+
+    #检查行业个数：一个行业+双指标
+    if  ((isinstance(industries,str) | (isinstance(industries,list) & (len(industries) == 1)))) \
+        & (isinstance(vtypes,list) & (len(vtypes) >= 2)):
+        
+        if isinstance(industries,str): 
+            industry=industries
+        elif isinstance(industries,list): 
+            industry=industries[0]
+        
+        if isinstance(vtypes,str): 
+            ivtypelist=[vtypes]
+        elif isinstance(vtypes,list): 
+            ivtypelist=vtypes[:2]
+        
+        import pandas as pd
+        df=pd.DataFrame()        
+        for t in ivtypelist:
+            
+            dft=industry_valuation_history_sw(industry,start=start,end=end,vtype=t,graph=False)
+            
+            if len(df) == 0:
+                df=dft
+            else:
+                df=pd.merge(df,dft,how='outer',left_index=True,right_index=True)
+                
+        #绘图
+        if graph:
+            titletxt="行业估值趋势对比："+industry   
+            
+            footnote0="注：申万宏源行业指数，"
+            footnote1=''
+            import datetime
+            today = datetime.date.today()
+            footnote2="数据来源: 申万宏源/韭圈儿，"+str(today)
+            footnote=footnote0+footnote1+footnote2
+            
+            collist=list(df)
+            colname1=label1=collist[0]
+            colname2=label2=collist[1]
+            
+            plot_line2(df,'',colname1,label1, \
+                       df,'',colname2,label2, \
+                           ylabeltxt='',titletxt=titletxt,footnote=footnote, \
+                           twinx=True, \
+                           resample_freq='D',loc1=loc1,loc2=loc2, \
+                           color1='red',color2='blue')
+            
+        return df
+
+    #检查行业个数：一个行业+一个指标
+    if  ((isinstance(industries,str) | (isinstance(industries,list) & (len(industries) == 1)))) \
+        & ((isinstance(vtypes,str) | (isinstance(vtypes,list) & (len(vtypes) == 1)))):
+
+        if isinstance(industries,str): 
+            industry=industries
+        elif isinstance(industries,list): 
+            industry=industries[0]
+        
+        if isinstance(vtypes,str): 
+            ivtype=vtypes
+        elif isinstance(vtypes,list): 
+            ivtype=vtypes[0]
+            
+        df=industry_valuation_history_sw(industry,start,end,vtype=ivtype, \
+                                          graph=graph,loc=loc1)
+        return df
+            
+if __name__=='__main__':
+    df=compare_industry_valuation_sw(industries=['纺织服饰','国防军工','食品饮料'], \
+                                     start='2017-1-1',end='2022-11-15', \
+                                     vtypes='PE',loc1='lower left',loc2='upper right')
+
+    df=compare_industry_valuation_sw(industries=['纺织服饰'], \
+                                     start='2017-1-1',end='2022-11-15', \
+                                     vtypes='PE',loc1='lower left',loc2='upper right')
+
+    df=compare_industry_valuation_sw(industries=['纺织服饰'], \
+                                     start='2017-1-1',end='2022-11-15', \
+                                     vtypes=['PE','PB'],loc1='lower left',loc2='upper right')
+
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    end='2022-11-18'
+    start=date_adjust(end,-365*5)
+    valuation=['PE','PB','dividend']
+    return_delay=['Annual','Quarterly','Monthly']
+    industries='all'
+    
+    lo_est_betas_list,betas_list,idfall=valuation2return_sw(start,end,valuation=valuation,return_delay=return_delay)
+
+def valuation2return_sw2(start,end,valuation=['PE','PB','dividend'], \
+                        return_delay=['Annual','Quarterly','Monthly'], \
+                            industries='all'):
+    """
+    废弃！!!
