关于图学习node2vec代码问题的请教

Question

从github上下载了开源的node2vec代码，其中node2vec部分如下图所示：

# -*- coding: utf-8 -*-
"""
@Time ： 2021/12/17 21:44
@Author ：KI
@File ：node2vec.py
@Motto：Hungry And Humblec
"""
import numpy as np
import sys
sys.path.append('../')
import numpy.random as npr
from gensim.models import Word2Vec
 
class node2vec:
    def __init__(self, args, G):
        self.G = G
        self.args = args
        self.init_transition_prob()
 
#以下是正常调用
    def init_transition_prob(self):
        """
        :return:Normalized transition probability matrix
        """
        g = self.G
        nodes_info, edges_info = {}, {}
        for node in g.nodes():
            nbs = sorted(g.neighbors(node))
            probs = [g[node][n] for n in nbs]
            # Normalized
            norm = sum(probs)
            normalized_probs = [float(n) / norm for n in probs]
            nodes_info[node] = self.alias_setup(normalized_probs)
 
        for edge in g.edges():
            # directed graph
            if g.is_directed():
                edges_info[edge] = self.get_alias_edge(edge[0], edge[1])
            # undirected graph
            else:
                edges_info[edge] = self.get_alias_edge(edge[0], edge[1])
                edges_info[(edge[1], edge[0])] = self.get_alias_edge(edge[1], edge[0])
 
        self.nodes_info = nodes_info
        self.edges_info = edges_info
 
#对于不是第二次采样的情况，需要利用get_alias_edge来得到一条边( t , v ) (t, v)(t,v)中v
    # v到其邻居节点转移概率的Alias数组和Prob数组
    def get_alias_edge(self, t, v):
        """
        Get the alias edge setup lists for a given edge.
        """
        g = self.G
        unnormalized_probs = []
        for v_nbr in sorted(g.neighbors(v)):
            if v_nbr == t:
                unnormalized_probs.append(g[v][v_nbr]['weight'] / self.args.p)
            elif g.has_edge(v_nbr, t):
                unnormalized_probs.append(g[v][v_nbr]['weight'])
            else:
                unnormalized_probs.append(g[v][v_nbr]['weight'] / self.args.q)
        norm_const = sum(unnormalized_probs)
        normalized_probs = [float(u_prob) / norm_const for u_prob in unnormalized_probs]
 
        return self.alias_setup(normalized_probs)
 
 
 
#以下是Alias Sample具体实现
    def alias_setup(self, probs):
        """
        :probs: probability
        :return: Alias and Prob
        """
        K = len(probs)
        q = np.zeros(K)
        J = np.zeros(K, dtype=np.int)
        # Sort the data into the outcomes with probabilities
        # that are larger and smaller than 1/K.
        smaller = []  #
        larger = []  #
        for kk, prob in enumerate(probs):
            q[kk] = K * prob  #
            if q[kk] < 1.0:
                smaller.append(kk)
            else:
                larger.append(kk)
 
        # Loop though and create little binary mixtures that
        # appropriately allocate the larger outcomes over the
        # overall uniform mixture.
 
        #
        while len(smaller) > 0 and len(larger) > 0:
            small = smaller.pop()
            large = larger.pop()
 
            J[small] = large  #
            q[large] = q[large] - (1.0 - q[small])  #
 
            if q[large] < 1.0:
                smaller.append(large)
            else:
                larger.append(large)
 
        return J, q
 
#构造采样函数
#kk或者J[kk]就是我们最终采样的结果
    def alias_draw(self, J, q):
        """
        in: Prob and Alias
        out: sampling results
        """
        K = len(J)
        # Draw from the overall uniform mixture.
        kk = int(np.floor(npr.rand() * K))  # random
 
        # Draw from the binary mixture, either keeping the
        # small one, or choosing the associated larger one.
        if npr.rand() < q[kk]:  # compare
            return kk
        else:
            return J[kk]
 
#有了转移概率以及采样策略后，我们就能轻松实现node2vecWalk了：
    def node2vecWalk(self, u):
        g = self.G
        walk = [u]
        nodes_info, edges_info = self.nodes_info, self.edges_info
        while len(walk) < self.args.l:
            curr = walk[-1]
            v_curr = sorted(g.neighbors(curr))
            if len(v_curr) > 0:
                if len(walk) == 1:
                    # print(adj_info_nodes[curr])
                    # print(alias_draw(adj_info_nodes[curr][0], adj_info_nodes[curr][1]))
                    walk.append(v_curr[self.alias_draw(nodes_info[curr][0], nodes_info[curr][1])])
                else:
                    prev = walk[-2]
                    ne = v_curr[self.alias_draw(edges_info[(prev, curr)][0], edges_info[(prev, curr)][1])]
                    walk.append(ne)
            else:
                break
 
        return walk
 
#以下为Learn Feature部分：
#walks用来存储随机游走，先初始化为空
#一共要进行r rr次循环，每一次循环要为图中每个节点都生成一个长度为l ll的游走序列
#第iter次循环中对所有节点都利用node2vec算法生成一个随机游走序列walk，然后将其加入walks
#得到所有节点的r rr个随机游走序列后，利用SGD方法得到每一个节点的向量表示
 
    def learning_features(self):
        walks = []
        g = self.G
        nodes = list(g.nodes())
        for t in range(self.args.r):
            np.random.shuffle(nodes)
            for node in nodes:
                walk = self.node2vecWalk(node)
                walks.append(walk)
        # embedding
        walks = [list(map(str, walk)) for walk in walks]
        # print(walks[0])
        # print(walks[1])
        model = Word2Vec(sentences=walks, vector_size=self.args.d, window=self.args.k, min_count=0, sg=1, workers=3)
        f = model.wv
        res = [f[x] for x in nodes]
        return res
 
 
 
 

然而总是报错，提示为：

问题出在line 32, in init_transition_prob，即 norm = sum(probs)部分

请问一下是什么原因，如何解决？感谢！

Answer 1

引自强大而免费的微信小程序：皆我百晓生。

---

根据您提供的信息，问题出现在norm = sum(probs)这一行，报错信息显示无法从node2vec模块中导入Node2Vec。出现这种情况可能有以下几个可能性：

1. 您没有正确导入所需的node2vec模块。请确认您是否从正确的路径导入了node2vec模块，并且路径设置正确。
2. 您可能没有安装所需的依赖库。请确保您安装了gensim库和其他可能需要的依赖库。可以通过运行pip install gensim来安装gensim库。
3. 您的node2vec模块可能已损坏或不完整。尝试重新下载并替换node2vec模块，确保您下载的是最新版本，并且没有损坏的文件。

解决这个问题的方法是确保导入正确的模块，并正确安装所需的依赖库。如果问题仍然存在，您可以检查代码中是否有其他错误或问题。