system HF staff commited on
Commit
e700b36
0 Parent(s):

Update files from the datasets library (from 1.2.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.2.0

.gitattributes ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bin.* filter=lfs diff=lfs merge=lfs -text
5
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.model filter=lfs diff=lfs merge=lfs -text
12
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
13
+ *.onnx filter=lfs diff=lfs merge=lfs -text
14
+ *.ot filter=lfs diff=lfs merge=lfs -text
15
+ *.parquet filter=lfs diff=lfs merge=lfs -text
16
+ *.pb filter=lfs diff=lfs merge=lfs -text
17
+ *.pt filter=lfs diff=lfs merge=lfs -text
18
+ *.pth filter=lfs diff=lfs merge=lfs -text
19
+ *.rar filter=lfs diff=lfs merge=lfs -text
20
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
21
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
22
+ *.tflite filter=lfs diff=lfs merge=lfs -text
23
+ *.tgz filter=lfs diff=lfs merge=lfs -text
24
+ *.xz filter=lfs diff=lfs merge=lfs -text
25
+ *.zip filter=lfs diff=lfs merge=lfs -text
26
+ *.zstandard filter=lfs diff=lfs merge=lfs -text
27
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,254 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language_creators:
5
+ - found
6
+ languages:
7
+ - en
8
+ licenses:
9
+ - unknown
10
+ multilinguality:
11
+ - monolingual
12
+ size_categories:
13
+ Ade_corpos_v2_classificaion:
14
+ - 10K<n<100K
15
+ Ade_corpos_v2_drug_ade_relation:
16
+ - 1K<n<10K
17
+ Ade_corpos_v2_drug_dosage_relation:
18
+ - n<1K
19
+ source_datasets:
20
+ - original
21
+ task_categories:
22
+ Ade_corpos_v2_classificaion:
23
+ - text-classification
24
+ Ade_corpos_v2_drug_ade_relation:
25
+ - structure-prediction
26
+ Ade_corpos_v2_drug_dosage_relation:
27
+ - structure-prediction
28
+ task_ids:
29
+ Ade_corpos_v2_classificaion:
30
+ - fact-checking
31
+ Ade_corpos_v2_drug_ade_relation:
32
+ - coreference-resolution
33
+ Ade_corpos_v2_drug_dosage_relation:
34
+ - coreference-resolution
35
+ ---
36
+
37
+ # Dataset Card for [Needs More Information]
38
+
39
+ ## Table of Contents
40
+ - [Dataset Card for [Needs More Information]](#dataset-card-for-needs-more-information)
41
+ - [Table of Contents](#table-of-contents)
42
+ - [Dataset Description](#dataset-description)
43
+ - [Dataset Summary](#dataset-summary)
44
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
45
+ - [Languages](#languages)
46
+ - [Dataset Structure](#dataset-structure)
47
+ - [Data Instances](#data-instances)
48
+ - [Data Fields](#data-fields)
49
+ - [Data Splits](#data-splits)
50
+ - [Dataset Creation](#dataset-creation)
51
+ - [Curation Rationale](#curation-rationale)
52
+ - [Source Data](#source-data)
53
+ - [Initial Data Collection and Normalization](#initial-data-collection-and-normalization)
54
+ - [Who are the source language producers?](#who-are-the-source-language-producers)
55
+ - [Annotations](#annotations)
56
+ - [Annotation process](#annotation-process)
57
+ - [Who are the annotators?](#who-are-the-annotators)
58
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
59
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
60
+ - [Social Impact of Dataset](#social-impact-of-dataset)
61
+ - [Discussion of Biases](#discussion-of-biases)
62
+ - [Other Known Limitations](#other-known-limitations)
63
+ - [Additional Information](#additional-information)
64
+ - [Dataset Curators](#dataset-curators)
65
+ - [Licensing Information](#licensing-information)
66
+ - [Citation Information](#citation-information)
67
+
68
+ ## Dataset Description
69
+
70
+ - **Homepage:** https://www.sciencedirect.com/science/article/pii/S1532046412000615
71
+ - **Repository:** [Needs More Information]
72
+ - **Paper:** https://www.sciencedirect.com/science/article/pii/S1532046412000615
73
+ - **Leaderboard:** [Needs More Information]
74
+ - **Point of Contact:** [Needs More Information]
75
+
76
+ ### Dataset Summary
77
+
78
+ ADE-Corpus-V2 Dataset: Adverse Drug Reaction Data.
79
+ This is a dataset for Classification if a sentence is ADE-related (True) or not (False) and Relation Extraction between Adverse Drug Event and Drug.
80
+ DRUG-AE.rel provides relations between drugs and adverse effects.
81
+ DRUG-DOSE.rel provides relations between drugs and dosages.
82
+ ADE-NEG.txt provides all sentences in the ADE corpus that DO NOT contain any drug-related adverse effects.
