NLTK：如何用python获得循环中数组的特定内容？

Question

我正在尝试使用NLTK获取语料库的统计数据，我想知道如何获取特定标记旁边的标记列表。 例如，我希望在标记DTDEF之后有标记列表

我试着按照https://www.nltk.org/book/ch05.html上的教程进行操作，并根据自己的需要进行调整。你知道吗

这里，代码在数组“tags”中存储单词'ny'之后的所有标记，或者我想在标记DTDEF之后存储标记（DTDEF是单词'ny'的标记）。你知道吗

import nltk
from nltk.corpus.reader import TaggedCorpusReader
reader = TaggedCorpusReader('cookbook', r'.*\.pos')
train_sents=reader.tagged_sents()
for sent in train_sents:
    tags = [tag[1] for (word, tag) in nltk.bigrams(sent) if word[0]=='ny']

    #0 is for the word and 1 is for the tag, so tag[0] get you the word and 
    #tag[1] the tag, the same with word[0] and word[1]


fd = nltk.FreqDist(tags)
fd.tabulate()

为了得到我想要的结果，我将代码改为：

import nltk
from nltk.corpus.reader import TaggedCorpusReader
reader = TaggedCorpusReader('cookbook', r'.*\.pos')
train_sents=reader.tagged_sents()
for sent in train_sents:
    #i change the line here
    tags = [tag[1] for (word, tag) in nltk.bigrams(sent) if tag[1]=='DTDEF']

fd = nltk.FreqDist(tags)
fd.tabulate()

我期望在tag DTDEF后面有tag的列表，但是我得到了tag DTDEF的所有出现。 DTDEF公司 150个

所以我尝试了这个，但是python的问题是我不能做这样的事情：

import nltk
from nltk.corpus.reader import TaggedCorpusReader
reader = TaggedCorpusReader('cookbook', r'.*\.pos')
train_sents=reader.tagged_sents()
tags=[]
count=0
for sent in train_sents:
    for (word,tag) in sent:
        #if tag is DTDEF i want to get the tag after it
        if tag=="DTDEF":
            tags[count]=tag[acutalIndex+1]
            count+=1


fd = nltk.FreqDist(tags)
fd.tabulate()

所以这就是我提问的原因。你知道吗

提前感谢您的回答和建议。你知道吗

Answer 1

感谢#CrazySqueak的帮助，我使用了他的代码并编辑了一些部分来获得：

import nltk
from nltk.corpus.reader import TaggedCorpusReader
reader = TaggedCorpusReader('cookbook', r'.*\.pos')
train_sents=reader.tagged_sents()
tags = []
foundit=False
for sent in train_sents:
    #i change the line here
    for (word,tag) in nltk.bigrams(sent):
        if foundit: #If the entry is after 'DTDEF'
            tags.append(tag[1]) #Add it to the resulting list of tags, i change
                                #tag [1] instead, if you use only tag, it will 
                                #store not only the tag but the word as well 
            #of foundit
            foundit=False #I need to make it false again, cause it will store again even 
                          #if the tag is != of DTDEF
        if tag[1]=='DTDEF': #If the entry is 'DTDEF'
            foundit=True #Set the 'After DTDEF' flag.

fd = nltk.FreqDist(tags)
fd.tabulate()

再次感谢你的建议和回答。你知道吗

Answer 2

我不是100%确定我能理解，但是如果您希望在一个特定条目之后获得列表中的所有条目，最简单的方法是：

foundthing=False
result = []
for i in list:
    if foundthing:
        result.append(i)
    if i == "Thing I'm Looking For":
        foundthing = True

将此添加到代码中会导致：

import nltk
from nltk.corpus.reader import TaggedCorpusReader
reader = TaggedCorpusReader('cookbook', r'.*\.pos')
train_sents=reader.tagged_sents()
tags = []
foundit=False
for sent in train_sents:
    #i change the line here
    for (word,tag) in nltk.bigrams(sent):
        if foundit: #If the entry is after 'DTDEF'
            tags.append(foundit) #Add it to the resulting list of tags.
        if tag[1]=='DTDEF': #If the entry is 'DTDEF'
            foundit=True #Set the 'After DTDEF' flag.

fd = nltk.FreqDist(tags)
fd.tabulate()

希望这有帮助。你知道吗