Pierre Lepagnol commited on
Commit
dcf22df
1 Parent(s): 4fa2464

Add and check other datasets

Browse files
Files changed (2) hide show
  1. README.md +28 -16
  2. WRENCH.py +31 -19
README.md CHANGED
@@ -71,60 +71,72 @@ dataset_info:
71
  dtype: string
72
  - name: label
73
  dtype: int8
 
 
 
 
74
  - name: weak_labels
75
  sequence: int8
76
  splits:
77
  - name: train
78
- num_bytes: 2056619
79
  num_examples: 8430
80
  - name: validation
81
- num_bytes: 218124
82
  num_examples: 920
83
  - name: test
84
- num_bytes: 1087330
85
  num_examples: 4673
86
  download_size: 11036213
87
- dataset_size: 3362073
88
  - config_name: chemprot
89
  features:
90
  - name: text
91
  dtype: string
92
  - name: label
93
  dtype: int8
 
 
 
 
94
  - name: weak_labels
95
  sequence: int8
96
  splits:
97
  - name: train
98
- num_bytes: 3103153
99
  num_examples: 12861
100
  - name: validation
101
- num_bytes: 389127
102
  num_examples: 1607
103
  - name: test
104
- num_bytes: 386810
105
  num_examples: 1607
106
  download_size: 15743249
107
- dataset_size: 3879090
108
  - config_name: semeval
109
  features:
110
  - name: text
111
  dtype: string
112
  - name: label
113
  dtype: int8
 
 
 
 
114
  - name: weak_labels
115
  sequence: int8
116
  splits:
117
  - name: train
118
- num_bytes: 494734
119
  num_examples: 1749
120
  - name: validation
121
- num_bytes: 50525
122
  num_examples: 178
123
  - name: test
124
- num_bytes: 171950
125
  num_examples: 600
126
  download_size: 2295058
127
- dataset_size: 717209
128
  - config_name: sms
129
  features:
130
  - name: text
@@ -179,16 +191,16 @@ dataset_info:
179
  sequence: int8
180
  splits:
181
  - name: train
182
- num_bytes: 634070
183
  num_examples: 4965
184
  - name: validation
185
- num_bytes: 63436
186
  num_examples: 500
187
  - name: test
188
- num_bytes: 59113
189
  num_examples: 500
190
  download_size: 2277067
191
- dataset_size: 756619
192
  - config_name: youtube
193
  features:
194
  - name: text
 
71
  dtype: string
72
  - name: label
73
  dtype: int8
74
+ - name: entity1
75
+ dtype: string
76
+ - name: entity2
77
+ dtype: string
78
  - name: weak_labels
79
  sequence: int8
80
  splits:
81
  - name: train
82
+ num_bytes: 2318545
83
  num_examples: 8430
84
  - name: validation
85
+ num_bytes: 246252
86
  num_examples: 920
87
  - name: test
88
+ num_bytes: 1229627
89
  num_examples: 4673
90
  download_size: 11036213
91
+ dataset_size: 3794424
92
  - config_name: chemprot
93
  features:
94
  - name: text
95
  dtype: string
96
  - name: label
97
  dtype: int8
98
+ - name: entity1
99
+ dtype: string
100
+ - name: entity2
101
+ dtype: string
102
  - name: weak_labels
103
  sequence: int8
104
  splits:
105
  - name: train
106
+ num_bytes: 3474637
107
  num_examples: 12861
108
  - name: validation
109
+ num_bytes: 435850
110
  num_examples: 1607
111
  - name: test
112
+ num_bytes: 434031
113
  num_examples: 1607
114
  download_size: 15743249
115
+ dataset_size: 4344518
116
  - config_name: semeval
117
  features:
118
  - name: text
119
  dtype: string
120
  - name: label
121
  dtype: int8
122
+ - name: entity1
123
+ dtype: string
124
+ - name: entity2
125
+ dtype: string
126
  - name: weak_labels
127
  sequence: int8
128
  splits:
129
  - name: train
130
+ num_bytes: 532785
131
  num_examples: 1749
132
  - name: validation
133
+ num_bytes: 54373
134
  num_examples: 178
135
  - name: test
136
+ num_bytes: 184826
137
  num_examples: 600
138
  download_size: 2295058
139
+ dataset_size: 771984
140
  - config_name: sms
141
  features:
142
  - name: text
 
