Python networkx.connected_component_subgraphs() Examples

def getNetworkGraph(segments,segmentlengths):
    """
    Builds a networkx graph from the network file, inluding segment length taken from arcpy.
    It selects the largest connected component of the network (to prevent errors from routing between unconnected parts)
    """
    #generate the full network path for GDAL to be able to read the file
    path =str(os.path.join(arcpy.env.workspace,segments))
    print path
    if arcpy.Exists(path):
        g = nx.read_shp(path)
        #This selects the largest connected component of the graph
        sg = list(nx.connected_component_subgraphs(g.to_undirected()))[0]
        print "graph size (excluding unconnected parts): "+str(len(g))
        # Get the length for each road segment and append it as an attribute to the edges in the graph.
        for n0, n1 in sg.edges():
            oid = sg[n0][n1]["OBJECTID"]
            sg[n0][n1]['length'] = segmentlengths[oid]
        return sg
    else:
        print "network file not found on path: "+path 

def read_net(fname, weighted, directed, log):
    if weighted:
        G = nx.read_edgelist(inodetype=int, data=(('weight', float),),
                             create_using=nx.DiGraph())
    else:
        G = nx.read_edgelist(fname, nodetype=int, create_using=nx.DiGraph())
        for edge in G.edges():
            G[edge[0]][edge[1]]['weight'] = 1

    if not directed:
        G = G.to_undirected()

    log.info('N: %d E: %d' % (G.number_of_nodes(), G.number_of_edges()))
    log.info('CC: %d' % nx.number_connected_components(G))
    giant = max(nx.connected_component_subgraphs(G), key=len)
    log.info('N: %d E: %d' % (giant.number_of_nodes(), giant.number_of_edges()))
    return giant 

def get_seed_scaffold():

    g_idx = 1
    seed_scaffolds = {} #this stores initial long scaffolds
    to_merge = set()

    for subg in nx.connected_component_subgraphs(G):
        p = []
        for node in subg.nodes():
            if subg.degree(node) == 1:
                p.append(node)

        #If this is 2 then we have found the path!
        if len(p) == 2:
            path = nx.shortest_path(subg,p[0],p[1])
            seed_scaffolds[g_idx] = path
            g_idx += 1


        #else try to insert these contigs in the long scaffolds generated previously
        else:
            for node in subg.nodes():
                to_merge.add(node.split(':')[0])

    return seed_scaffolds, to_merge 

def getNetworkGraph(segments,segmentlengths):
    """
    Builds a networkx graph from the network file, inluding segment length taken from arcpy.
    It selects the largest connected component of the network (to prevent errors from routing between unconnected parts)
    """
    #generate the full network path for GDAL to be able to read the file
    path =str(os.path.join(arcpy.env.workspace,segments))
    print path
    if arcpy.Exists(path):
        g = nx.read_shp(path)
        #This selects the largest connected component of the graph
        sg = list(nx.connected_component_subgraphs(g.to_undirected()))[0]
        print "graph size (excluding unconnected parts): "+str(len(g))
        # Get the length for each road segment and append it as an attribute to the edges in the graph.
        for n0, n1 in sg.edges():
            oid = sg[n0][n1]["OBJECTID"]
            sg[n0][n1]['length'] = segmentlengths[oid]
        return sg
    else:
        print "network file not found on path: "+path 

def caveman_special(c=2,k=20,p_path=0.1,p_edge=0.3):
    p = p_path
    path_count = max(int(np.ceil(p * k)),1)
    G = nx.caveman_graph(c, k)
    # remove 50% edges
    p = 1-p_edge
    for (u, v) in list(G.edges()):
        if np.random.rand() < p and ((u < k and v < k) or (u >= k and v >= k)):
            G.remove_edge(u, v)
    # add path_count links
    for i in range(path_count):
        u = np.random.randint(0, k)
        v = np.random.randint(k, k * 2)
        G.add_edge(u, v)
    G = max(nx.connected_component_subgraphs(G), key=len)
    return G 