83
+
84
+ ### Supported Tasks and Leaderboards
85
+
86
+ Sentiment classification, Relation Extraction
87
+
88
+ ### Languages
89
+
90
+ English
91
+
92
+ ## Dataset Structure
93
+
94
+ ### Data Instances
95
+
96
+ #### Config - `Ade_corpos_v2_classificaion`
97
+ ```
98
+ {
99
+ 'label': 1,
100
+ 'text': 'Intravenous azithromycin-induced ototoxicity.'
101
+ }
102
+
103
+ ```
104
+
105
+ #### Config - `Ade_corpos_v2_drug_ade_relation`
106
+
107
+ ```
108
+ {
109
+ 'drug': 'azithromycin',
110
+ 'effect': 'ototoxicity',
111
+ 'indexes': {
112
+ 'drug': {
113
+ 'end_char': [24],
114
+ 'start_char': [12]
115
+ },
116
+ 'effect': {
117
+ 'end_char': [44],
118
+ 'start_char': [33]
119
+ }
120
+ },
121
+ 'text': 'Intravenous azithromycin-induced ototoxicity.'
122
+
123
+ }
124
+
125
+ ```
126
+
127
+ #### Config - `Ade_corpos_v2_drug_dosage_relation`
128
+
129
+ ```
130
+ {
131
+ 'dosage': '4 times per day',
132
+ 'drug': 'insulin',
133
+ 'indexes': {
134
+ 'dosage': {
135
+ 'end_char': [56],
136
+ 'start_char': [41]
137
+ },
138
+ 'drug': {
139
+ 'end_char': [40],
140
+ 'start_char': [33]}
141
+ },
142
+ 'text': 'She continued to receive regular insulin 4 times per day over the following 3 years with only occasional hives.'
143
+ }
144
+
145
+ ```
146
+
147
+
148
+ ### Data Fields
149
+
150
+ #### Config - `Ade_corpos_v2_drug_ade_relation`
151
+
152
+ - `text` - Input text.
153
+ - `label` - Whether the adverse drug effect(ADE) related (1) or not (0).
154
+ -
155
+ #### Config - `Ade_corpos_v2_drug_ade_relation`
156
+
157
+ - `text` - Input text.
158
+ - `drug` - Name of drug.
159
+ - `effect` - Effect caused by the drug.
160
+ - `indexes.drug.start_char` - Start index of `drug` string in text.
161
+ - `indexes.drug.end_char` - End index of `drug` string in text.
162
+ - `indexes.effect.start_char` - Start index of `effect` string in text.
163
+ - `indexes.effect.end_char` - End index of `effect` string in text.
164
+
165
+ #### Config - `Ade_corpos_v2_drug_dosage_relation`
166
+
167
+ - `text` - Input text.
168
+ - `drug` - Name of drug.
169
+ - `dosage` - Dosage of the drug.
170
+ - `indexes.drug.start_char` - Start index of `drug` string in text.
171
+ - `indexes.drug.end_char` - End index of `drug` string in text.
172
+ - `indexes.dosage.start_char` - Start index of `dosage` string in text.
173
+ - `indexes.dosage.end_char` - End index of `dosage` string in text.
174
+
175
+
176
+ ### Data Splits
177
+
178
+ | Train |
179
+ | ------ |
180
+ | 23516 |
181
+
182
+ ## Dataset Creation
183
+
184
+ ### Curation Rationale
185
+
186
+ [Needs More Information]
187
+
188
+ ### Source Data
189
+
190
+ #### Initial Data Collection and Normalization
191
+
192
+ [Needs More Information]
193
+
194
+ #### Who are the source language producers?
195
+
196
+ [Needs More Information]
197
+
198
+ ### Annotations
199
+
200
+ #### Annotation process
201
+
202
+ [Needs More Information]
203
+
204
+ #### Who are the annotators?
