python-社交网络数据处理

边表node值非连续改为连续

import networkx as nx
from time import strftime, localtime

def updateFile(file,old_str,new_str):
    file_data = ""
    with open(file, "r", encoding="utf-8") as f:
        for line in f:
            for i in range(len(old_str)):
                if old_str[i] in line:
                    line = line.replace(old_str[i],new_str[i])
            file_data += line
    with open(file,"w",encoding="utf-8") as f:
        f.write(file_data)


def main():
    address = "com-amazon.txt"
    G = nx.read_edgelist(address, create_using=nx.Graph())  # 读取文件，构建网络
    nodes = list(G.nodes)

    nodes_keep = []#保留的结点
    nodes_new = []#被替换的结点
    nodes_need = []

    for node in nodes:
        if int(node)<len(nodes):
            nodes_keep.append(node)
        else:
            nodes_new.append(node)

    nodes_temp = []
    for ij in range(len(nodes)):
        nodes_temp.append(str(ij))
    nodes_need = list(set(nodes_temp) - set(nodes_keep))


    print(len(nodes_need))
    print(len(nodes_new))

    updateFile(r"com-amazon.txt", nodes_new, nodes_need)

    # for iy in range(len(nodes_new)):
    #     if iy%5000 == 0:
    #         print(strftime("%Y-%m-%d %H:%M:%S", localtime()))
    #     updateFile(r"C:\\Users\mkc\Desktop\dataset\\com-dblp.txt", nodes_new[iy], nodes_need[iy])


if __name__ == '__main__':
    main()