siat 2.13.43__py3-none-any.whl → 2.14.1__py3-none-any.whl

siat/grafix.py

@@ -30,10 +30,10 @@ plt.rcParams['xtick.labelsize']=11 #横轴字体大小
 plt.rcParams['ytick.labelsize']=11 #纵轴字体大小
 
 title_txt_size=16
-ylabel_txt_size=14
-xlabel_txt_size=14
-legend_txt_size=14
-annotate_size=9
+ylabel_txt_size=12
+xlabel_txt_size=12
+legend_txt_size=12
+annotate_size=11
 
 #设置绘图风格：网格虚线
 plt.rcParams['axes.grid']=False
@@ -67,15 +67,39 @@ if czxt in ['linux']: #website Jupyter
 # 解决保存图像时'-'显示为方块的问题
 plt.rcParams['axes.unicode_minus'] = False 
 #==============================================================================
+if __name__ =="__main__":
+    df0=get_price('000001.SS','2023-1-1','2024-3-22')
+    
+    colname='Close'
+    collabel='Close'
+    ylabeltxt='Close'
+    titletxt='Title'
+    footnote='footnote'
+    datatag=False
+    power=0
+    zeroline=False
+    average_value=False
+    resample_freq='H'
+    loc='best'
+    date_range=False
+    date_freq=False
+    date_fmt='%Y-%m-%d'
+    mark_top=1
+    mark_bottom=1
+    
+    plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,mark_top=1,mark_bottom=1)
+
 def plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,datatag=False, \
               power=0,zeroline=False,average_value=False, \
               resample_freq='H',loc='best', \
-              date_range=False,date_freq=False,date_fmt='%Y-%m-%d'):
+              date_range=False,date_freq=False,date_fmt='%Y-%m-%d', \
+              mark_top=0,mark_bottom=0,mark_end=True):
     """
     功能：绘制折线图。如果power=0不绘制趋势图，否则绘制多项式趋势图
     假定：数据表有索引，且已经按照索引排序
     输入：数据表df，数据表中的列名colname，列名的标签collabel；y轴标签ylabeltxt；
     标题titletxt，脚注footnote；是否在图中标记数据datatag；趋势图的多项式次数power
+    mark_top,mark_bottom：标记最高最低点的个数
     输出：折线图
     返回值：无
     注意：需要日期类型作为df索引
@@ -118,9 +142,10 @@ def plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,datatag=False, \
         collabel=''
         if ylabeltxt == "stooq_MB":
             ylabeltxt=''
-            
+
+    lwadjust=linewidth_adjust(df)
     plt.plot(df.index,df[colname],'-',label=collabel, \
-             linestyle='-',color='blue', linewidth=2)
+             linestyle='-',color='blue', linewidth=lwadjust)
         
     haveLegend=True
     if collabel == '':
@@ -129,22 +154,46 @@ def plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,datatag=False, \
     #绘制数据标签
     if datatag:
         for x, y in zip(df.index, df[colname]):
-            plt.text(x,y+0.1,'%.2f' % y,ha='center',va='bottom',color='black')        
+            plt.text(x,y*1.001,'%.2f' % y,ha='center',va='bottom',color='black')        
+
+    #标记最高点/最低点
+    if mark_top > 0 or mark_bottom > 0:
+        df_mark=df[[colname]].copy() #避免影响原df
+        df_mark.sort_values(by=colname,ascending=False,inplace=True)
+        
+        high_poit=df_mark[colname].head(1).values[0]
+        low_poit=df_mark[colname].tail(1).values[0]
+        high_low=high_poit - low_poit
+        if mark_top > 0:
+            df_mark_top=df_mark[:mark_top]
+            for x, y in zip(df_mark_top.index, df_mark_top[colname]):
+                #plt.text(x,y+0.1,'%.2f' % y,ha='center',va='bottom',color='red')
+                y1=round(y+high_low*0.01,2)
+                s='%.0f' if y >= 100 else '%.2f'
+                plt.text(x,y1,s % y,ha='center',va='bottom',color='red')
+                plt.scatter(x,y, color='red',marker='8',s=70)
+            
+        if mark_bottom > 0:
+            df_mark_bottom=df_mark[-mark_bottom:]
+            for x, y in zip(df_mark_bottom.index, df_mark_bottom[colname]):
+                #plt.text(x,y-0.1,'%.2f' % y,ha='center',va='bottom',color='black')  
+                y1=round(y-high_low*0.055,2) #标记位置对应y1的底部
+                s='%.0f' if y >= 100 else '%.2f'
+                plt.text(x,y1,s % y,ha='center',va='bottom',color='seagreen')
+                plt.scatter(x,y, color='seagreen',marker='8',s=70)
+
+    #标记曲线末端数值
+    if mark_end:
+        df_end=df.tail(1)
+        y_end = df_end[colname].min()    # 末端的y坐标
+        x_end = df_end[colname].idxmin() # 末端值的x坐标 
+        
+        y1=str(int(y_end)) if y_end >= 100 else str(round(y_end,2))
+        plt.annotate(text=' '+y1, 
+                     xy=(x_end, y_end),
+                     xytext=(x_end, y_end),fontsize=annotate_size)
     
