如何解决Nivre基于Arc-Eager Transition的依赖解析器:TypeError:预期的str,字节或os.PathLike对象,而不是NoneType
我一直试图在通用TreeBank上运行基于Nivres Transition的解析器https://github.com/dpressel/arcs-py。这是读取树库的代码,
import csv
WORD = 0
POS = 1
HEAD = 2
LABEL = 3
def read_conll_deps(f):
sentences = []
with open(f) as csvfile:
reader = csv.reader(csvfile,delimiter='\t',quoting=csv.QUOTE_NONE)
sentence = []
for row in reader:
if len(row) == 0:
sentence = [tok if tok[HEAD] is not -1 else (tok[WORD],tok[POS],len(sentence),tok[LABEL]) for tok in sentence]
sentences.append(sentence)
sentence = []
continue
sentence.append((row[1].lower(),row[3],int(row[6]) - 1,row[7]))
return sentences
运行主解析器时出现以下错误:
Traceback (most recent call last):
File "C:\Users\LENOVO\Desktop\WorkfromHome\CNN\Nivre_Parser.py",line 565,in
<module>
gold = filter_non_projective(fileio.read_conll_deps(opts.train))
File "C:\Users\LENOVO\Desktop\WorkfromHome\CNN\fileio.py",line 13,in read_conll_deps
with open(f) as csvfile:
TypeError: expected str,bytes or os.PathLike object,not NoneType
版权声明:本文内容由互联网用户自发贡献,该文观点与技术仅代表作者本人。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容, 请发送邮件至 dio@foxmail.com 举报,一经查实,本站将立刻删除。