def caveman_special(c=2,k=20,p_path=0.1,p_edge=0.3):
    p = p_path
    path_count = max(int(np.ceil(p * k)),1)
    G = nx.caveman_graph(c, k)
    # remove 50% edges
    p = 1-p_edge
    for (u, v) in list(G.edges()):
        if np.random.rand() < p and ((u < k and v < k) or (u >= k and v >= k)):
            G.remove_edge(u, v)
    # add path_count links
    for i in range(path_count):
        u = np.random.randint(0, k)
        v = np.random.randint(k, k * 2)
        G.add_edge(u, v)
    G = max(nx.connected_component_subgraphs(G), key=len)
    return G 

def n_community(c_sizes, p_inter=0.01):
    graphs = [nx.gnp_random_graph(c_sizes[i], 0.7, seed=i) for i in range(len(c_sizes))]
    G = nx.disjoint_union_all(graphs)
    communities = list(nx.connected_component_subgraphs(G))
    for i in range(len(communities)):
        subG1 = communities[i]
        nodes1 = list(subG1.nodes())
        for j in range(i+1, len(communities)):
            subG2 = communities[j]
            nodes2 = list(subG2.nodes())
            has_inter_edge = False
            for n1 in nodes1:
                for n2 in nodes2:
                    if np.random.rand() < p_inter:
                        G.add_edge(n1, n2)
                        has_inter_edge = True
            if not has_inter_edge:
                G.add_edge(nodes1[0], nodes2[0])
    #print('connected comp: ', len(list(nx.connected_component_subgraphs(G))))
    return G 

def caveman_special(c=2,k=20,p_path=0.1,p_edge=0.3):
    p = p_path
    path_count = max(int(np.ceil(p * k)),1)
    G = nx.caveman_graph(c, k)
    # remove 50% edges
    p = 1-p_edge
    for (u, v) in list(G.edges()):
        if np.random.rand() < p and ((u < k and v < k) or (u >= k and v >= k)):
            G.remove_edge(u, v)
    # add path_count links
    for i in range(path_count):
        u = np.random.randint(0, k)
        v = np.random.randint(k, k * 2)
        G.add_edge(u, v)
    G = max(nx.connected_component_subgraphs(G), key=len)
    return G 

def n_community(c_sizes, p_inter=0.01):
    graphs = [nx.gnp_random_graph(c_sizes[i], 0.7, seed=i) for i in range(len(c_sizes))]
    G = nx.disjoint_union_all(graphs)
    communities = list(nx.connected_component_subgraphs(G))
    for i in range(len(communities)):
        subG1 = communities[i]
        nodes1 = list(subG1.nodes())
        for j in range(i+1, len(communities)):
            subG2 = communities[j]
            nodes2 = list(subG2.nodes())
            has_inter_edge = False
            for n1 in nodes1:
                for n2 in nodes2:
                    if np.random.rand() < p_inter:
                        G.add_edge(n1, n2)
                        has_inter_edge = True
            if not has_inter_edge:
                G.add_edge(nodes1[0], nodes2[0])
    #print('connected comp: ', len(list(nx.connected_component_subgraphs(G))))
    return G 

def Partition(UnclusteredGraph, code, type, scale):
    # Make edges on Unclustered graph
    # between all nodes separated by distance 'scale'
    # on Dual Lattice
    for node1 in UnclusteredGraph.nodes():
        for node2 in UnclusteredGraph.nodes():
            if node1 != node2:
                d = code.distance(type, node1, node2)
                if d <= scale:
                    UnclusteredGraph.add_edge(*(node1, node2), weight=d)
    Clusters = []
    # Networkx connected components analysis
    subgraphs = nx.connected_component_subgraphs(UnclusteredGraph)
    for i, sg in enumerate(subgraphs):
        Clusters.append(sg.nodes(data=True))
            

    return Clusters


# Choose fixed node for cluster fusion 

def strategy_connected_sequential(G, colors, traversal='bfs'):
    """
    Connected sequential ordering (CS). Yield nodes in such an order, that
    each node, except the first one, has at least one neighbour in the
    preceeding sequence. The sequence can be generated using both BFS and
    DFS search (using the strategy_connected_sequential_bfs and
    strategy_connected_sequential_dfs method). The default is bfs.
    """
    for component_graph in nx.connected_component_subgraphs(G):
        source = component_graph.nodes()[0]

        yield source  # Pick the first node as source

        if traversal == 'bfs':
            tree = nx.bfs_edges(component_graph, source)
        elif traversal == 'dfs':
            tree = nx.dfs_edges(component_graph, source)
        else:
            raise nx.NetworkXError(
                'Please specify bfs or dfs for connected sequential ordering')