     #是否绘制水平线
-    #if zeroline and (min(df[colname]) < 0):
-    """
-    if zeroline:
-        if isinstance(zeroline,bool):
-            hline=0
-        elif isinstance(zeroline,float):
-            hline=zeroline
-        elif isinstance(zeroline,int):
-            hline=zeroline
-        else:
-            hline=0
-        plt.axhline(y=hline,ls=":",c="black",linewidth=2)
-    """
     if isinstance(zeroline,bool):#若zeroline为True
         if zeroline: 
             hline=0
@@ -171,7 +220,6 @@ def plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,datatag=False, \
         #trend_txt=''
         if lang == 'English':
             trend_txt='Trend line'           
-        
         try:
             #生成行号，借此将横轴的日期数量化，以便拟合
             df['id']=range(len(df))
@@ -194,7 +242,6 @@ def plot_line(df0,colname,collabel,ylabeltxt,titletxt,footnote,datatag=False, \
     plt.ylabel(ylabeltxt,fontsize=ylabel_txt_size)
     plt.xlabel(footnote,fontsize=xlabel_txt_size)
     plt.title(titletxt,fontweight='bold',fontsize=title_txt_size)
-
     
     plt.show()
     plt.close()
@@ -355,8 +402,9 @@ def plot_line2_coaxial(df01,ticker1,colname1,label1, \
         else:
             label1txt=codetranslate(ticker1)+'('+label1+')'    
     
+    lwadjust=linewidth_adjust(df1)
     plt.plot(df1.index,df1[colname1],'-',label=label1txt, \
-             linestyle='-',linewidth=2.5,color=color1)
+             linestyle='-',linewidth=lwadjust,color=color1)
     #证券1：绘制数据标签
     if datatag1:
         for x, y in zip(df1.index, df1[colname1]):
@@ -408,8 +456,9 @@ def plot_line2_coaxial(df01,ticker1,colname1,label1, \
         else:
             label2txt=codetranslate(ticker2)+'('+label2+')'    
     
+    lwadjust=linewidth_adjust(df2)
     plt.plot(df2.index,df2[colname2],'-',label=label2txt, \
-             linestyle='-.',linewidth=2.5,color=color2)
+             linestyle='-.',linewidth=lwadjust,color=color2)
     #证券2：绘制数据标签
     if datatag2:
         for x, y in zip(df2.index, df2[colname2]):
@@ -526,8 +575,9 @@ def plot_line2_coaxial2(df01,ticker1,colname1,label1, \
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))    
 
+    lwadjust=linewidth_adjust(df1)
     plt.plot(df1.index,df1[colname1],'-',label=label1txt, \
-             linestyle='-',linewidth=2.5,color=color1)
+             linestyle='-',linewidth=lwadjust,color=color1)
     #证券1：绘制数据标签
     if datatag1:
         for x, y in zip(df1.index, df1[colname1]):
@@ -589,8 +639,9 @@ def plot_line2_coaxial2(df01,ticker1,colname1,label1, \
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))
 
