python – 在NetworkX中组合两个加权图

概述我正在使用 python多处理来创建多个不同的NetworkX图,然后使用下面的函数来组合图.但是,虽然此函数适用于小图形,但对于较大的图形,它使用大量内存并挂在我的系统和内存繁重的AWS系统上(仅占系统总内存的三分之一).有没有更有效的方法来执行以下功能？ def combine_graphs(graph1, graph2, graph2_weight = 1): ''' Giv 我正在使用 python多处理来创建多个不同的NetworkX图,然后使用下面的函数来组合图.但是,虽然此函数适用于小图形,但对于较大的图形,它使用大量内存并挂在我的系统和内存繁重的AWS系统上(仅占系统总内存的三分之一).有没有更有效的方法来执行以下功能？

def combine_graphs(graph1,graph2,graph2_weight = 1):    '''    Given two graphs of different edge (but same node) structure (and the same type),combine the two graphs,summing all edge attributes and multiplying the second one's    attributes by the desired weights.     E.g. if graph1.edge[a][b] = {'a': 1,'b':2} and     graph2.edge[a][b] = {'a': 3,'c': 4},with a weight of 1 the final graph edge should be     final_graph.edge[a][b] = {'a': 4,'b': 2,'c': 4} and with a weight     of .5 the final graph edge should be {'a': 2.5,'c': 2}.    inputs: Two graphs to be combined and a weight to give to the second graph    '''    if type(graph1) != type(graph2) or len(set(graph2.nodes()) - set(graph1.nodes())) > 0:        raise Exception('Graphs must have the same type and graph 2 cannot have nodes that graph 1 does not have.')    # make a copy of the new graph to ensure that it doesn't change    new_graph = graph1.copy()    # iterate over graph2's edges,adding them to graph1    for node1,node2 in graph2.edges():        # if that edge already exists,Now iterate over the attributes        if new_graph.has_edge(node1,node2):            for attr in graph2.edge[node1][node2]:                # if that attribute exists,sum the values,otherwise,simply copy attrs                if new_graph.edge[node1][node2].get(attr) is not None:                    # try adding weighted value: if it fails,it's probably not numeric so add the full value (the only other option is a List)                    try:                        new_graph.edge[node1][node2][attr] += graph2.edge[node1][node2][attr] * graph2_weight                    except:                        new_graph.edge[node1][node2][attr] += graph2.edge[node1][node2][attr]                else:                    try:                        new_graph.edge[node1][node2][attr] = graph2.edge[node1][node2][attr] * graph2_weight                    except:                        new_graph.edge[node1][node2][attr] = graph2.edge[node1][node2][attr]        # otherwise,add the new edge with all its atributes -- first,iterate through those attributes to weight them        else:            attr_dict = graph2.edge[node1][node2]            for item in attr_dict:                try:                    attr_dict[item] = attr_dict[item] * graph2_weight                except:                    continue            new_graph.add_edge(node1,node2,attr_dict = attr_dict)    return new_graph

解决方法 内存将在您的代码中扩展两个位置：

1)制作graph1的副本(也许你需要保留一份副本)

2)使用graph2.edges()生成内存中所有边的列表,graph2.edges_iter()迭代边而不创建新列表

您也可以通过不同方式处理边缘数据来加快速度.您可以在迭代边缘时获取数据对象,而不必像字典查找那样执行：

def combined_graphs_edges(G,H,weight = 1.0):    for u,v,hdata in H.edges_iter(data=True):        # multply attributes of H by weight        attr = dict( (key,value*weight) for key,value in hdata.items())        # get data from G or use empty dict if no edge in G        gdata = G[u].get(v,{})        # add data from g        # sum shared items        shared = set(gdata) & set(hdata)        attr.update(dict((key,attr[key] + gdata[key]) for key in shared))        # non shared items        non_shared = set(gdata) - set(hdata)        attr.update(dict((key,gdata[key]) for key in non_shared))        yIEld u,attr    returnif __name__ == '__main__':    import networkx as nx    G = nx.Graph([('a','b',{'a': 1,'b':2})])    H = nx.Graph([('a',{'a': 3,'c':4})])    print List(combined_graphs_edges(G,weight=0.5))    # or to make a new graph     graph = G.copy()    graph.add_edges_from(combined_graphs_edges(G,weight=0.5))

总结

以上是内存溢出为你收集整理的python – 在NetworkX中组合两个加权图全部内容，希望文章能够帮你解决python – 在NetworkX中组合两个加权图所遇到的程序开发问题。

如果觉得内存溢出网站内容还不错，欢迎将内存溢出网站推荐给程序员好友。