        for (_, end) in tree:
            # Then yield nodes in the order traversed by either BFS or DFS
            yield end 

def setUp(self):
        self.undirected = [
            nx.connected_component_subgraphs,
            nx.biconnected_component_subgraphs,
        ]
        self.directed = [
            nx.weakly_connected_component_subgraphs,
            nx.strongly_connected_component_subgraphs,
            nx.attracting_component_subgraphs,
        ]
        self.subgraph_funcs = self.undirected + self.directed

        self.D = nx.DiGraph()
        self.D.add_edge(1, 2, eattr='red')
        self.D.add_edge(2, 1, eattr='red')
        self.D.nodes[1]['nattr'] = 'blue'
        self.D.graph['gattr'] = 'green'

        self.G = nx.Graph()
        self.G.add_edge(1, 2, eattr='red')
        self.G.nodes[1]['nattr'] = 'blue'
        self.G.graph['gattr'] = 'green' 

def setUp(self):
        self.undirected = [
            nx.connected_component_subgraphs,
            nx.biconnected_component_subgraphs,
        ]
        self.directed = [
            nx.weakly_connected_component_subgraphs,
            nx.strongly_connected_component_subgraphs,
            nx.attracting_component_subgraphs,
        ]
        self.subgraph_funcs = self.undirected + self.directed
        
        self.D = nx.DiGraph()
        self.D.add_edge(1, 2, eattr='red')
        self.D.add_edge(2, 1, eattr='red')
        self.D.nodes[1]['nattr'] = 'blue'
        self.D.graph['gattr'] = 'green'

        self.G = nx.Graph()
        self.G.add_edge(1, 2, eattr='red')
        self.G.nodes[1]['nattr'] = 'blue'
        self.G.graph['gattr'] = 'green' 

def test_connected_raise(self):
        assert_raises(NetworkXNotImplemented, nx.connected_components, self.DG)
        assert_raises(NetworkXNotImplemented, nx.number_connected_components, self.DG)
        assert_raises(NetworkXNotImplemented, nx.connected_component_subgraphs, self.DG)
        assert_raises(NetworkXNotImplemented, nx.node_connected_component, self.DG,1)
        assert_raises(NetworkXNotImplemented, nx.is_connected, self.DG)
        assert_raises(nx.NetworkXPointlessConcept, nx.is_connected, nx.Graph()) 

def is_bipartite_node_set(G,nodes):
    """Returns True if nodes and G/nodes are a bipartition of G.

    Parameters
    ----------
    G : NetworkX graph

    nodes: list or container
      Check if nodes are a one of a bipartite set.

    Examples
    --------
    >>> from networkx.algorithms import bipartite
    >>> G = nx.path_graph(4)
    >>> X = set([1,3])
    >>> bipartite.is_bipartite_node_set(G,X)
    True

    Notes
    -----
    For connected graphs the bipartite sets are unique.  This function handles
    disconnected graphs.
    """
    S=set(nodes)
    for CC in nx.connected_component_subgraphs(G):
        X,Y=sets(CC)
        if not ( (X.issubset(S) and Y.isdisjoint(S)) or
                 (Y.issubset(S) and X.isdisjoint(S)) ):
            return False
    return True 

def is_forest(G):
    """
    Returns True if `G` is a forest.

    A forest is a graph with no undirected cycles.

    For directed graphs, `G` is a forest if the underlying graph is a forest.
    The underlying graph is obtained by treating each directed edge as a single
    undirected edge in a multigraph.

    Parameters
    ----------
    G : graph
        The graph to test.

    Returns
    -------
    b : bool
        A boolean that is True if `G` is a forest.

    Notes
    -----
    In another convention, a directed forest is known as a *polyforest* and
    then *forest* corresponds to a *branching*.

    See Also
    --------
    is_branching

    """
    if len(G) == 0:
        raise nx.exception.NetworkXPointlessConcept('G has no nodes.')

    if G.is_directed():
        components = nx.weakly_connected_component_subgraphs
    else:
        components = nx.connected_component_subgraphs

    return all(len(c) - 1 == c.number_of_edges() for c in components(G)) 

def strategy_connected_sequential(G, colors, traversal='bfs'):
    """Returns an iterable over nodes in ``G`` in the order given by a
    breadth-first or depth-first traversal.

