0

我不知道是什么破坏了我的代码，但是当我尝试将网络对象导出到 gephi 对象时，我不断收到关键错误 numpy.int64。我正在遍历组合列表并将网络对象放入 write_gexf 函数中。

Combnol = 
[[1992, <networkx.classes.graph.Graph at 0x10dfc5b50>],
 [2000, <networkx.classes.graph.Graph at 0x10dc58650>],

这是网络对象的样子

combnol[1].edge
{'Alliances  Partnerships': {'Education': {'difference': 1.0,
   'weight': 1,
   'year': 2001},

combnol[1].node
{'Alliances  Partnerships': {'color': 'y'},
 'Education': {'color': 'y'},
 'Entrepreneurship': {'color': 'y'},

#save file 
for i, no in enumerate(combnol):
    print repr(no[1])
    nx.write_gexf(no[1], 'no{0}'.format(i))
    break

追溯

KeyError                                  Traceback (most recent call last)
<ipython-input-1536-80e0a534cec9> in <module>()
      2 for i, no in enumerate(combnol):
      3     print repr(no[1])
----> 4     nx.write_gexf(no[1], 'no')
      5     break

//anaconda/lib/python2.7/site-packages/networkx/readwrite/gexf.pyc in write_gexf(G, path, encoding, prettyprint, version)

//anaconda/lib/python2.7/site-packages/networkx/utils/decorators.pyc in _open_file(func, *args, **kwargs)
    261         # Finally, we call the original function, making sure to close the fobj.
    262         try:
--> 263             result = func(*new_args, **kwargs)
    264         finally:
    265             if close_fobj:

//anaconda/lib/python2.7/site-packages/networkx/readwrite/gexf.pyc in write_gexf(G, path, encoding, prettyprint, version)
     75     writer = GEXFWriter(encoding=encoding,prettyprint=prettyprint,
     76                         version=version)
---> 77     writer.add_graph(G)
     78     writer.write(path)
     79 

//anaconda/lib/python2.7/site-packages/networkx/readwrite/gexf.pyc in add_graph(self, G)
    286         self.graph_element=graph_element
    287         self.add_nodes(G,graph_element)
--> 288         self.add_edges(G,graph_element)
    289         self.xml.append(graph_element)
    290 

//anaconda/lib/python2.7/site-packages/networkx/readwrite/gexf.pyc in add_edges(self, G, graph_element)
    361             edge_data=self.add_viz(edge_element,edge_data)
    362             edge_data=self.add_attributes("edge", edge_element,
--> 363                                           edge_data, default)
    364             edges_element.append(edge_element)
    365         graph_element.append(edges_element)

//anaconda/lib/python2.7/site-packages/networkx/readwrite/gexf.pyc in add_attributes(self, node_or_edge, xml_obj, data, default)
    379             if k == 'key':
    380                 k='networkx_key'
--> 381             attr_id = self.get_attr_id(make_str(k), self.xml_type[type(v)],
    382                                        node_or_edge, default, mode)
    383             if type(v)==list:

KeyError: <type 'numpy.int64'>

更新 - 回答问题：

print type(combnol[2][1].edge['Education']['Foundations']['difference'])
<type 'float'>

那么是否只是确保我添加整数而不是浮点数？int(combnol[2][1].edge['Education']['Foundations']['difference'])

这是我当前的实现，其中添加了值。我应该在下面的 1 周围添加一个 int() 吗？g[pair[0]][pair[1]]['weight'] += int(1); 整数（1 / 重量）

        # adjusting weight and difference if edge already exists
            #print pair[i][0], pair[i][1]
            if g.has_edge(pair[0], pair[1]):
                g[pair[0]][pair[1]]['weight'] += 1

            # adding edge if non-existant with initial weight = 1
            else:
                weight = 1
                g.add_edge(pair[0], pair[1], weight = weight, difference = 1. / weight, year= year)
4

0 回答 0