我具有以下功能,该功能接收单词标记列表,以WordNet可读的格式收集词性标记,并使用它来对每个标记进行词形化-我将其应用于单词标记列表:
from nltk import pos_tag
from nltk.stem import WordNetLemmatizer
from nltk.corpus import wordnet as wn
def getWordNetPOS (POStag):
def is_noun(POStag):
return POStag in ['NN', 'NNS', 'NNP', 'NNPS']
def is_verb(POStag):
return POStag in ['VB', 'VBD', 'VBG', 'VBN', 'VBP', 'VBZ']
def is_adverb(POStag):
return POStag in ['RB', 'RBR', 'RBS']
def is_adjective(POStag):
return POStag in ['JJ', 'JJR', 'JJS']
if is_noun(POStag):
return wn.NOUN
elif is_verb(POStag):
return wn.VERB
elif is_adverb(POStag):
return wn.ADV
elif is_adjective(POStag):
return wn.ADJ
else:
# if not noun, verb, adverb or adjective, return noun
return wn.NOUN
# lemmatize word tokens
def lemmas (wordtokens):
lemmatizer = WordNetLemmatizer()
POStag = pos_tag(wordtokens)
wordtokens = [lemmatizer.lemmatize(token[0], getWordNetPOS(token[1]))
for token in POStag]
return wordtokens
lemmatizedList = []
mylist = [['this','is','my','first','sublist'],['this','is','my','second','sublist']]
for ls in mylist:
x = lemmas(ls)
lemmatizedList.append(x)
我想找到一种方法,以将限制词法限制为设置长度(即2)的令牌,但至关重要的是,我还想保留小于此阈值的任何单词的原始形式。我得到的最接近的是在if len(token[0])>2
函数内的wordtokens
列表理解的末尾添加lemmas
,但这仅返回经过词法化的标记。同样,我尝试在else token for token in POStag
语句后添加类似于if
的内容,但是出现语法错误。为了清楚起见,这就是我的意思:
wordtokens = [lemmatizer.lemmatize(token[0], getWordNetPOS(token[1]))
for token in POStag if len(token[0])>2
else token for token in POStag]
我希望它是一个简单的错误,并希望在我自己的身上留下一些Python盲点。
这实际上只是一个“盲点”。