    ``traversal`` must be one of the strings ``'dfs'`` or ``'bfs'``,
    representing depth-first traversal or breadth-first traversal,
    respectively.

    The generated sequence has the property that for each node except
    the first, at least one neighbor appeared earlier in the sequence.

    ``G`` is a NetworkX graph. ``colors`` is ignored.

    """
    if traversal == 'bfs':
        traverse = nx.bfs_edges
    elif traversal == 'dfs':
        traverse = nx.dfs_edges
    else:
        raise nx.NetworkXError("Please specify one of the strings 'bfs' or"
                               " 'dfs' for connected sequential ordering")
    for component in nx.connected_component_subgraphs(G):
        source = arbitrary_element(component)
        # Yield the source node, then all the nodes in the specified
        # traversal order.
        yield source
        for (_, end) in traverse(component, source):
            yield end 

def _check_separating_sets(G):
    for Gc in nx.connected_component_subgraphs(G):
        if len(Gc) < 3:
            continue
        node_conn = nx.node_connectivity(Gc)
        for cut in nx.all_node_cuts(Gc):
            assert_equal(node_conn, len(cut))
            H = Gc.copy()
            H.remove_nodes_from(cut)
            assert_false(nx.is_connected(H)) 

def test_weakly_connected_component_subgraphs(self):
        wcc = nx.weakly_connected_component_subgraphs
        cc = nx.connected_component_subgraphs
        for G, C in self.gc:
            U = G.to_undirected()
            w = {frozenset(g) for g in wcc(G)}
            c = {frozenset(g) for g in cc(U)}
            assert_equal(w, c) 

def is_bipartite_node_set(G, nodes):
    """Returns True if nodes and G/nodes are a bipartition of G.

    Parameters
    ----------
    G : NetworkX graph

    nodes: list or container
      Check if nodes are a one of a bipartite set.

    Examples
    --------
    >>> from networkx.algorithms import bipartite
    >>> G = nx.path_graph(4)
    >>> X = set([1,3])
    >>> bipartite.is_bipartite_node_set(G,X)
    True

    Notes
    -----
    For connected graphs the bipartite sets are unique.  This function handles
    disconnected graphs.
    """
    S = set(nodes)
    for CC in nx.connected_component_subgraphs(G):
        X, Y = sets(CC)
        if not ((X.issubset(S) and Y.isdisjoint(S)) or
                (Y.issubset(S) and X.isdisjoint(S))):
            return False
    return True 

def test_weakly_connected_component_subgraphs(self):
        wcc = nx.weakly_connected_component_subgraphs
        cc = nx.connected_component_subgraphs
        for G, C in self.gc:
            U = G.to_undirected()
            w = {frozenset(g) for g in wcc(G)}
            c = {frozenset(g) for g in cc(U)}
            assert_equal(w, c) 

def test_connected_raise(self):
        assert_raises(NetworkXNotImplemented, nx.connected_components, self.DG)
        assert_raises(NetworkXNotImplemented, nx.number_connected_components, self.DG)
        assert_raises(NetworkXNotImplemented, nx.connected_component_subgraphs, self.DG)
        assert_raises(NetworkXNotImplemented, nx.node_connected_component, self.DG,1)
        assert_raises(NetworkXNotImplemented, nx.is_connected, self.DG)
        assert_raises(nx.NetworkXPointlessConcept, nx.is_connected, nx.Graph()) 

def test_connected_component_subgraphs(self):
        wcc = nx.weakly_connected_component_subgraphs
        cc = nx.connected_component_subgraphs
        for G, C in self.gc:
            U = G.to_undirected()
            w = {frozenset(g) for g in wcc(G)}
            c = {frozenset(g) for g in cc(U)}
            assert_equal(w, c) 

def _check_separating_sets(G):
    for Gc in nx.connected_component_subgraphs(G):
        if len(Gc) < 3:
            continue
        for cut in nx.all_node_cuts(Gc):
            assert_equal(nx.node_connectivity(Gc), len(cut))
            H = Gc.copy()
            H.remove_nodes_from(cut)
            assert_false(nx.is_connected(H)) 

def is_bipartite_node_set(G,nodes):
    """Returns True if nodes and G/nodes are a bipartition of G.