+    lwadjust=linewidth_adjust(df2)
     plt.plot(df2.index,df2[colname2],'-',label=label2txt, \
-             linestyle='-.',linewidth=2.5,color=color2)
+             linestyle='-.',linewidth=lwadjust,color=color2)
     #证券2：绘制数据标签
     if datatag2:
         for x, y in zip(df2.index, df2[colname2]):
@@ -673,10 +724,11 @@ def plot_line2_twinx(df01,ticker1,colname1,label1, \
         if label1 == '':
             label1txt=codetranslate(ticker1)
         else:
-            label1txt=codetranslate(ticker1)+'('+label1+')'    
-    
+            label1txt=codetranslate(ticker1)+'('+label1+')'   
+            
+    lwadjust=linewidth_adjust(df1)
     ax.plot(df1.index,df1[colname1],'-',label=label1txt, \
-             linestyle='-',color=color1,linewidth=2)   
+             linestyle='-',color=color1,linewidth=lwadjust)   
     #证券1：绘制数据标签
     if datatag1:
         for x, y in zip(df1.index, df1[colname1]):
@@ -712,10 +764,11 @@ def plot_line2_twinx(df01,ticker1,colname1,label1, \
         if label2 == '':
             label2txt=codetranslate(ticker2)
         else:
-            label2txt=codetranslate(ticker2)+'('+label2+')'    
-    
+            label2txt=codetranslate(ticker2)+'('+label2+')'   
+            
+    lwadjust=linewidth_adjust(df2)
     ax2.plot(df2.index,df2[colname2],'-',label=label2txt, \
-             linestyle='-.',color=color2,linewidth=2.5)
+             linestyle='-.',color=color2,linewidth=lwadjust)
     #证券2：绘制数据标签
     if datatag2:
         for x, y in zip(df2.index, df2[colname2]):
@@ -844,9 +897,10 @@ def plot_line2_twinx2(df01,ticker1,colname1,label1, \
     if date_range and date_freq:
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))    
-        
+    
+    lwadjust=linewidth_adjust(df1)        
     ax.plot(df1.index,df1[colname1],'-',label=label1txt, \
-             linestyle='-',color=color1,linewidth=2)   
+             linestyle='-',color=color1,linewidth=lwadjust)   
     #证券1：绘制数据标签
     if datatag1:
         for x, y in zip(df1.index, df1[colname1]):
@@ -895,9 +949,10 @@ def plot_line2_twinx2(df01,ticker1,colname1,label1, \
     if date_range and date_freq:
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))        
-        
+    
+    lwadjust=linewidth_adjust(df2)        
     ax2.plot(df2.index,df2[colname2],'-',label=label2txt, \
-             linestyle='-.',color=color2,linewidth=2.5)
+             linestyle='-.',color=color2,linewidth=lwadjust)
     #证券2：绘制数据标签
     if datatag2:
         for x, y in zip(df2.index, df2[colname2]):
@@ -922,6 +977,7 @@ def plot_line2_twinx2(df01,ticker1,colname1,label1, \
             label2txt=''
         else:
             label2txt=codetranslate(ticker2)+"("+trend_txt+")"
+        
         ax2.plot(df2.index, f(df2.id),"c--", label=label2txt,linewidth=1)        
         
     ax.set_xlabel(footnote,fontsize=xlabel_txt_size)
@@ -959,7 +1015,8 @@ def plot_line2_twinx2(df01,ticker1,colname1,label1, \
 #==============================================================================
 def draw_lines(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
                data_label=True,resample_freq='H',smooth=True,linewidth=1.5, \
-               loc='best',annotate=False,annotate_value=False,plus_sign=False):
+               loc='best',annotate=False,annotate_value=False,plus_sign=False, \
+               mark_top=0,mark_bottom=0,mark_end=False):
     """
     函数功能：根据df的内容绘制折线图
     输入参数：
@@ -1003,21 +1060,14 @@ def draw_lines(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
     
