def load_normal(dir_path):
file_names = os.listdir(dir_path)
lines = []
for file_name in file_names:
temp_lines = data.read_corpus_preprocess(dir_path + file_name)
lines += temp_lines
return lines
def load_Chinese_Literature_NER_RE_data():
data_dir = '../../data/corpus./Chinese-Literature-NER-RE-Dataset/'
file_names = os.listdir(data_dir)
lines = []
for file_name in file_names:
temp_lines = data.read_corpus_preprocess(data_dir + file_name)
lines += temp_lines
#print(lines[0])
for line in lines:
for i in range(len(line[0])):
if line[1][i][1:] == '_Person':
line[1][i] = line[1][i].replace('_Person', '-PER')
elif line[1][i][1:] == '_Location':
line[1][i] = line[1][i].replace('_Location', '-LOC')
elif line[1][i][1:] == '_Organization':
line[1][i] = line[1][i].replace('_Organization', '-ORG')
else:
line[1][i] = "O"
new_data_list = []
for [chars, nertags, _] in lines:
for i in range(len(chars)):
if nertags[i][-3:] in ['GPE', 'GEO']:
nertags[i] = nertags[i][:2] + 'LOC'
new_data_list.append([chars, nertags, data_dir])
#print("load_Chinese_Literature_NER_RE_data ", lines[:3])
return lines
def load_weibo_data():
ner_tag_set = set({'PER', "LOC", "ORG"})
data_dir = '../../data/corpus/Weibo/'
file_names = os.listdir(data_dir)
lines = []
for file_name in file_names:
temp_lines = data.read_corpus_preprocess(data_dir + file_name)
lines += temp_lines
lines = list(
map(lambda x: [list(map(lambda y: y[0], x[0])), x[1], x[2]], lines))
#print(lines[0])
for line in lines:
for i in range(len(line[0])):
if '.NAM' in line[1][i]:
line[1][i] = line[1][i].replace('.NAM', '')
else:
line[1][i] = "O"
new_data_list = []
for [chars, nertags, _] in lines:
for i in range(len(chars)):
if nertags[i][-3:] in ['GPE', 'GEO']:
nertags[i] = nertags[i][:2] + 'LOC'
new_data_list.append([chars, nertags, data_dir])
return lines
def load_ProHiryu_bert_chinese_ner_data():
scr_dir = '../../data/corpus/ProHiryu_bert-chinese-ner_data/'
file_names = os.listdir(scr_dir)
data_list = []
for file_name in file_names:
data_list += data.read_corpus_preprocess(scr_dir + file_name)
#print("数据量是", len(data_list), data_list[:2])
return data_list
def load_crownpku_ner_data():
data_dir = '../../data/corpus/crownpku_ner_data/'
file_names = os.listdir(data_dir)
lines = []
for file_name in file_names:
temp_lines = data.read_corpus_preprocess(data_dir + file_name)
lines += temp_lines
return lines
def load_crownpku_Small_Chinese_Corpus():
scr_dir = '../../data/corpus/crownpku_Small-Chinese-Corpus/'
file_names = os.listdir(scr_dir)
data_list = []
for file_name in file_names:
data_list += data.read_corpus_preprocess(scr_dir + file_name)
#print("数据量是", len(data_list), data_list[:2])
return data_list
def load_boson_data():
scr_dir = '../../data/corpus/boson_data/boson_ner_format.txt'
data_list = data.read_corpus_preprocess(scr_dir)
new_data_list = []
for [chars, nertags, _] in data_list:
for i in range(len(chars)):
#print(nertags[i])
if nertags[i][-4:] not in ['-ORG', '-LOC', '-PER']:
nertags[i] = "O"
new_data_list.append([chars, nertags, scr_dir])
# print("数据量是", len(data_list), new_data_list[:2])
return new_data_list
def load_ner_data_LatticeLSTM():
'''这份数据的命名实体标记与常见的不同，需要替换一下:GPE-LOC, GEO-LOC(https://github.com/yanqiangmiffy/ner-english)'''
scr_dir = '../../data/corpus/ner_data_LatticeLSTM/'
file_names = os.listdir(scr_dir)
data_list = []
for file_name in file_names:
data_list += data.read_corpus_preprocess(scr_dir + file_name)
new_data_list = []
for [chars, nertags, _] in data_list:
for i in range(len(chars)):
if nertags[i][-3:] in ['GPE', 'GEO']:
nertags[i] = nertags[i][:2] + 'LOC'
if nertags[i][0] == "M":
nertags[i] = 'I' + nertags[i][1:]
new_data_list.append([chars, nertags, scr_dir])
#print("数据量是", len(data_list), new_data_list[:2])
return new_data_list
def load_mhcao916_ner_data():
data_dir = '../../data/corpus/mhcao916_ner_data/'
file_names = os.listdir(data_dir)
lines = []
for file_name in file_names:
temp_lines = data.read_corpus_preprocess(data_dir + file_name)
lines += temp_lines
new_data_list = []
for [chars, nertags, _] in lines:
for i in range(len(chars)):
#print(nertags[i])
if nertags[i][-4:] in ['-SCE', '-DLO']:
nertags[i] = nertags[i][0] + '-LOC'
elif nertags[i][-4:] in ['-SCE', '-HOT']:
nertags[i] = nertags[i][0] + '-ORG'
elif nertags[i][-4:] in ['-PER', '-ORG']:
pass
else:
nertags[i] = "O"
new_data_list.append([chars, nertags, data_dir])
#print(new_data_list[:5])
return new_data_list