    Parameters
    ----------
    G : NetworkX graph

    nodes: list or container
      Check if nodes are a one of a bipartite set.

    Examples
    --------
    >>> from networkx.algorithms import bipartite
    >>> G = nx.path_graph(4)
    >>> X = set([1,3])
    >>> bipartite.is_bipartite_node_set(G,X)
    True

    Notes
    -----
    For connected graphs the bipartite sets are unique.  This function handles
    disconnected graphs.
    """
    S=set(nodes)
    for CC in nx.connected_component_subgraphs(G):
        X,Y=sets(CC)
        if not ( (X.issubset(S) and Y.isdisjoint(S)) or
                 (Y.issubset(S) and X.isdisjoint(S)) ):
            return False
    return True 

def establish_groups(self):
        d = list(nx.connected_component_subgraphs(self.trail_network))
        for i, group in enumerate(d):
            for node in group:
                self.path_groups[node] = i
            self.group_list.append(i) 

def attackable(classifier, s2v_g, x = None, y = None):
    g = s2v_g.to_networkx()
    comps = [c for c in nx.connected_component_subgraphs(g)]
    set_id = {}

    for i in range(len(comps)):
        for j in comps[i].nodes():
            set_id[j] = i
    
    if x is not None:
        r_i = [x]
    else:
        r_i = range(len(g) - 1)

    g_list = []    
    for i in r_i:
        if y is not None:
            assert x is not None
            r_j = [y]
        else:
            if x is not None:
                r_j = range(len(g) - 1)
            else:
                r_j = range(i + 1, len(g))
        for j in r_j:
            if set_id[i] != set_id[j]:
                continue
            g2 = g.copy()
            g2.add_edge(i, j)
            assert nx.number_connected_components(g2) == s2v_g.label
            g_list.append(S2VGraph(g2, s2v_g.label))
    if len(g_list) == 0:
        print(x, y)
        print(g.edges(), s2v_g.label)
        print(set_id)
    assert len(g_list)
    _, _, acc = classifier(g_list)

    return np.sum(acc.view(-1).numpy()) < len(g_list) 

def bannedActions(self, g, node_x):        
        comps = [c for c in nx.connected_component_subgraphs(g)]
        set_id = {}
        for i in range(len(comps)):
            for j in comps[i].nodes():
                set_id[j] = i

        banned_actions = set()
        for i in range(len(g)):
            if set_id[i] != set_id[node_x] or i == node_x:
                banned_actions.add(i)
        return banned_actions 

def citeseer_ego():
    _, _, G = data.Graph_load(dataset='citeseer')
    G = max(nx.connected_component_subgraphs(G), key=len)
    G = nx.convert_node_labels_to_integers(G)
    graphs = []
    for i in range(G.number_of_nodes()):
        G_ego = nx.ego_graph(G, i, radius=3)
        if G_ego.number_of_nodes() >= 50 and (G_ego.number_of_nodes() <= 400):
            graphs.append(G_ego)
    return graphs 

def propose_attack(model, s2v_g, num_added=1):
    g = s2v_g.to_networkx()
    comps = [c for c in nx.connected_component_subgraphs(g)]
    set_id = {}

    for i in range(len(comps)):
        for j in comps[i].nodes():
            set_id[j] = i

    cand = []
    for i in range(len(g) - 1):
        for j in range(i + 1, len(g)):
            if set_id[i] != set_id[j] or i == j:
                continue
            cand.append('%d %d' % (i, j))
    
    if cmd_args.rand_att_type == 'random':
        added = np.random.choice(cand, num_added)
        added = [(int(w.split()[0]), int(w.split()[1])) for w in added]
        g.add_edges_from(added)
        return S2VGraph(g, s2v_g.label)
    elif cmd_args.rand_att_type == 'exhaust':
        g_list = []
        for c in cand:
            x, y = [int(w) for w in c.split()]
            g2 = g.copy()
            g2.add_edge(x, y)
            g_list.append(S2VGraph(g2, s2v_g.label))
        _, _, acc = model(g_list)
        ans = g_list[0]
        for i in range(len(g_list)):
            if acc.numpy()[i] < 1:
                ans = g_list[i]
                break
        return ans
    else:
        raise NotImplementedError