     # 所有字段转换为数值类型，以防万一
     for c in collist:
-        df[c]=df[c].astype('float')
+        try:
+            df[c]=df[c].astype('float')
+        except:
+            del df[c]
     
-    """
-    # 计算所有列中的最大最小差距
-    dfmax=0; dfmin=0
-    for c in collist:
-        cmax=df[c].max()
-        if cmax > dfmax:
-            dfmax=cmax
-        
-        cmin=df[c].min()
-        if cmin < dfmin:
-            dfmin=cmin
-    dfspread=(dfmax - dfmin)/10.0
-    """
+    # 计算所有列中的最大最小差距，所有列必须为数值型！
+    dfmax=df.max().max(); dfmin=df.min().min()
+    high_low=dfmax - dfmin
     
     # 将末端值最大的排在第一列，优先绘图
     dftt=df.T
@@ -1052,6 +1102,7 @@ def draw_lines(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
             continue
         
         #plt.plot(dfg,label=c,linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
+        lwadjust=linewidth_adjust(dfg)
         if not annotate:
             """
             注意：许多传入的df字段名已经不是证券代码，此处调用codetranslate将会导致
@@ -1060,36 +1111,20 @@ def draw_lines(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
             本函数仅负责绘图，不负责翻译证券名称。
             """
             #plt.plot(dfg,label=codetranslate(c),linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
-            plt.plot(dfg,label=c,linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
+            plt.plot(dfg,label=c,linewidth=lwadjust,ls=lsc,marker=mkc,markersize=3)
         else:
             #plt.plot(dfg[c],label=codetranslate(c),linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
             #plt.plot(dfg,label=codetranslate(c),linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
-            plt.plot(dfg,label=c,linewidth=linewidth,ls=lsc,marker=mkc,markersize=3)
+            plt.plot(dfg,label=c,linewidth=lwadjust,ls=lsc,marker=mkc,markersize=3)
             
             df_end=dfg.tail(1)
             # df_end[c]必须为数值类型，否则可能出错
             y_end = df_end[c].min()    # 末端的y坐标
             x_end = df_end[c].idxmin() # 末端值的x坐标 
             
