按照添加边的顺序获取networkx中的连接组件

Question

doc2 中的每个句子都显示为图表。现在，边以 s-o-v 的形式从各自的 subject_list、object_list 和 verb_list 添加。 我试图显示连接的组件。但是它显示句子的顺序不是添加边的顺序。

# This Python file uses the following encoding: utf-8
%matplotlib notebook
import codecs
import itertools
import re
import networkx as nx
import matplotlib.pyplot as pl
from matplotlib.font_manager import FontProperties


prop = FontProperties()
graph = nx.Graph()

labels = {}
each_one = []
list_of_sentences = []
subject_list = []
object_list = []
verb_list = []
newDict = {}

with codecs.open('doc2.txt', encoding='utf-8') as f:
    text = f.read()
    sentences = re.split(r' *[\.\?!][\'"\)\]]* *', text)
    for stuff in sentences:
        list_of_sentences.append(stuff)

new_list_of_sentences = []
for d in list_of_sentences:
    s = d.replace(u'वतीन', '').replace(u'आनी', '').replace(u'हिणें', '').replace(',', '')
    new_list_of_sentences.append(s)

f = open('doc2_tag.txt', 'r')
for line in f:
    k, v = line.strip().split('/')
    newDict[k.strip().decode('utf-8')] = v.strip()
f.close()

for sentence in new_list_of_sentences:
    a = b = c = ""
    sentence_word_list = sentence.split()
    for word in sentence_word_list:
        if newDict[word] == 'N-NNP':
            a += word + " "
        if newDict[word] == 'N-NN':
            b += word + " "
        if newDict[word] == 'JJ':
            b += word + " "
        if newDict[word] == 'QT-QTC':
            b += word + " "
        if newDict[word] == 'RB':
            b += word + " "
        if newDict[word] == 'N-NST':
            b += word + " "
        if newDict[word] == 'PR-PRP':
            b += word + " "
        if newDict[word] == 'PSP':
            b += word + " "
        if newDict[word] == 'CC-CCD':
            b += word + " "
        if newDict[word] == 'V-VM-VF':
            c += word + " "

    subject_list.append(a)
    object_list.append(b)
    verb_list.append(c)


konkani_dict = {u'सनरायझर्साक': u'सनरायझर्स', u'सनरायझर्सान': u'सनरायझर्स', u'सनरायझर्साच्या': u'सनरायझर्स'}
for idx, sub in enumerate(subject_list):
    temp_list = sub.split(" ")
    for i in temp_list:
        if i in konkani_dict:
            new_sub = sub.replace(i, konkani_dict[i])
            subject_list[idx] = new_sub


for s in subject_list:
    if s is not "":
        graph.add_node(s)
        labels[s] = s


for o in object_list:
    if o is not "":
        graph.add_node(o)
        labels[b] = b

for v in verb_list:
    if v is not "":
        graph.add_node(v)
        labels[v] = v



for (s, o, v) in zip(subject_list, object_list, verb_list):
    if s and o is not "":
        graph.add_edge(s, o)
    if o and v is not "":
        graph.add_edge(o, v)

pos=nx.spring_layout(graph,k=0.15,iterations=20)
nx.draw(graph, with_labels = True, font_family = "Nirmala UI", node_size = 40, font_size = 9 ,node_color = "darkblue")

pl.show()

sentences=[]
for component in nx.connected_components(graph):

    g=(
        filter(
            lambda x: x[0] in component and x[1] in component,
            graph.edges
        )
    )

    p=[]
    p= ''.join(item for tuple_ in g for item in tuple_)
    print p
    sentences.append(p)
    print sentences
output=[]
for i in sentences:
    inputWords = i.split(" ") 

    inputWords=inputWords[-1::-1]
    output = ' '.join(inputWords) 
    print output

预期的输出是这样的：

शिखर धवनान सगळ्यांत चड ४५ धांवड्यो  केल्यो ,
सनरायझर्स दीपर हुडा जैतांत पर्जळ्ळो

这是我得到的输出：sentences displayed

Answer 1

networkx 不存储创建节点/边的顺序，因为这些信息大多是无用的。如果您想拥有此信息，则应手动添加。例如，在您的程序中（对于边缘）：

edge_index = 0

for (s, o, v) in zip(subject_list, object_list, verb_list):
    if s and o is not "":
        graph.add_edge(s, o, index=edge_index)
        edge_index += 1
    if o and v is not "":
        graph.add_edge(o, v, index=edge_index)
        edge_index += 1

那么你应该打印排序的边缘：

sorted(  # Sorted list of edges
    list(g.edges.data('index')),  # With 'index' data
    key=lambda x: x[2]  # Sorted by 'index' data
)```