将Python列表转换为networkx图

Question

目前我有一个清单:

[['???? ????', 10, '????'],
 ['???? ????', 10, '????', 5, '?????'],
 ['???? ????', 10, '????', 3, '?????'],
 ['???? ????', 10, '????', 3, '?????', 1, '??????']]

我需要以某种方式将其转换为Networkx边缘,其中单词对应该成为图形的节点,并且成为权重之间的整数:

G = nx.Graph()
G.add_edge('???? ????', '????', weight=10)
G.add_edge('????', '?????', weight=5)
G.add_edge('????', '?????', weight=3)
G.add_edge('?????', '??????', weight=1)

目前我被困,没有想法.任何帮助,将不胜感激!

Answer 1

由于您有重复的信息，我建议首先创建一个字典以确保唯一性。我添加了一项检查，以确保重复元素的距离不会发生冲突。

\n\n

In [1]:\ndistances = {}\nfor row in l:\n    for i in range(0, len(row)-1, 2):\n        key_tuple = (row[i], row[i+2])\n        d = row[i+1]\n        if key_tuple in distances.keys():\n            if distances[key_tuple] != d:\n                print("Warning: Found a conflicting distance for {}: {} and "\n                      "{}. Using last".format(key_tuple, distances[key_tuple], d))\n        distances[key_tuple] = d\n\nIn [2]: distances\nOut[2]:\n {(\'\xd0\x9c\xd0\xb0\xd0\xbc\xd0\xb0 \xd0\xbc\xd1\x8b\xd1\x82\xd1\x8c\', \'\xd1\x80\xd0\xb0\xd0\xbc\xd0\xb0\'): 10,\n (\'\xd1\x80\xd0\xb0\xd0\xbc\xd0\xb0\', \'\xd0\xb4\xd0\xbe\xd0\xbb\xd0\xb3\xd0\xbe\'): 5,\n (\'\xd1\x80\xd0\xb0\xd0\xbc\xd0\xb0\', \'\xd0\xb2\xd1\x87\xd0\xb5\xd1\x80\xd0\xb0\'): 3,\n (\'\xd0\xb2\xd1\x87\xd0\xb5\xd1\x80\xd0\xb0\', \'\xd0\xbf\xd0\xbe\xd0\xb7\xd0\xb4\xd0\xbd\xd0\xbe\'): 1}\n

\n\n

然后您可以使用该字典创建边缘。

\n\n

In [3]:\nimport networkx as nx\nG = nx.Graph()\nfor k, v in distances.items():\n    G.add_edge(k[0], k[1], weight=v)\n

\n