-            """
-            if firstline:
-                # 直接绘图
-                firstline=False
-                y_prev=y_end
-            else:
-                distance=y_prev - y_end
-                if distance < dfspread:
-                    y_end=y_end - dfspread
-                    y_prev=y_end
-            y_end_list=y_end_list+[y_end]
-            """
-            """
-            plt.annotate(text=codetranslate(c)+':'+str(round(y_end,2)), 
-                                 xy=(x_end, y_end),
-                                 xytext=(x_end, y_end),fontsize=9)   
-            """
             if annotate_value: #在标记曲线名称的同时标记其末端数值
-                plt.annotate(text=c+':'+str(round(y_end,2)), 
+                y1=str(int(y_end)) if y_end >= 100 else str(round(y_end,2))
+                plt.annotate(text=c+':'+y1, 
                                  xy=(x_end, y_end),
                                  xytext=(x_end, y_end),fontsize=annotate_size)
             else:
@@ -1101,10 +1136,44 @@ def draw_lines(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
         #plt.plot(df[c],label=c,linewidth=1.5,marker=mkc,markersize=3)
         #为折线加数据标签
         if data_label==True:
-            for a,b in zip(df2.index,df2[c]):
+            for a,b in zip(dfg.index,df2[c]):
                 plt.text(a,b+0.02,str(round(b,2)), \
                          ha='center',va='bottom',fontsize=7)
-    
+
+        #标记最高点/最低点
+        if mark_top > 0 or mark_bottom > 0:
+            df_mark=dfg[[c]].copy() #避免影响原df
+            df_mark.dropna(inplace=True)
+            df_mark.sort_values(by=c,ascending=False,inplace=True)
+            
+            if mark_top > 0:
+                df_mark_top=df_mark[:mark_top]
+                for x, y in zip(df_mark_top.index, df_mark_top[c]):
+                    y1=round(y+high_low*0.01,2)
+                    s='%.0f' if y >= 100 else '%.2f'
+                    plt.text(x,y1,s % y,ha='center',va='bottom',color='red')
+                    plt.scatter(x,y, color='red',marker='8',s=70)
+                
+            if mark_bottom > 0:
+                df_mark_bottom=df_mark[-mark_bottom:]
+                for x, y in zip(df_mark_bottom.index, df_mark_bottom[c]):
+                    y1=round(y-high_low*0.055,2) #标记位置对应y1的底部
+                    s='%.0f' if y >= 100 else '%.2f'
+                    plt.text(x,y1,s % y,ha='center',va='bottom',color='seagreen')
+                    plt.scatter(x,y, color='seagreen',marker='8',s=70)
+
+        #标记曲线末端数值
+        if mark_end:
+            df_end=dfg.tail(1)
+            y_end = df_end[c].min()    # 末端的y坐标
+            x_end = df_end[c].idxmin() # 末端值的x坐标 
+            
+            y1=str(int(y_end)) if y_end >= 100 else str(round(y_end,2))
+            plt.annotate(text=' '+y1, 
+                         xy=(x_end, y_end),
+                         xytext=(x_end, y_end),fontsize=annotate_size)
+
+
     #绘制水平辅助线
     if axhline_label !="":
         if '零线' in axhline_label:
@@ -1148,7 +1217,8 @@ def draw_lines2(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
                 data_label=False,resample_freq='6H',smooth=True, \
                 date_range=False,date_freq=False,date_fmt='%Y-%m-%d', \
                 colorlist=[],lslist=[],lwlist=[], \
-                band_area='',loc='best',annotate=False,annotate_value=False):
+                band_area='',loc='best',annotate=False,annotate_value=False, \
+                mark_top=0,mark_bottom=0,mark_end=False):
     """
     函数功能：根据df的内容绘制折线图
     输入参数：
@@ -1184,6 +1254,17 @@ def draw_lines2(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
     else:
         df=df0
 
+    # 所有字段转换为数值类型，以防万一
+    for c in list(df):
+        try:
+            df[c]=df[c].astype('float')
+        except:
+            del df[c]
+
+   # 计算所有列中的最大最小差距，假设所有列均为数值型！
+    dfmax=df.max().max(); dfmin=df.min().min()
+    high_low=dfmax - dfmin
+
     #定义横轴标签：显示完整开始、结束日期
     import matplotlib.dates as mdate
     ax=plt.gca()    
@@ -1201,6 +1282,7 @@ def draw_lines2(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
         
     #取得df字段名列表
     collist=df.columns.values.tolist()  
+    collist3=collist[:3] #专用于绘制布林带，取前3个字段
     
     #绘制折线图    
     for c in collist:
@@ -1210,7 +1292,8 @@ def draw_lines2(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
             lls=lslist[pos]
             llw=lwlist[pos]
         except:
-            plt.plot(df[c],label=c,linewidth=2)
+            lwadjust=linewidth_adjust(df)
+            plt.plot(df[c],label=c,linewidth=lwadjust)
         else:
             plt.plot(df[c],label=c,linewidth=llw,ls=lls,color=lcolor)
         