+    功能：测试估值指标对滞后一段时间收益率的影响。若正向（负向）影响，行业估值未低估（高估）
+    start, end: 测试期间
+    valuation: 估值指标，可为单项指标或列表。市盈率PE, 市净率PB, 股息率dividend
+    return_delay: 滞后时间长度，可为单项指标或列表。Monthly一个月=21天，Quarterly一个季度=63天，Annual一年=252天
+    """
+    # 检查日期的合理性
+
+    # 检查估值指标的类型
+    valuationlist=['pe','pb','dividend']
+    if isinstance(valuation,str):
+        valuation_list=[valuation]
+    elif isinstance(valuation,list):
+        valuation_list=valuation
+    for v in valuation_list:
+        if not (v.lower() in valuationlist):
+            print("  #Warning(valuation2return_sw): unsupported type of valuation:",v)
+            print("  supported types of valuation:",valuationlist)
+            return None
+
+    # 检查估值指标的类型
+    return_delaylist=['annual','quarterly','monthly']
+    measurelist=['Annual Ret','Quarterly Ret','Monthly Ret']
+    shiftlist=[252,63,21]
+    
+    if isinstance(return_delay,str):
+        return_delay_list=[return_delay]
+    elif isinstance(return_delay,list):
+        return_delay_list=return_delay
+        
+    measure_list=[]
+    shift_days_list=[]
+    for v in return_delay_list:
+        if not (v.lower() in return_delaylist):
+            print("  #Warning(valuation2return_sw): unsupported type of return delay:",v)
+            print("  supported types of return delay:",return_delaylist)
+            return None
+        
+        pos=return_delaylist.index(v.lower())
+        measure=measurelist[pos]
+        measure_list=measure_list+[measure]
+        shift_days=shiftlist[pos]
+        shift_days_list=shift_days_list+[shift_days]
+
+    # 获取行业历史数据，本步骤所需时间较长=========================================
+    industry_data=get_industry_sw('I')
+    if not (industries.lower() == 'all'):
+        industry_codes=industry_sw_codes(industries)
+    else:
+        industry_codes=list(set(list(industry_data['ticker'])))
+    
+    # 计算基础数据，本步骤所需时间较长============================================
+    idf,idfall=calc_industry_sw(industry_data,start,end)
+
+    # 构造回归数据，进行回归，记录回归结果
+    import pandas as pd
+    from scipy import stats
+
+    #屏蔽函数内print信息输出的类
+    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
+    
+    betas_list=[]
+    lo_est_betas_list=[]
+    print("Calculating industry valuations, it may take great time, please wait ...")
+    for m in measure_list:
+        
+        #计算收益率
+        ret_df=compare_industry_sw(idfall,industry_codes,measure=m,graph=False)
+        industry_names=list(ret_df)
+        pos=measure_list.index(m)
+        d=shift_days_list[pos]
+    
+        ret_collist=list(ret_df)
+        ret_reg=pd.DataFrame()
+        # 构造用于回归的数据结构    
+        for i in ret_collist:
+            tmpdf=ret_df[[i]]
+            tmpdf['行业']=i
+            tmpdf.rename(columns={i:'ret'},inplace=True)
+            
+            if len(ret_reg)==0:
+                ret_reg=tmpdf
+            else:
+                try:
+                    ret_reg=ret_reg.append(tmpdf)
+                except:
+                    ret_reg=ret_reg._append(tmpdf)
+        
+        # 计算估值指标
+        for val in valuation_list:
+            with HiddenPrints():
+                val_df=compare_industry_valuation_sw(industry_names,start=start,end=end,
+                                                 vtypes=val,graph=False)
+            
+            # 滞后估值指标
+            val_df2=val_df.shift(d)
+            val_collist=list(val_df2)
+            val_reg=pd.DataFrame()
+            
+            for i in val_collist:
+                tmpdf=val_df2[[i]]
+                tmpdf['行业']=i
+                tmpdf.rename(columns={i:val},inplace=True)
+                
+                if len(val_reg)==0:
+                    val_reg=tmpdf
+                else:
+                    try:
+                        val_reg=val_reg.append(tmpdf)
+                    except:
+                        val_reg=val_reg._append(tmpdf)
+        
+            #合成    
+            val_reg['日期']=val_reg.index
+            ret_reg['日期']=ret_reg.index
+            
