根据顶点名称 Python igraph 执行图的并集 [英] Perform union of graphs based on vertex names Python igraph

问题描述

这个问题已经在 github 上提交了，大约 6 个月前，但自从它尚未修复，我想知道是否有我遗漏的快速修复.

This issue has been filed on github something like 6 months ago, but since it has not yet been fixed I'm wondering whether there is a quick fix that I am missing.

我想根据名称合并两个图:

I want to merge two graphs based on their names:

g1 = igraph.Graph()
g2 = igraph.Graph()

# add vertices
g1.add_vertices(["A","B"])
g2.add_vertices(["B","C","D"])

for vertex in g1.vs:
    print vertex.index
0
1

for vertex in g2.vs:
    print vertex.index
0
1
2

然而，当我执行联合时，igraph 使用顶点 ID 而不是名称，所以我最终得到三个顶点而不是四个(如果它基于名称).我猜是因为B在g2中有索引0，所以它与g1<的A合并/代码>.以类似的方式，g2的C与g1的B合并.

However when I perform the union, igraph uses the vertex IDs rather than the names, so I end up with three vertices instead of four (if it was based on names). I guess that because B has index 0 in g2, it is merged with A of g1. And in a similar way, C of g2 is merged with B of g1.

g_union = igraph.Graph.union(g1,g2)

g_union.vs['name'] # of course
KeyError: 'Attribute does not exist'

for vertex in g_union.vs:
    print vertex.index
0
1
2

知道如何绕过这个问题吗?这是可能的，因为它是在 igraph 的 R 实现中完成的.

Any idea on how to bypass this issue? This is possible, since it was done in the R implementation of igraph.

推荐答案

只需创建一个新图形，并按名称添加顶点.当然，这会消除其他节点属性，您也必须手动添加这些属性.

Simply make a new graph, and add vertices by name. Of course, this would eliminate other node properties, which you would also have to add manually.

g1 = igraph.Graph()
g2 = igraph.Graph()

# add vertices
g1.add_vertices(["A","B"])
g2.add_vertices(["B","C","D"])

g3 = igraph.Graph()
verts_to_add = []
for v in g1.vs:
    if v['name'] not in verts_to_add:
        verts_to_add.append(v['name'])
for v in g2.vs:
    if v['name'] not in verts_to_add:
        verts_to_add.append(v['name'])

g3.add_vertices(verts_to_add)

for v in g3.vs:
    print(v['name'])

#A
#B
#C
#D

这篇关于根据顶点名称 Python igraph 执行图的并集的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！