如何在networkx和pandas中使用字符串相似性而不是精确字符串匹配G.neighbors连接节点

问题描述

G1 = nx.from_pandas_edgelist(df,'Name','State')
G2 = nx.from_pandas_edgelist(df,'State','Name')
GT = nx.compose(G1,G2)

print(GT.neighbors(node))

如何通过相似节点（字符串匹配）而不是匹配相同节点来连接节点？

解决方法

import networkx as nx
from fuzzywuzzy import fuzz

def hot_insert_edge_str_match(G):
    for node in G.nodes():
        for non_neighbor in list(nx.non_neighbors(G,node)):
            if fuzz.ratio(node,non_neighbor) > 90:
                print("string matched",node,non_neighbor)
                G.add_edge(node,non_neighbor)

解决方法是手动循环遍历所有节点和非邻居，如果两个字符串匹配且高于阈值，则连接它们