# 把词语表通过OneHot转换为特征向量
# 输入：一维词表，定义好了的向量词典
def ConvertWordsToTensor(words, dic):
l = len(dic)
z = np.zeros([l, 1], dtype=np.float)
for w in words:
if w in dic:
z[dic[w]] = 1
else:
z[dic['UNKNOWN']] = 1
return z
fr = FR.OneHotBuilder(R'data/1998-01-2003版-带音.txt', "19980101", "19980120")
# 把y[i]建成一个numpy向量列表？
X = []
for w in fr.linkedWord:
z = ConvertWordsToTensor(w, fr.oneHotDic)
X.append(torch.from_numpy(z))
Y = []
for w in fr.linkedWord:
if w in fr.entityWord:
# Y.append(1)
Y.append(torch.tensor([1.]))
else:
# Y.append(0)
