wangsong commited on
Commit
8b359ff
·
1 Parent(s): b8d6196

Upload jd21.py

Browse files
Files changed (1) hide show
  1. jd21.py +102 -0
jd21.py ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ from datasets import Value, ClassLabel,Sequence
3
+ import datasets
4
+
5
+
6
+ _JD21_CITATION = """\
7
+
8
+ """
9
+
10
+ _JD21_DESCRIPTION = """\
11
+ GLUE, the General Language Understanding Evaluation benchmark
12
+ (https://gluebenchmark.com/) is a collection of resources for training,
13
+ evaluating, and analyzing natural language understanding systems.
14
+ """
15
+
16
+ class JD21Config(datasets.BuilderConfig):
17
+
18
+ def __init__(
19
+ self,
20
+ text_features,
21
+ label_column,
22
+ data_url,
23
+ data_dir,
24
+ citation,
25
+ url,
26
+ label_classes=None,
27
+ process_label=lambda x: x,
28
+ **kwargs,
29
+ ):
30
+ super(JD21Config, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
31
+
32
+ self.text_features = text_features
33
+ self.label_column = label_column
34
+ self.label_classes = label_classes
35
+ self.data_url = data_url
36
+ self.data_dir = data_dir
37
+ self.citation = citation
38
+ self.url = url
39
+ self.process_label = process_label
40
+
41
+ class JD21(datasets.GeneratorBasedBuilder):
42
+ domain_list = ['褪黑素', '维生素', '无线耳机', '蛋白粉', '游戏机', '电视', 'MacBook', '洗面奶', '智能手表', '吹风机', '小米手机', '红米手机', '护肤品',
43
+ '电动牙刷', 'iPhone', '海鲜', '酒', '平板电脑', '修复霜', '运动鞋', '智能手环']
44
+
45
+ BUILDER_CONFIGS = [
46
+ JD21Config(name=domain_name,
47
+ description= f'comments of JD {domain_name}.',
48
+ text_features={'sentence':'sentence', 'domain':'domain'},
49
+ label_classes=['POS','NEG'],
50
+ label_column='label',
51
+ citation="",
52
+ data_dir= r"D:\Personal\CodeBase\Continual Learning\continual-learning-framework-for-NLP\datasets\jd21\\",
53
+ data_url = "",
54
+ url='https://github.com/ws719547997/LNB-DA')
55
+ for domain_name in domain_list
56
+ ]
57
+
58
+ def _info(self):
59
+ features = {'id':Value(dtype='int32', id=None),
60
+ 'domain':Value(dtype='string', id=None),
61
+ 'label':ClassLabel(num_classes=2, names=['POS', 'NEG'], names_file=None, id=None),
62
+ 'rank':Value(dtype='int32', id=None),
63
+ 'sentence':Value(dtype='string', id=None)}
64
+
65
+ return datasets.DatasetInfo(
66
+ description=_JD21_DESCRIPTION,
67
+ features=datasets.Features(features),
68
+ homepage=self.config.url,
69
+ citation=self.config.citation + "\n" + _JD21_CITATION,
70
+ )
71
+
72
+ def _split_generators(self, dl_manager):
73
+ test_file = rf'{self.config.data_dir}data\test\{self.config.name}.txt'
74
+ dev_file = rf'{self.config.data_dir}data\dev\{self.config.name}.txt'
75
+ train_file = rf'{self.config.data_dir}data\train\{self.config.name}.txt'
76
+ return [datasets.SplitGenerator(name=datasets.Split.TEST,
77
+ gen_kwargs={
78
+ "data_file": test_file,
79
+ "split": "test",
80
+ },),
81
+ datasets.SplitGenerator(name=datasets.Split.VALIDATION,
82
+ gen_kwargs={
83
+ "data_file": dev_file,
84
+ "split": "dev",
85
+ },),
86
+ datasets.SplitGenerator(name=datasets.Split.TRAIN,
87
+ gen_kwargs={
88
+ "data_file": train_file,
89
+ "split": "train",
90
+ },)]
91
+
92
+ def _generate_examples(self, data_file, split):
93
+ with open(data_file, 'r', encoding='utf-8') as f:
94
+ for line in f:
95
+ lin = line.strip()
96
+ if not lin:
97
+ continue
98
+ lin_sp = lin.split('\t')
99
+ if len(lin_sp) < 5:
100
+ continue
101
+ # id, {example}
102
+ yield lin_sp[0], {'sentence':lin_sp[4],'domain':lin_sp[1], 'label':lin_sp[2], 'id':lin_sp[0], 'rank':lin_sp[3]}