+            df_reg=val_reg.merge(ret_reg,how='inner',on=['日期','行业'])
+            df_reg.set_index('日期',inplace=True)
+            
+            df_reg.dropna(inplace=True)
+            
+            output=stats.linregress(df_reg[val],df_reg['ret'])
+            (beta,alpha,r_value,p_value,std_err)=output
+            
+            if p_value < 0.001: siglevel='***'
+            elif p_value < 0.01: siglevel='** '
+            elif p_value < 0.05: siglevel='*  '
+            else: siglevel='   '
+            r_sqr=round(r_value**2,4)
+            
+            betas=pd.DataFrame(columns=('估值指标','收益率指标','行业','截距','估值系数','p值','显著性','R-sqr'))
+            row=pd.Series({'估值指标':val,'收益率指标':m,'行业':'全行业','截距':alpha,'估值系数':beta, \
+                           'p值':p_value,'显著性':siglevel,'R-sqr':r_sqr})
+            try:
+                betas=betas.append(row,ignore_index=True)
+            except:
+                betas=betas._append(row,ignore_index=True)
+            
+            industry_list=list(set(list(df_reg['行业']))) 
+            for i in industry_list:
+                dftmp=df_reg[df_reg['行业']==i]
+                
+                output=stats.linregress(dftmp[val],dftmp['ret'])
+                (beta,alpha,r_value,p_value,std_err)=output
+                
+                if p_value < 0.001: siglevel='***'
+                elif p_value < 0.01: siglevel='** '
+                elif p_value < 0.05: siglevel='*  '
+                else: siglevel='   '
+                r_sqr=round(r_value**2,4)
+                
+                row=pd.Series({'估值指标':val,'收益率指标':m,'行业':i,'截距':alpha,'估值系数':beta, \
+                               'p值':p_value,'显著性':siglevel,'R-sqr':r_sqr})
+                try:
+                    betas=betas.append(row,ignore_index=True)
+                except:
+                    betas=betas._append(row,ignore_index=True)
+            
+            if val.lower() in ['pe','pb']:
+                betas['估值判断']=betas['估值系数'].apply(lambda x: '可能高估' if x <0 else '可能低估')
+                betas.sort_values(by=['估值系数','显著性'],ascending=[False,False],inplace=True)
+            elif val.lower() in ['dividend']:
+                betas['估值判断']=betas['估值系数'].apply(lambda x: '可能高估' if x >0 else '可能低估')
+                betas.sort_values(by=['估值系数','显著性'],ascending=[True,False],inplace=True)
+            
+            lo_est_betas=betas[(betas['行业']=='全行业') | (betas['估值判断']=='可能低估')]    
+            
+            betas.reset_index(drop=True,inplace=True)
+            betas.index=betas.index + 1
+            betas_list=betas_list+[betas]
+            lo_est_betas_list=lo_est_betas_list+[lo_est_betas]
+    
+        # 整理各个行业的综合评价
+        # 合并betas
+        allbetas=pd.DataFrame()
+        for b in betas_list:
+            try:
+                allbetas=allbetas.append(b)
+            except:
+                allbetas=allbetas._append(b)
+        
+        valtable=pd.DataFrame(columns=('行业', \
+                                    ('PE','Annual Ret'), \
+                                    ('PE','Quarterly Ret'), \
+                                    ('PE','Monthly Ret'), \
+                                    ('PB','Annual Ret'), \
+                                    ('PB','Quarterly Ret'), \
+                                    ('PB','Monthly Ret'), \
+                                    ('dividend','Annual Ret'), \
+                                    ('dividend','Quarterly Ret'), \
+                                    ('dividend','Monthly Ret'), \
+                                    ))
+        valtable_list=list(valtable)
+        valtable_list.remove('行业')
+        industry_names=list(set(list(allbetas['行业'])))
+        for i in industry_names:
+            pos=industry_names.index(i)
+            row=pd.Series({'行业':i})
+            try:
+                valtable=valtable.append(row,ignore_index=True)
+            except:
+                valtable=valtable._append(row,ignore_index=True)
+            
+            for v in valtable_list:
+                val,ret=v
+                try:
