从字符串而不是文件中读取networkx中的点图

Question

根据此处的文档，该read_dot()方法将文件句柄作为输入。

即使我这样做也能正常工作，

g = nx.drawing.nx_agraph.read_dot(dotFile)

dotFile文件句柄在哪里。但是，如果我dot在变量中格式化了图形，如何以 nx 图形格式导入？

编辑 - 我的变量看起来像这样，

dotFormat = """
digraph G{
edge [dir=forward]
node [shape=plaintext]
0 [label="0 (None)"]
0 -> 7 [label="root"]
1 [label="1 (The)"]
4 [label="4 (great Indian Circus)"]
4 -> 4 [label="compound"]
4 -> 1 [label="det"]
4 -> 4 [label="amod"]
5 [label="5 (is)"]
6 [label="6 (in)"]
7 [label="7 (Mumbai)"]
7 -> 6 [label="case"]
7 -> 5 [label="cop"]
7 -> 4 [label="nsubj"]
}
"""

编辑2 - 我还尝试阅读py_dot然后转换为这样的networkx图。

pg = pydot.graph_from_dot_data(dotFormat)
g = nx.nx_pydot.from_pydot(pg)

然后我得到以下错误

if P.get_strict(None): # pydot bug: get_strict() shouldn't take argument
AttributeError: 'list' object has no attribute 'get_strict'

我应该怎么办？

Answer 1

您可以通过显式使用 pygraphviz 加载图形，然后将其转换为 networkx 图（在本例中为 MultiDiGraph）来完成此操作。

In [1]: import pygraphviz

In [2]: from networkx.drawing import nx_agraph

In [3]: dotFormat = """
digraph G{
edge [dir=forward]
node [shape=plaintext]
0 [label="0 (None)"]
0 -> 7 [label="root"]
1 [label="1 (The)"]
4 [label="4 (great Indian Circus)"]
4 -> 4 [label="compound"]
4 -> 1 [label="det"]
4 -> 4 [label="amod"]
5 [label="5 (is)"]
6 [label="6 (in)"]
7 [label="7 (Mumbai)"]
7 -> 6 [label="case"]
7 -> 5 [label="cop"]
7 -> 4 [label="nsubj"]
}
"""

In [4]: G = nx_agraph.from_agraph(pygraphviz.AGraph(dotFormat))