191
  sequence: int8
192
  splits:
193
  - name: train
194
+ num_bytes: 610244
195
  num_examples: 4965
196
  - name: validation
197
+ num_bytes: 61048
198
  num_examples: 500
199
  - name: test
200
+ num_bytes: 56479
201
  num_examples: 500
202
  download_size: 2277067
203
+ dataset_size: 727771
204
  - config_name: youtube
205
  features:
206
  - name: text
WRENCH.py CHANGED
@@ -24,6 +24,8 @@ class WrenchSequenceTaggingConfig(datasets.BuilderConfig):
24
  class Wrench(datasets.GeneratorBasedBuilder):
25
  """WRENCH classification dataset."""
26
 
 
 
27
  SEQUENCETAGGING_DATASET = [
28
  {"name": "conll", "dataset_path": "./conll"},
29
  {"name": "laptopreview", "dataset_path": "./laptopreview"},
@@ -47,10 +49,7 @@ class Wrench(datasets.GeneratorBasedBuilder):
47
  {"name": "yelp", "dataset_path": "./yelp"},
48
  {"name": "youtube", "dataset_path": "./youtube"},
49
  ]
50
- # configs_st = [
51
- # WrenchSequenceTaggingConfig(name=i["name"], dataset_path=i["dataset_path"])
52
- # for i in SEQUENCETAGGING_DATASET
53
- # ]
54
  configs_clf = [
55
  WrenchClassifConfig(name=i["name"], dataset_path=i["dataset_path"]) for i in CLASSIF_DATASETS
56
  ]
@@ -58,7 +57,7 @@ class Wrench(datasets.GeneratorBasedBuilder):
58
 
59
  def _info(self):
60
  name = self.config.name
61
- if name == "spouse":
62
  return datasets.DatasetInfo(
63
  features=datasets.Features(
64
  {
@@ -105,7 +104,7 @@ class Wrench(datasets.GeneratorBasedBuilder):
105
  with open(filepath, encoding="utf-8") as f:
106
  json_data = json.load(f)
107
 
108
- if name == "spouse":
109
  for idx in json_data:
110
  data = json_data[idx]
111
 
@@ -123,16 +122,29 @@ class Wrench(datasets.GeneratorBasedBuilder):
123
  "weak_labels": weak_labels,
124
  }
125
  else:
126
- for idx in json_data:
127
- # print(json_data[idx])
128
- data = json_data[idx]
129
-
130
- text = data["data"]["text"]
131
- weak_labels = data["weak_labels"]
132
- label = data["label"]
133
-
134
- yield int(idx), {
135
- "text": text,
136
- "label": label,
137
- "weak_labels": weak_labels,
138
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  class Wrench(datasets.GeneratorBasedBuilder):
25
  """WRENCH classification dataset."""
26
 
27
+ ENTITY_RELATED_DATASETS = ["spouse", "cdr", "semeval", "chemprot"]
28
+
29
  SEQUENCETAGGING_DATASET = [
30
  {"name": "conll", "dataset_path": "./conll"},
31
  {"name": "laptopreview", "dataset_path": "./laptopreview"},
 
49
  {"name": "yelp", "dataset_path": "./yelp"},
50
  {"name": "youtube", "dataset_path": "./youtube"},
51
  ]
52
+
 
 
 
53
  configs_clf = [
54
  WrenchClassifConfig(name=i["name"], dataset_path=i["dataset_path"]) for i in CLASSIF_DATASETS
55
  ]
 
57
 
58
  def _info(self):
59
  name = self.config.name
60
+ if name in self.ENTITY_RELATED_DATASETS:
61
  return datasets.DatasetInfo(
62
  features=datasets.Features(
63
  {
 
104
  with open(filepath, encoding="utf-8") as f:
105
  json_data = json.load(f)
106
 
107
+ if name in self.ENTITY_RELATED_DATASETS:
108
  for idx in json_data:
109
  data = json_data[idx]
110
 
 
122
  "weak_labels": weak_labels,
123
  }
124
  else:
125
+ if name == "trec":
126
+ for idx in json_data:
127
+ data = json_data[idx]
128
+ text = " ".join(data["data"]["text"].split(" ")[1:])
129
+ weak_labels = data["weak_labels"]
130
+ label = data["label"]
131
+
132
+ yield int(idx), {
133
+ "text": text,
134
+ "label": label,
135
+ "weak_labels": weak_labels,
136
+ }
137
+ else:
138
+ for idx in json_data:
139
+ # print(json_data[idx])
140
+ data = json_data[idx]
141
+
142
+ text = data["data"]["text"]
143
+ weak_labels = data["weak_labels"]
144
+ label = data["label"]
145
+
146
+ yield int(idx), {
147
+ "text": text,
148
+ "label": label,
149
+ "weak_labels": weak_labels,
150
+ }