+                    val_value=allbetas[(allbetas['估值指标']==val) & (allbetas['收益率指标']==ret) & (allbetas['行业']==i)]['估值判断'].values[0]
+                    if val_value == '可能低估':
+                        val_value1='低估'
+                    else:
+                        val_value1='高估'
+                        
+                    sig_value=allbetas[(allbetas['估值指标']==val) & (allbetas['收益率指标']==ret) & (allbetas['行业']==i)]['显著性'].values[0]
+                    vsvalue=val_value1+sig_value
+                    
+                    valtable.at[pos,v]=vsvalue
+                except:
+                    continue
+    
+    valtable.fillna('不确定',inplace=True)
+    
+    # 排序，低估在前，PE优先
+    fld1=valtable_list[0]
+    fld2=valtable_list[1]
+    fld3=valtable_list[2]
+    valtable.sort_values(by=[fld1,fld2,fld3],ascending=[True,True,True],inplace=True)
+    valtable.reset_index(drop=True,inplace=True)
+
+    print("Successfully valuated",len(valtable),'industries')
+    print("Valuation completed by mixing PE/PB/dividend with Annual/Quarterly/Monthly")
+    
+    return valtable,betas_list
+
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    end='2022-11-22'
+    start=date_adjust(end,-365*5)
+    itype='I'
+    industries='all'
+    
+    lo_est_betas_list,betas_list,idfall=valuation2return_sw(start,end,valuation=valuation,return_delay=return_delay)
+
+def valuation2return_sw(start,end,itype='1',industries='all'):
+    """
+    功能：测试三种估值指标对滞后一段时间收益率的影响。
+    测试行业哑元变量对估值指标的调节作用，借此判断。若正向（负向）影响，行业估值未低估（高估）
+    start, end: 测试期间
+    itype: 申万指数种类，默认行业类别I, 市场表征F, 投资风格F，全部A
+    industries: 指定具体的指数列表，用于节省处理时间，默认all
+    """
+    #设定相关参数
+    # 估值指标，市盈率PE, 市净率PB, 股息率dividend，三项合用
+    valuation=['PE','PB','dividend']
+    # 估值指标对收益率影响的滞后时间长度
+    # Monthly一个月=21天，Quarterly一个季度=63天，Annual一年=252天。分别回归，用于观察期间长短的影响
+    # 用于判断过去某个时点的估值指标能够对当前的收益率产生影响，以及何种影响
+    return_delay=['Annual','Quarterly','Monthly']
+    
+    # 检查日期的合理性
+    flag,start1,end1=check_period(start,end)
+    if not flag:
+        print("  #Error(valuation2return_sw): invalid date period",start,end)
+        return None
+
+    # 检查估值指标的类型
+    valuationlist=['pe','pb','dividend']
+    if isinstance(valuation,str):
+        valuation_list=[valuation]
+    elif isinstance(valuation,list):
+        valuation_list=valuation
+    for v in valuation_list:
+        if not (v.lower() in valuationlist):
+            print("  #Warning(valuation2return_sw): unsupported type of valuation:",v)
+            print("  supported types of valuation:",valuationlist)
+            return None
+
+    # 检查估值指标的类型
+    return_delaylist=['annual','quarterly','monthly']
+    measurelist=['Annual Ret','Quarterly Ret','Monthly Ret']
+    shiftlist=[252,63,21]
+    
+    if isinstance(return_delay,str):
+        return_delay_list=[return_delay]
+    elif isinstance(return_delay,list):
+        return_delay_list=return_delay
+        
+    measure_list=[]
+    shift_days_list=[]
+    for v in return_delay_list:
+        if not (v.lower() in return_delaylist):
+            print("  #Warning(valuation2return_sw): unsupported type of return delay:",v)
+            print("  supported types of return delay:",return_delaylist)
+            return None
+        
+        pos=return_delaylist.index(v.lower())
+        measure=measurelist[pos]
+        measure_list=measure_list+[measure]
+        shift_days=shiftlist[pos]
+        shift_days_list=shift_days_list+[shift_days]
+
+    #屏蔽函数内print信息输出的类
+    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
+
+    # 步骤1：获取行业历史数据，本步骤所需时间较长==================================
+    print("Step1: retrieving industry information, it may take up to hours ...")
+    industry_data=get_industry_sw(itype=itype)
+    if not (industries.lower() == 'all'):
+        industry_codes=industry_sw_codes(industries)
+    else:
+        industry_codes=list(set(list(industry_data['ticker'])))
+    
+    # 步骤2：计算基础数据，本步骤所需时间较长======================================
+    print("Step2: Calculating industry valuations, it may take great time ...")
