错误"未找到GraphViz的可执行文件"在iPython笔记本中从NetworkX调用GraphViz布局时

时间:2014-10-02 01:59:39

标签: python graphviz ipython-notebook networkx enthought

我收到错误" InvocationException:未找到GraphViz的可执行文件,错误GraphViz的可执行文件未找到"当我在我的代码中尝试使用graphviz_layout方法时。当graphviz_layout被spring_layout替换时,代码工作正常。从throwback异常信息中,它看起来像是调用pydot_layout。我通过Canopy的软件包管理器安装了pydot模块,但导入它并没有帮助。我也已经安装了graphviz并导入它没有运气。

这是我的代码:

import networkx as nx
import matplotlib.pyplot as plt

keywordTreeFile = open('decode_wordnet/keywordTreeFile.TXT','r') #keyword generation file
keywordTreeFileLineData = keywordTreeFile.readlines()

G = nx.Graph()
pairData = []

for i in range(0,len(keywordTreeFileLineData)):
    pairData = pairData + [keywordTreeFileLineData[i].split('\t')]
    pairData[i][1] = pairData[i][1].rstrip('\n')
    G.add_edge(pairData[i][0],pairData[i][1])

pos = nx.graphviz_layout(G)

nx.draw(G,
        pos=pos,
        width = 1.0,
        with_labels = True,
        font_size = 3,
        linewidths=.1
        )
plt.savefig("graph.pdf")

以下是回归异常信息:

---------------------------------------------------------------------------
InvocationException                       Traceback (most recent call last)
<ipython-input-1-e85a11cf6191> in <module>()
     15 
     16 #pos = nx.spring_layout(G,k=.15,iterations=50,scale=100)
---> 17 pos = nx.graphviz_layout(G)
     18 #labels = nx.draw_networkx_labels(G,pos)
     19 nx.draw(G,

/Users/scott/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/networkx/drawing/nx_pydot.pyc in graphviz_layout(G, prog, root, **kwds)
    245     This is a wrapper for pydot_layout.
    246     """
--> 247     return pydot_layout(G=G,prog=prog,root=root,**kwds)
    248 
    249 

/Users/scott/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/networkx/drawing/nx_pydot.pyc in pydot_layout(G, prog, root, **kwds)
    269         P.set("root",make_str(root))
    270 
--> 271     D=P.create_dot(prog=prog)
    272 
    273     if D=="":  # no data returned

/Users/scott/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/pydot.pyc in <lambda>(f, prog)
   1800             self.__setattr__(
   1801                 'create_'+frmt,
-> 1802                 lambda f=frmt, prog=self.prog : self.create(format=f, prog=prog))
   1803             f = self.__dict__['create_'+frmt]
   1804             f.__doc__ = '''Refer to the docstring accompanying the 'create' method for more information.'''

/Users/scott/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/pydot.pyc in create(self, prog, format)
   1951             if self.progs is None:
   1952                 raise InvocationException(
-> 1953                     'GraphViz\'s executables not found' )
   1954 
   1955         if not self.progs.has_key(prog):

InvocationException: GraphViz's executables not found

系统信息:我在OS X 10.9.5的64位Mac上使用了Enthought Canopy的最新版本和模块。另外,我使用的是iPython笔记本。

3 个答案:

答案 0 :(得分:1)

我通过使用macports在我的Mac上安装GraphViz解决了这个问题:

sudo port install graphviz

在这种情况下安装python包是不够的。尽管如此,这是Enthought应该采用的依赖。

答案 1 :(得分:0)

通常,您应该从全新安装的Macports开始

  1. 删除旧的Macport

    sudo port -f uninstall installed
    sudo port clean --work --archive all

  2. 导入,安装和编译最近的Macports源文件

    cd〜
    卷曲-O https://distfiles.macports.org/MacPorts/MacPorts-2.3.3.tar.bz2
    tar xf MacPorts-2.3.3.tar.bz2
    cd MacPorts-2.3.3 /
    的./configure
    使
    sudo make install
    港口版本

  3. (可选)更新Macports

    sudo port selfupdate
    sudo port升级过时了

  4. 使用新的Macports安装Graphviz

    sudo port install graphviz
    sudo pip install --upgrade graphviz

答案 2 :(得分:0)

对于ubuntu用户:-

sudo apt-get install -y graphviz libgraphviz-dev

这将安装python显示图所需的graphviz可执行文件