python实现造100组图结构数据p（造SimGNN数据集）

python实现造100组图结构数据p（造SimGNN数据集）

import random

import numpy as np
import torch
from rdkit import Chem
import networkx as nx
import os
import json


def Get_GED(arr1, arr2):
    G1 = nx.DiGraph()
    G2 = nx.DiGraph()
    max1 = -1
    max2 = -1

    for i in range(0, len(arr1)):
        max1 = max(max1, arr1[i][0], arr1[i][1])
    for i in range(0, len(arr2)):
        max2 = max(max2, arr2[i][0], arr2[i][1])
    for i in range(0, max1 + 1):
        G1.add_node(i)

    for i in range(0, max2 + 1):
        G2.add_node(i)

    for i in range(0, len(arr1)):
        G1.add_edges_from([(arr1[i][0], arr1[i][1])])
        G1.add_edges_from([(arr1[i][1], arr1[i][0])])

    for i in range(0, len(arr2)):
        G2.add_edges_from([(arr2[i][0], arr2[i][1])])
        G2.add_edges_from([(arr2[i][1], arr2[i][0])])
    # 返回两个图的GED
    return nx.graph_edit_distance(G1, G2)

def new_graph(node_num):
    max_edge_num = node_num * (node_num - 1)
    arr = np.zeros((node_num, node_num))
    node_du = np.zeros(node_num, dtype=int)
    edge_arr = []

    for i in range(0, node_num):
        for j in range(0, node_num):
            if arr[i][j] == 1 or arr[j][i] == 1:
                continue
            if i == j:
                continue
            float = np.random.uniform(0, 1)
            if float <= 0.5:
                edge_arr.append(i)
                edge_arr.append(j)
                arr[i][j] = arr[j][i] = 1
    tmp=edge_arr
    prin_edge_arr = np.array(edge_arr)
    for i in prin_edge_arr:
        node_du[i] += 1


    for i in range(0, len(node_du)):
        if node_du[i] == 0:
            tmp.append(i)
            x=0
            while x == i:
                x=random.randint(0, len(node_du)-1)
            tmp.append(x)
            node_du[i] += 1
            node_du[x] += 1


    prin_edge_arr = np.array(tmp).reshape(-1, 2)
    return node_du, prin_edge_arr


def new_ged():
    node_num1 = np.random.randint(3, 8)
    nodedu1, edge1 = new_graph(node_num1)
    node_num2 = np.random.randint(3, 8)
    nodedu2, edge2 = new_graph(node_num2)

    Ged = Get_GED(edge1, edge2)

    return Ged, nodedu1, nodedu2, edge1, edge2


for i in range(0, 100):
    path = '../dataset/newtrain/' + str(i) + '.txt'
    f = open(path, 'w')

    Ged, du1, du2, e1, e2 = new_ged()
    dd2=np.array(du2,dtype=str)
    dd1 = np.array(du1, dtype=str)
    mp = {"graph_1": e1, "ged": Ged, "graph_2": e2, "labels_2": dd2, "labels_1": dd1}
    ans = str(mp)
    ans = ans.replace("\n", "")
    ans = ans.replace("       ", " ")
    ans = ans.replace("\'", "\"")
    ans = ans.replace("array(", "")
    ans = ans.replace(")", "")
    ans = ans.replace(", dtype=\"<U11\"", "")


    print(ans)
    print(ans, file=f)

    f.close()
for i in range(0, 100):
    pat = '../dataset/newtrain/' + str(i) + '.txt'
    filename = os.path.split(pat)
    fliesion = filename[1]
    newname = str(i) + '.json'
    os.rename('../dataset/newtrain/' + fliesion, '../dataset/newtrain/' + newname)