@@ -1219,23 +1302,56 @@ def draw_lines2(df0,y_label,x_label,axhline_value,axhline_label,title_txt, \
             for a,b in zip(df.index,df[c]):
                 plt.text(a,b+0.02,str(round(b,2)), \
                          ha='center',va='bottom',fontsize=7)
-    
-        #曲线末端标记
+
+        #标记最高点/最低点
+        if mark_top > 0 or mark_bottom > 0:
+            df_mark=df[[c]].copy() #避免影响原df
+            df_mark.dropna(inplace=True)
+            df_mark.sort_values(by=c,ascending=False,inplace=True)
+            
+            if mark_top > 0:
+                df_mark_top=df_mark[:mark_top]
+                for x, y in zip(df_mark_top.index, df_mark_top[c]):
+                    y1=round(y+high_low*0.01,2)
+                    s='%.0f' if y >= 100 else '%.2f'
+                    plt.text(x,y1,s % y,ha='center',va='bottom',color='red')
+                    plt.scatter(x,y, color='red',marker='8',s=70)
+                
+            if mark_bottom > 0:
+                df_mark_bottom=df_mark[-mark_bottom:]
+                for x, y in zip(df_mark_bottom.index, df_mark_bottom[c]):
+                    y1=round(y-high_low*0.055,2) #标记位置对应y1的底部
+                    s='%.0f' if y >= 100 else '%.2f'
+                    plt.text(x,y1,s % y,ha='center',va='bottom',color='seagreen')
+                    plt.scatter(x,y, color='seagreen',marker='8',s=70)
+    
+        #曲线末端标记：不建议用于布林带
         if annotate:
             df_end=df.tail(1)
             y_end = df_end[c].min()    # 末端的y坐标
             x_end = df_end[c].idxmin() # 末端值的x坐标 
             
             if annotate_value: #在标记曲线名称的同时标记其末端数值
-                plt.annotate(text=c+':'+str(round(y_end,2)), 
+                y1=str(int(y_end)) if y_end >= 100 else str(round(y_end,2))
+                plt.annotate(text=c+':'+y1, 
                              xy=(x_end, y_end),
                              xytext=(x_end, y_end),fontsize=annotate_size)
             else:
                 plt.annotate(text=c, 
                              xy=(x_end, y_end),
                              xytext=(x_end, y_end),fontsize=annotate_size)
-            
     
+        #处理布林带的mark_end，仅标记上中下线
+        if mark_end & (c in collist3):
+            df_end=df.tail(1)
+            y_end = df_end[c].min()    # 末端的y坐标
+            x_end = df_end[c].idxmin() # 末端值的x坐标 
+            
+            y1=str(int(y_end)) if y_end >= 100 else str(round(y_end,2))
+            plt.annotate(text=y1, 
+                         xy=(x_end, y_end),
+                         xytext=(x_end, y_end),fontsize=annotate_size)
+        
     #绘制带状区域
     if band_area != '' and isinstance(band_area,list) and len(band_area)>=2:
        plt.fill_between(df.index,df[band_area[0]],df[band_area[1]],alpha=0.5,label='') 
@@ -1445,7 +1561,8 @@ def plot_2lines(df01,colname1,label1, \
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))
     
-    plt.plot(df1.index,df1[colname1],label=label1,linestyle='-',linewidth=2)
+    lwadjust=linewidth_adjust(df1)
+    plt.plot(df1.index,df1[colname1],label=label1,linestyle='-',linewidth=lwadjust)
     
     #证券2：先绘制折线图
     date_start=df2.index[0]
@@ -1462,8 +1579,9 @@ def plot_2lines(df01,colname1,label1, \
         ax=plt.gca()
         ax.xaxis.set_major_formatter(mdate.DateFormatter(date_fmt))
         plt.xticks(pd.date_range(date_start,date_end,freq=date_freq))
-        
-    plt.plot(df2.index,df2[colname2],label=label2,linestyle='-.',linewidth=2.5)
+    
+    lwadjust=linewidth_adjust(df2)        
+    plt.plot(df2.index,df2[colname2],label=label2,linestyle='-.',linewidth=lwadjust)
     
     #是否绘制水平虚线
     if not (hline == 0):
@@ -1535,7 +1653,7 @@ def df_smooth(df):
         #print('  #Warning(df_smooth): resampling failed for frequency',freqlist)
         dfh=df
     
-    #插值
+    #重采样后插值
     methodlist=['pchip','nearest','cubic','quadratic','slinear','linear','zero','time','index', \
             'piecewise_polynomial','akima','from_derivatives','spline','polynomial']
     methodlist_order=['spline','polynomial']
@@ -1562,19 +1680,19 @@ def df_smooth(df):
     
