如何在Matplotlib中使用Font Awesome符号作为标记

Question

是否可以使用matplotlib在散点图中使用字体真棒图标作为标记？或者是否可以将其用作字体并将图标作为"文本"？

---

^{这个问题最初是在这里提出的,但由于未知原因而被关闭.由于我认为这是一个有效且有用的问题,而Stackoverflow上的任何地方都没有解决这个问题,而且肯定值得回答,我只想再问一遍.}

Answer 1

FontAwesome可从这里获得.它提供其图标作为矢量图形和otf字体.

使用FontAwesome otf字体

Matplotlib不能原生地读取矢量图形,但它可以加载otf字体.下载FontAwesome软件包后,您可以通过matplotlib.font_manager.FontProperties对象访问字体,例如

fp = FontProperties(fname=r"C:\Windows\Fonts\Font Awesome 5 Free-Solid-900.otf") 

创建文本

该FontProperties可为matplotlib文本对象输入

plt.text(.6, .4, "\uf16c", fontproperties=fp)

不幸的是,使用FontAwesome连字是不可能的.因此,需要通过UTF8密钥访问各个符号.这有点麻烦,但是这里的备忘单可以派上用场.将那些需要的符号存储在具有有意义名称的字典中可能是有意义的.

例:

from matplotlib.font_manager import FontProperties
import matplotlib.pyplot as plt

fp1 = FontProperties(fname=r"C:\Windows\Fonts\Font Awesome 5 Brands-Regular-400.otf")
fp2 = FontProperties(fname=r"C:\Windows\Fonts\Font Awesome 5 Free-Solid-900.otf")

symbols = dict(cloud = "\uf6c4", campground = "\uf6bb", hiking = "\uf6ec",
               mountain = "\uf6fc", tree = "\uf1bb", fish = "\uf578",
               stackoverflow = "\uf16c")

fig, (ax, ax2) = plt.subplots(ncols=2, figsize=(6.2, 2.2), sharey=True)
ax.text(.5, .5, symbols["stackoverflow"], fontproperties=fp1, size=100, 
         color="orange", ha="center", va="center")


ax2.stackplot([0,.3,.55,.6,.65,1],[.1,.2,.2,.2,.2,.15],[.3,.2,.2,.3,.2,.2],
              colors=["paleturquoise", "palegreen"])
ax2.axis([0,1,0,1])
ax2.text(.6, .4, symbols["mountain"], fontproperties=fp2, size=16, ha="center")
ax2.text(.09, .23, symbols["campground"], fontproperties=fp2, size=13)
ax2.text(.22, .27, symbols["hiking"], fontproperties=fp2, size=14)
ax2.text(.7, .24, symbols["tree"], fontproperties=fp2, size=14,color="forestgreen")
ax2.text(.8, .33, symbols["tree"], fontproperties=fp2, size=14,color="forestgreen")
ax2.text(.88, .28, symbols["tree"], fontproperties=fp2, size=14,color="forestgreen")
ax2.text(.35, .03, symbols["fish"], fontproperties=fp2, size=14,)
ax2.text(.2, .7, symbols["cloud"], fontproperties=fp2, size=28,)

plt.show()

创建标记

创建像上面这样的很多文本并不是很方便.将图标作为标记对于某些应用程序来说会更好.Matplotlib确实能够使用utf符号作为标记,但是,只能通过mathtext功能.在我的试验中,在matplotlib中使用otf字体作为mathfont是不成功的.

另一种方法是matplotlib.path.Path从符号中创建一个.这可以通过一个matplotlib.textpath.TextToPath实例完成,遗憾的是没有文档.该TextToPath有一个方法get_text_path采取fontproperty和串作为输入,并返回顶点和从哪个码来创建Path.A Path可以用作标记,例如用于scatter绘图.

v, codes = TextToPath().get_text_path(fp, \uf6fc)
path = Path(v, codes, closed=False)
plt.scatter(..., marker=path)

一些例子:

import numpy as np; np.random.seed(32)
from matplotlib.path import Path
from matplotlib.textpath import TextToPath
from matplotlib.font_manager import FontProperties
import matplotlib.pyplot as plt

fp = FontProperties(fname=r"C:\Windows\Fonts\Font Awesome 5 Free-Solid-900.otf")

symbols = dict(cloud = "\uf6c4", campground = "\uf6bb", hiking = "\uf6ec",
               mountain = "\uf6fc", tree = "\uf1bb", fish = "\uf578",
               stackoverflow = "\uf16c")

fig, ax = plt.subplots()

def get_marker(symbol):
    v, codes = TextToPath().get_text_path(fp, symbol)
    v = np.array(v)
    mean = np.mean([np.max(v,axis=0), np.min(v, axis=0)], axis=0)
    return Path(v-mean, codes, closed=False)

x = np.random.randn(4,10)
c = np.random.rand(10)
s = np.random.randint(120,500, size=10)
plt.scatter(*x[:2], s=s, c=c, marker=get_marker(symbols["cloud"]), 
            edgecolors="none", linewidth=2)
plt.scatter(*x[2:], s=s, c=c, marker=get_marker(symbols["fish"]), 
            edgecolors="none", linewidth=2)   

plt.show()