Seaborn对于x countplot设置图例值

Question

我ploting一个明确的数据和值数由sns.countplot()

我想对x值的传说添加到图如下：手柄设置x值的，标签是x值的描述。

ax = sns.countplot(x = df.GARAGE_DOM) 
handles, labels = ax.get_legend_handles_labels() 

handles = ["VP", "BC", "GC", "GP", "JC", "PO"] 
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"] 
by_label = OrderedDict(zip(handles,labels)) 
ax.legend(by_label.keys(), by_label.values()) 

但是，我警告

UserWarning：

联想不支持 '副总裁' 的情况。 可以使用代理艺术家代替。 见：http://matplotlib.org/users/legend_guide.html#using-proxy-artist

我读过代理艺术家的文档，但我没有找到我的案例。

感谢您的帮助。

Answer 1

这是一个可能的解决方案，创建一个文本字段作为图例处理程序。 以下内容将创建一个TextHandler用于创建图例艺术家，这是一个简单的matplotlib.text.Text实例。图例的手柄以（文本，颜色）元组给出，TextHandler创建所需的Text。

import seaborn as sns 
import matplotlib.pyplot as plt 
from matplotlib.legend_handler import HandlerBase 
from matplotlib.text import Text 
import numpy as np 
import pandas as pd 

class TextHandler(HandlerBase): 
    def create_artists(self, legend, tup ,xdescent, ydescent, 
         width, height, fontsize,trans): 
     tx = Text(width/2.,height/2,tup[0], fontsize=fontsize, 
        ha="center", va="center", color=tup[1], fontweight="bold") 
     return [tx] 


a = np.random.choice(["VP", "BC", "GC", "GP", "JC", "PO"], size=100, 
        p=np.arange(1,7)/21.) 
df = pd.DataFrame(a, columns=["GARAGE_DOM"]) 

ax = sns.countplot(x = df.GARAGE_DOM) 


handltext = ["VP", "BC", "GC", "GP", "JC", "PO"] 
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"] 


t = ax.get_xticklabels() 
labeldic = dict(zip(handltext, labels)) 
labels = [labeldic[h.get_text()] for h in t] 
handles = [(h.get_text(),c.get_fc()) for h,c in zip(t,ax.patches)] 

ax.legend(handles, labels, handler_map={tuple : TextHandler()}) 

plt.show() 

---

上述溶液是低于原始版本，这似乎更复杂的更新版本。 以下是原始解决方案，它使用 TextArea和 AnchoredOffsetbox将文本放置在图例中。

import seaborn.apionly as sns 
import matplotlib.pyplot as plt 
import matplotlib.patches as patches 
from matplotlib.offsetbox import TextArea, AnchoredOffsetbox 
from matplotlib.transforms import TransformedBbox, Bbox 
from matplotlib.legend_handler import HandlerBase 
import numpy as np 
import pandas as pd 

class TextHandler(HandlerBase): 
    def __init__(self, text, color="k"): 
     self.text = text 
     self.color = color 
     super(TextHandler, self).__init__() 

    def create_artists(self, legend, orig_handle,xdescent, ydescent, 
         width, height, fontsize,trans): 
     bb = Bbox.from_bounds(xdescent,ydescent, width,height) 
     tbb = TransformedBbox(bb, trans) 
     textbox = TextArea(self.text, textprops={"weight":"bold","color":self.color}) 
     ab = AnchoredOffsetbox(loc=10,child=textbox, bbox_to_anchor=tbb, frameon=False) 
     return [ab] 


a = np.random.choice(["VP", "BC", "GC", "GP", "JC", "PO"], size=100, 
        p=np.arange(1,7)/21.) 
df = pd.DataFrame(a, columns=["GARAGE_DOM"]) 

ax = sns.countplot(x = df.GARAGE_DOM) 


handltext = ["VP", "BC", "GC", "GP", "JC", "PO"] 
labels = ["Voie Publique", "box", "Garage couvert", "garage particulier clos", "Jardin clos", "parking ouvert"] 

handles = [ patches.Rectangle((0,0),1,1) for h in handltext] 
t = ax.get_xticklabels() 
labeldic = dict(zip(handltext, labels)) 
labels = [labeldic[h.get_text()] for h in t] 
handlers = [TextHandler(h.get_text(),c.get_fc()) for h,c in zip(t,ax.patches)] 
handlermap = dict(zip(handles, handlers)) 
ax.legend(handles, labels, handler_map=handlermap,) 

plt.show()