inheriting networkx Graph and using nx.connected_component_subgraphs

Question

I am trying to subclass a networkx Graph object. My __init__ has one variable passed to it. However, this means that when I try to use the following method which calls connected_component_iter ,

def connected_component_iter(self):
    """
    Yields connected components.
    """
    assert self.is_built is True
    for subgraph in nx.connected_component_subgraphs(self):
        yield subgraph

I get this error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "src/unitigGraph.py", line 163, in connected_component_iter
    def connected_component_iter(self):
  File "/Library/Python/2.7/site-packages/networkx/algorithms/components/connected.py", line 94, in connected_component_subgraphs
    yield G.subgraph(c).copy()
  File "/Library/Python/2.7/site-packages/networkx/classes/graph.py", line 1486, in subgraph
    H = self.__class__()
TypeError: __init__() takes exactly 2 arguments (1 given)

I would really prefer not to remove my initialization class variable. Is there a way I can still use the connected_component_iter method from Graph ?

Answer 1

You could work around the problem by giving your new initialization variable, val , a default value:

class MyGraph(nx.Graph):
    def __init__(self, data=None, val=None, **attr):
        super(MyGraph, self).__init__()
        self.val = val

Above, the default value for val is None. So

H = self.__class__()

would initialize a new subgraph with val equal to None .

However, it seems likely that you'd like the subgraph to inherit the same value of val as the parent MyGraph. In that case, we'd need to change

    H = self.__class__()

to

    H = self.__class__(val=self.val)

We can do this by overriding the subgraph method by defining our slightly altered version in MyGraph . For example, the code might look something like:

import networkx as nx
class MyGraph(nx.Graph):
    def __init__(self, data=None, val=None, **attr):
        super(MyGraph, self).__init__()
        self.val = val
        self.is_built = True

    def connected_component_iter(self):
        """
        Yields connected components.
        """
        assert self.is_built is True
        for subgraph in nx.connected_component_subgraphs(self):
            yield subgraph

    def subgraph(self, nbunch):
        bunch =self.nbunch_iter(nbunch)
        # create new graph and copy subgraph into it
        H = self.__class__(val=self.val)
        # copy node and attribute dictionaries
        for n in bunch:
            H.node[n]=self.node[n]
        # namespace shortcuts for speed
        H_adj=H.adj
        self_adj=self.adj
        # add nodes and edges (undirected method)
        for n in H.node:
            Hnbrs={}
            H_adj[n]=Hnbrs
            for nbr,d in self_adj[n].items():
                if nbr in H_adj:
                    # add both representations of edge: n-nbr and nbr-n
                    Hnbrs[nbr]=d
                    H_adj[nbr][n]=d
        H.graph=self.graph
        return H

G = MyGraph(val='val')
G.add_edges_from([(0, 1), (1, 2), (1, 3), (3, 5), (3, 6), (3, 7), (4, 8), (4, 9)])

for subgraph in G.connected_component_iter():
    print(subgraph.nodes(), subgraph.val)