+    idf,idfall=calc_industry_sw(industry_data,start,end)
+
+    # 步骤3：构造回归数据，进行回归，记录回归结果
+    import pandas as pd
+    coefdflist=[]
+    print("Step3: Analyzing industry performance, it may need quite some time ...")
+    
+    total=len(measure_list)*len(valuation_list)
+    for m in measure_list:
+        
+        # （1）计算收益率
+        #print("  Processing measure",m)
+        ret_df=compare_industry_sw(idfall,industry_codes,measure=m,graph=False)
+        industry_names=list(ret_df)
+        pos=measure_list.index(m)
+        d=shift_days_list[pos]
+    
+        ret_collist=list(ret_df)
+        ret_reg=pd.DataFrame()
+        obs_num=0
+        # 构造用于回归的数据结构    
+        for i in ret_collist:
+            tmpdf=ret_df[[i]]
+            tmpdf['行业']=i
+            tmpdf.rename(columns={i:'ret'},inplace=True)
+            
+            if len(ret_reg)==0:
+                ret_reg=tmpdf
+            else:
+                try:
+                    ret_reg=ret_reg.append(tmpdf)
+                except:
+                    ret_reg=ret_reg._append(tmpdf)
+        
+        # （2）处理估值指标
+        df=pd.DataFrame()
+        for val in valuation_list:
+            
+            #print("    Handling valuation",val,'with measure',m)
+            
+            # （a）计算估值指标
+            with HiddenPrints():
+                val_df=compare_industry_valuation_sw(industry_names,start=start,end=end,
+                                                 vtypes=val,graph=False)
+            # （b）滞后估值指标
+            val_df2=val_df.shift(d)
+            val_collist=list(val_df2)
+            val_reg=pd.DataFrame()
+            
+            for i in val_collist:
+                tmpdf=val_df2[[i]]
+                tmpdf['行业']=i
+                tmpdf.rename(columns={i:val},inplace=True)
+                
+                if len(val_reg)==0:
+                    val_reg=tmpdf
+                else:
+                    try:
+                        val_reg=val_reg.append(tmpdf)
+                    except:
+                        val_reg=val_reg._append(tmpdf)
+        
+            # （c）合成滞后的估值指标和收益率    
+            val_reg['日期']=val_reg.index
+            ret_reg['日期']=ret_reg.index
+            
+            df_reg=val_reg.merge(ret_reg,how='inner',on=['日期','行业'])
+            df_reg.set_index('日期',inplace=True)
+            df_reg2=index2date(df_reg,date_field='date')
+            
+            df_reg2.dropna(inplace=True)
+            
+            if len(df)==0:
+                df=df_reg2
+            else:
+                df=pd.merge(df,df_reg2,how='outer',on=['date','行业'])
+
+            # 本步骤所需时间漫长，显示当前进度
+            current=measure_list.index(m)*len(valuation_list)+valuation_list.index(val)
+            print_progress_percent(current,total,steps=5,leading_blanks=2)
+            
+        # （3）增加额外的自变量：PB/PE表示净资产产生利润的能力
+        #print("Model regression for the impact on industrial returns by prior valuation ...")