         
 #==============================================================================
-def df_smooth_manual(df,method='pchip',resample_freq='H',order=3):
+def df_smooth_manual(df,method='linear',resample_freq='H',order=3):
     """
     功能：对df中的第一个数值列样本进行插值，以便绘制的折线图相对平滑。
     要求：df的索引为pandas的datetime日期型
     注意1：如果样本数量较多，例如多于100个，平滑效果不明显。
     注意2：order阶数仅对'spline'和'polynomial'方法有效，其中'polynomial'方法的阶数只能为奇数。
-    注意3：pchip方法经常失败，改为cubic
+    注意3：pchip方法经常失败，可改为cubic
     """
     
     #如果样本个数多于100个，没必要进行平滑，完全看不出效果
     if len(df) >= 100: return df
     
-    #检查平滑方法是否支持
+    #检查插值方法是否支持
     methodlist=['quadratic','cubic','slinear','linear','zero','nearest','time','index', \
             'piecewise_polynomial','pchip','akima','from_derivatives','spline','polynomial']
     if not (method in methodlist): return df
@@ -1599,13 +1717,14 @@ def df_smooth_manual(df,method='pchip',resample_freq='H',order=3):
         return df
     df.set_index('date',inplace=True)    
     
+    #重新采样
     try:
         dfh=df.resample(resample_freq)
     except:
         print('  #Warning(df_smooth): resampling failed for frequency',resample_freq)
         return df
     
-    #重新采样
+    #重采样后插值(不然太多nan)：是否methodlist_o里面的特别插值方法
     methodlist_o=['spline','polynomial']
     if method in methodlist_o:
         try:
@@ -1617,6 +1736,7 @@ def df_smooth_manual(df,method='pchip',resample_freq='H',order=3):
         #成功返回经过重采样的df
         return dfm
     
+    #重采样后插值：其他插值方法
     try:
         dfm=dfh.interpolate(method=method)
     except:

siat/risk_adjusted_return.py

@@ -339,7 +339,7 @@ if __name__=='__main__':
     graph=True    
     
 def rar_ratio_rolling(portfolio,start,end,ratio_name='treynor',RF=True, \
-                      window=30,graph=True,source='auto'):
+                      window=21,graph=True,source='auto'):
     """
     功能：滚动计算一个投资组合的风险调整后的收益率指数
     投资组合的结构：{'Market':('US','^GSPC'),'AAPL':0.5,'MSFT':0.3,'IBM':0.2}
@@ -1194,13 +1194,13 @@ if __name__=='__main__':
     trailing=20
 
 def compare_mrar(tickers,rar_name,start,end, \
-                 market="China",market_index="000300.SS",RF=False,window=60, \
+                 market="China",market_index="000300.SS",RF=False,window=63, \
                  axhline_value=0,axhline_label='零线',graph=True,printout=True, \
                  sortby='tpw_mean',source='auto',trailing=20,trend_threshhold=0.001, \
                  annotate=False):
     """
-    功能：计算多只股票的rar比率，并绘图对比
-    比率：支持夏普比率、特雷诺比率、索替诺比率、阿尔法比率等
+    功能：计算多只股票的rar比率，并绘图对比。多只股票必须处于同一个经济体的证券市场
+    比率：支持夏普比率、特雷诺比率、索替诺比率、阿尔法比率
     
     sortby: 
         tpw_mean(近期优先加权平均值降序排列)
@@ -1364,7 +1364,7 @@ if __name__=='__main__':
     trailing=20
 
 def compare_1security_mrar(ticker,rar_names,start,end, \
-                 market="China",market_index="000300.SS",RF=False,window=60, \
+                 market="China",market_index="000300.SS",RF=False,window=63, \
                  axhline_value=0,axhline_label='零线',graph=True,printout=False, \
                  sortby='tpw_mean',source='auto',trailing=20,trend_threshhold=0.001, \
                  annotate=False):