问题描述
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
from nltk.corpus import wordnet as wn
#from itertools import product
#variables that will be used
database_word_synset=[]
uploaded_sentence_synset=[]
uploaded_sentence_words_tokenized=[]
filtered_uploaded_sentences = []
database_sentence_words_tokenized=[]
filtered_database_sentence=[]
database_sentence_synset=[]
word_check=[0.0]
sentence_check=[0.0]
count_sentence=0
count_word=0
not_fond=0
#the given data
uploaded_sentence=" The issue of text semantics,such as word semantics and sentence semantics has received increasing attentions in recent years. However,rare research focuses on the document-level semantic matching due to its complexity. Long documents usually have sophisticated structure and massive information,which causes hardship to measure their semantic similarity. The semantic similarity between words,sentences,texts,and documents is widely studied in varIoUs fields,including natural language processing,document semantic comparison,artificial intelligence,semantic web,and semantic search engines. "
database_word=["car","complete",'run',"sleep"]
database_sentence="the earth is round not flat"
stopwords = stopwords.words('english')
uploaded_sentence_words_tokenized = word_tokenize(uploaded_sentence)
#filtering the sentence and synset
for word in uploaded_sentence_words_tokenized:
if word not in stopwords:
filtered_uploaded_sentences.append(word)
print (filtered_uploaded_sentences)
for sentences_are in filtered_uploaded_sentences:
uploaded_sentence_synset.append(wn.synsets(sentences_are))
print(uploaded_sentence_synset)
#for finding similrity in the words
for databasewords in database_word:
database_word_synset.append(wn.synsets(databasewords))
print(database_word_synset)
words_list_synset=list()
for t in database_word_synset:
for x in t:
words_list_synset.append(x)
print(words_list_synset)
#removing empty list element and making single dimension list
removing_empty_list_uploaded_sentence=list()
removing_empty_list_uploaded_sentence = [x for x in uploaded_sentence_synset if x != []]
up_list_sentence=list()
for t in removing_empty_list_uploaded_sentence:
for x in t:
up_list_sentence.append(x)
print(up_list_sentence)
#the similarity main function for words
#sims=[]
#for sense1,sense2 in product(database_word_synset,up_list_sentence):
# d = wn.wup_similarity(sense1,sense2)
# sims.append(d)
#print (sims)
#word_found=list()
for data in words_list_synset:
for sen in up_list_sentence :
if wn.wup_similarity(data,sen) is None or wn.wup_similarity(data,sen) <0.70:
not_fond=not_fond+1
else:
count_word=count_word+1
print (word_check)
print("\n words that are not found :",not_fond)
print("\n words that are found :",count_word)
#for finding similrity in the sentence
database_sentence_words_tokenized=word_tokenize(database_sentence)
for word in database_sentence_words_tokenized:
if word not in stopwords:
filtered_database_sentence.append(word)
print(filtered_database_sentence)
for sentence_synset in filtered_database_sentence:
database_sentence_synset.append(wn.synsets(sentence_synset))
print(database_sentence_synset)
#removing empty list element and making single dimension list
removing_empty_list_db=list()
removing_empty_list_db = [x for x in database_sentence_synset if x != []]
db_list_sentence=list()
for t in removing_empty_list_db:
for x in t:
db_list_sentence.append(x)
print(db_list_sentence)
#the similarity main function for sentence
for db_sentence in db_list_sentence:
for upl_sentence in up_list_sentence:
sentence_check.append(wn.wup_similarity(db_sentence,upl_sentence))
for sentence_checks in sentence_check:
if sentence_checks is None or sentence_checks <0.70:
not_fond=not_fond+1
else:
count_sentence=count_sentence+1
print (sentence_check)
print("\n words that are not found :",count_sentence)
在此项目中,我们使用chaquopy在我们的android项目中使用python,但是它存在一些问题,例如在导入库中,我还分别安装了Nltk,wordnet,停止单词和单词标记化,但是我无法访问这些库python文件,如果我们安装我们的应用程序,它将崩溃。
if (! Python.isstarted()) {
Python.start(new Androidplatform(this));
Python py = Python.getInstance();
final PyObject pyobj = py.getModule("sum");
b2.setonClickListener(new View.OnClickListener() {
@Override
public void onClick(View view) {
if (path==null) {
Toast.makeText(documentupload.this," plz upload the doc",Toast.LENGTH_SHORT).show();
//upload.setText(path);
// Intent intent= new Intent(documentupload.this,result.class);
//startActivity(intent);
}
else {
PyObject obj = pyobj.callAttr("main",Words.toString());
upload.setText(obj.toString());
Toast.makeText(documentupload.this,"uploaded" + Words,Toast.LENGTH_LONG).show();
// Toast.makeText(documentupload.this,Toast.LENGTH_LONG).show();
}
}
});
解决方法
我假设调用wn.synsets
时发生了崩溃?这是我看到的堆栈跟踪:
File "/data/user/0/com.chaquo.python.pkgtest3/files/chaquopy/AssetFinder/requirements/nltk/corpus/util.py",line 120,in __getattr__
File "/data/user/0/com.chaquo.python.pkgtest3/files/chaquopy/AssetFinder/requirements/nltk/corpus/util.py",line 85,in __load
File "/data/user/0/com.chaquo.python.pkgtest3/files/chaquopy/AssetFinder/requirements/nltk/corpus/util.py",line 80,in __load
File "/data/user/0/com.chaquo.python.pkgtest3/files/chaquopy/AssetFinder/requirements/nltk/data.py",line 585,in find
LookupError:
**********************************************************************
Resource [93mwordnet[0m not found.
我认为“ wordnet”和“ corpus” pip软件包与nltk没有任何关系。相反,您应该使用nltk.download
安装它们,就像错误消息所述一样。
由于仿真器错误,您可能需要按this answer中所述循环调用nltk.download
。