+        # PE*PB表示利润与净资产对于股价的联合支撑作用
+        df['PBdivPE']=df['PB']/df['PE']
+        df['PBxPE']=df['PB']*df['PE']
+        xList=['PE','PB','dividend','PBdivPE','PBxPE']
+        
+        # （4）构造哑元变量：行业，年度
+        df['year']=df['date'].apply(lambda x:x[:4])
+        df2,indDummies,yDummies=df_fe2(df,industry_field="行业",year_field='year')
+        obs_num=obs_num+len(df2)
+        
+        # （5）多元回归，记录行业哑元变量的回归系数和显著性
+        coefdf=multi_ols(df2,xList=xList,y='ret',industryDummies=indDummies,yearDummies=yDummies)
+        coefdf2=coefdf.T[indDummies].T
+        coefdf2['val']=coefdf2['coef'].apply(lambda x: '低估' if x >0 else '高估')                
+        coefdf2['估值判断']=coefdf2['val']+coefdf2['sig']
+        coefdf2['measure']=m                
+            
+        coefdflist=coefdflist+[coefdf2]    
+    
+    # 步骤4：制作多行业多期的估值评价矩阵
+    allbetas=pd.DataFrame()
+    for b in coefdflist:
+        b['行业0']=b.index
+        b['行业']=b['行业0'].apply(lambda x:x[1:])
+        if b['measure'].values[0]=='Annual Ret':
+            b.rename(columns={'估值判断': '年度估值判断'}, inplace=True)
+            b2=b[['行业','年度估值判断']]
+        elif b['measure'].values[0]=='Quarterly Ret':
+            b.rename(columns={'估值判断': '季度估值判断'},inplace=True)
+            b2=b[['行业','季度估值判断']]
+        else:
+            b.rename(columns={'估值判断': '月度估值判断'},inplace=True)
+            b2=b[['行业', '月度估值判断']]
+            
+        if len(allbetas)==0:
+            allbetas=b2
+        else:
+            allbetas=pd.merge(allbetas,b2,how='outer',on=['行业'])
+        
+    allbetas.fillna('未知',inplace=True)
+    
+    allbetas['score1']=allbetas['年度估值判断'].apply(lambda x: val_score(x))
+    allbetas['score2']=allbetas['季度估值判断'].apply(lambda x: val_score(x))
+    allbetas['score3']=allbetas['月度估值判断'].apply(lambda x: val_score(x))
+    allbetas['score']=allbetas['score1']+allbetas['score2']+allbetas['score3']
+    
+    loest=len(allbetas[allbetas['score']<0])
+    hiest=len(allbetas[allbetas['score']>0])
+    
+    allbetas.sort_values(by=['score','行业'],ascending=[True,True],inplace=True)
+    
+    allbetas.reset_index(drop=True,inplace=True)
+    allbetas2=allbetas[['行业', '年度估值判断', '季度估值判断', '月度估值判断']]
+    allbetas2.index=allbetas2.index+1
+
+    print("\nResults:")
+    modelstr='ret = lagged('+xList[0]
+    for x in xList[1:]:
+        modelstr=modelstr+' + '+x
+    modelstr=modelstr+')'+' + '+'Industry/Year dummies'
+    print("  Valuation model:",modelstr)
+    print("  Depenbdent: using Annual/Quarterly/Monthly ret respectively")
+    
+    print("  Sample period:",start,'to',end,'\b, total',obs_num,'observations for regression')
+    
+    print(" ",len(allbetas2),'industries valuated,',str(loest)+'('+str(hiest)+') might be under(over) estimated from future return perspective')
+    
+    # 在Jupyter Notebook可直接显示返回的变量，格式整齐
+    return allbetas2
+
+#==============================================================================
+def val_score(val_comment):
+    """
+    功能：基于估值判断给出评分，仅用于排序，无其他实际意义
+    """
+    #print(val_comment,len(val_comment))
+    if val_comment == '未知':
+        score=0
+        
+    if val_comment == '低估   ':
+        score=-1
+    if val_comment == '低估*  ':
+        score=-2
+    if val_comment == '低估** ':
+        score=-3
+    if val_comment == '低估***':
+        score=-4
+        
+    if val_comment == '高估   ':
+        score=1
+    if val_comment == '高估*  ':
+        score=2
+    if val_comment == '高估** ':
+        score=3
+    if val_comment == '高估***':
+        score=4
+        
+    return score
+        
+        
+        
+        
+#==============================================================================
+if __name__ == '__main__':
+    date_field='date'
+    
+def index2date(df,date_field='date'):
+    """
+    功能：从日期型df.index取出日期，类型YYYY-MM-DD，放在新字段date_field中。
+    """
+    
+    df[date_field+'pd']=df.index
+    df[date_field]=df[date_field+'pd'].apply(lambda x: x.strftime("%Y-%m-%d"))
+    del df[date_field+'pd']
+
+    return df    
+    
+#==============================================================================
+
+
+
+def df_fe2(df,industry_field,year_field):
+    """
+    功能：基于df做出industry_field和year_field的哑元变量'i'+industry_field和'y'+year_field
+    
+    """
+    ilist=[]
+    ylist=[]
+    
+    #生成行业哑元变量，全部预置为0和1
+    industry_list=list(set(list(df[industry_field])))
+    industry_list.sort(reverse=False)
+    for i in industry_list:
+        df['i'+i]=df[industry_field].apply(lambda x: 1 if x==i else 0)
+        ilist=ilist+['i'+i]
+
+    #生成年度哑元变量，全部预置为0和1
+    year_list=list(set(list(df[year_field])))
+    year_list.sort(reverse=False)
+    for i in year_list:
+        df['y'+i]=df[year_field].apply(lambda x: 1 if x==i else 0)
+        ylist=ylist+['y'+i]
+    
+    return df,ilist,ylist
+
+#==============================================================================
+def sig_level(p):
+    """
+    功能：基于p值给出显著性星星个数
+    
+    """
+    if p >=0.05:
+        sig="   "
+    elif 0.05 > p >= 0.01:
+        sig='*  '
+    elif 0.01 > p >= 0.001:
+        sig="** "
+    else:
+        sig="***"
+        
+    return sig
+
+#==============================================================================
+
+def multi_ols(df,xList,y,industryDummies,yearDummies):
+    """
+    功能：多元线性回归, y=f(X)，需要系数和显著性
+    df: 所有数据
+    xList: 自变量列表，不包括行业和年度哑元变量
+    y： 因变量
+    industryDummies： 行业哑元变量列表
+    yearDummies： 年度哑元变量列表
+    """
+    import statsmodels.formula.api as smf
+    
+    #构造模型表达式：y~x1+x2+.....
+    model=''
+    allXVars=xList+industryDummies+yearDummies
+    for x in allXVars:
+        if model == '':
+            model=y+'~'+x
+        else:
+            model=model+'+'+x
+    
+    #形成smf模型
+    reg = smf.ols(formula=model,data=df)    
+    
+    #线性回归
+    result=reg.fit()   
+    #print(result.summary())
+
+    #自变量各个系数及其显著性
+    import pandas as pd
+    coefMatrix=pd.DataFrame([result.params,result.pvalues],index=["coef","p"]).T
+    
+    coefMatrix['sig']=coefMatrix['p'].apply(lambda x: sig_level(x))
+    
+    #整个模型的显著性：F-test
+    modelSig=result.f_pvalue
+    
+    return coefMatrix
+
+#==============================================================================
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    sw_code='850831.SW'
+    sw_code='801193.SW'
+    indicator='PE'
+    start='2023-1-1'
+    end='2023-12-15'
+    top=10
+
+def valuation_industry_sw_generating(sw_code,indicator,start,end,top=5):
+    """
+    功能：模拟申万行业指数的估值，PE/PB/股息率等
+    sw_code：申万行业分类指数，各个级别
+    start/end：开始/结束日期
+    top：使用前几大成分股的估值进行合成
+    
+    注意：指数模拟出的估值曲线波动过大，缺乏实用价值！
+    """
+    import pandas as pd
+    #查找申万行业指数成分股
+    clist,cdf=industry_stock_sw(industry=sw_code,top=top)
+    
+    #查找成分股的历史估值
+    df=None
+    for t in clist:
+        dft=get_stock_valuation_cn_hk(ticker=t,indicators=indicator,start=start,end=end)
+        dft[t]=dft[indicator]
+        dft2=dft[[t]]
+        if dft2 is None: continue
+    
+        #将负数填充为0，不计入估值?整个成分股剔除？
+        dft2[t]=dft2[t].apply(lambda x: 0 if x<0 else x)
+    
+        if df is None:
+            df=dft2
+        else:
+            df=pd.merge(df,dft2,how='outer',left_index=True,right_index=True)
+    
+    #成分股权重
+    weight=list(cdf['最新权重'])
+                
+    #各行权重分别求和
+    dfw=df.copy()
+    collist=list(dfw)   
+    for c in collist:
+        dfw[c]=dfw[c].apply(lambda x: 0 if x<=0 else 1)
+    dfw['weight']=dfw.dot(weight)
+    dfw2=dfw[['weight']]
+    
+    #加权平均
+    df['weighted_total']=df.dot(weight)
+    df2=pd.merge(df,dfw2,left_index=True,right_index=True)
+    
+    df2['weighted_avg']=df2['weighted_total']/df2['weight']
+    df2['code']=sw_code
+    df3=df2[['code','weighted_avg']]
+    
+    #因有市盈率负数，不管如何处理都会导致加权平均后数值波动过大，不能实用
+    return df3
+    
+    
+    
+    
+
+#==============================================================================