205
+
206
+ [Needs More Information]
207
+
208
+ ### Personal and Sensitive Information
209
+
210
+ [Needs More Information]
211
+
212
+ ## Considerations for Using the Data
213
+
214
+ ### Social Impact of Dataset
215
+
216
+ [Needs More Information]
217
+
218
+ ### Discussion of Biases
219
+
220
+ [Needs More Information]
221
+
222
+ ### Other Known Limitations
223
+
224
+ [Needs More Information]
225
+
226
+ ## Additional Information
227
+
228
+ ### Dataset Curators
229
+
230
+ [Needs More Information]
231
+
232
+ ### Licensing Information
233
+
234
+ [Needs More Information]
235
+
236
+ ### Citation Information
237
+
238
+ ```
239
+ @article{GURULINGAPPA2012885,
240
+ title = "Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports",
241
+ journal = "Journal of Biomedical Informatics",
242
+ volume = "45",
243
+ number = "5",
244
+ pages = "885 - 892",
245
+ year = "2012",
246
+ note = "Text Mining and Natural Language Processing in Pharmacogenomics",
247
+ issn = "1532-0464",
248
+ doi = "https://doi.org/10.1016/j.jbi.2012.04.008",
249
+ url = "http://www.sciencedirect.com/science/article/pii/S1532046412000615",
250
+ author = "Harsha Gurulingappa and Abdul Mateen Rajput and Angus Roberts and Juliane Fluck and Martin Hofmann-Apitius and Luca Toldo",
251
+ keywords = "Adverse drug effect, Benchmark corpus, Annotation, Harmonization, Sentence classification",
252
+ abstract = "A significant amount of information about drug-related safety issues such as adverse effects are published in medical case reports that can only be explored by human readers due to their unstructured nature. The work presented here aims at generating a systematically annotated corpus that can support the development and validation of methods for the automatic extraction of drug-related adverse effects from medical case reports. The documents are systematically double annotated in various rounds to ensure consistent annotations. The annotated documents are finally harmonized to generate representative consensus annotations. In order to demonstrate an example use case scenario, the corpus was employed to train and validate models for the classification of informative against the non-informative sentences. A Maximum Entropy classifier trained with simple features and evaluated by 10-fold cross-validation resulted in the F1 score of 0.70 indicating a potential useful application of the corpus."
253
+ }
254
+ ```
ade_corpus_v2.py ADDED
@@ -0,0 +1,268 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+
16
+ # Lint as: python3
17
+
18
+ """ Adverse Drug Reaction Data: ADE-Corpus-V2 """
19
+
20
+ from __future__ import absolute_import, division, print_function
21
+
22
+ import re
23
+
24
+ import datasets
25
+
26
+
27
+ _CITATION = """\
28
+ @article{GURULINGAPPA2012885,
29
+ title = "Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports",
30
+ journal = "Journal of Biomedical Informatics",
31
+ volume = "45",
32
+ number = "5",
33
+ pages = "885 - 892",
34
+ year = "2012",
35
+ note = "Text Mining and Natural Language Processing in Pharmacogenomics",
36
+ issn = "1532-0464",
37
+ doi = "https://doi.org/10.1016/j.jbi.2012.04.008",
38
+ url = "http://www.sciencedirect.com/science/article/pii/S1532046412000615",
39
+ author = "Harsha Gurulingappa and Abdul Mateen Rajput and Angus Roberts and Juliane Fluck and Martin Hofmann-Apitius and Luca Toldo",
40
+ keywords = "Adverse drug effect, Benchmark corpus, Annotation, Harmonization, Sentence classification",
41
+ abstract = "A significant amount of information about drug-related safety issues such as adverse effects are published in medical case reports that can only be explored by human readers due to their unstructured nature. The work presented here aims at generating a systematically annotated corpus that can support the development and validation of methods for the automatic extraction of drug-related adverse effects from medical case reports. The documents are systematically double annotated in various rounds to ensure consistent annotations. The annotated documents are finally harmonized to generate representative consensus annotations. In order to demonstrate an example use case scenario, the corpus was employed to train and validate models for the classification of informative against the non-informative sentences. A Maximum Entropy classifier trained with simple features and evaluated by 10-fold cross-validation resulted in the F1 score of 0.70 indicating a potential useful application of the corpus."
42
+ }
43
+ """
44
+
45
+ _DESCRIPTION = """\
46
+ ADE-Corpus-V2 Dataset: Adverse Drug Reaction Data.
47
+ This is a dataset for Classification if a sentence is ADE-related (True) or not (False) and Relation Extraction between Adverse Drug Event and Drug.
48
+ DRUG-AE.rel provides relations between drugs and adverse effects.
49
+ DRUG-DOSE.rel provides relations between drugs and dosages.
50
+ ADE-NEG.txt provides all sentences in the ADE corpus that DO NOT contain any drug-related adverse effects.
51
+ """
52
+
53
+ _DOWNLOAD_URL = "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/{}-{}.{}"
54
+
55
+ # Different usage configs/
56
+ configs = {
57
+ "classification": "Ade_corpos_v2_classificaion",
58
+ "RE_ade": "Ade_corpos_v2_drug_ade_relation",
59
+ "RE_dosage": "Ade_corpos_v2_drug_dosage_relation",
60
+ }
61
+
62
+
63
+ class ADE_Corpus_V2Config(datasets.BuilderConfig):
64
+ """BuilderConfig for ADE_Corpus_V2."""
65
+
66
+ def __init__(self, **kwargs):
67
+ """BuilderConfig for ADE_Corpus_V2.
68
+ Args:
69
+ **kwargs: keyword arguments forwarded to super.
70
+ """
71
+ super(ADE_Corpus_V2Config, self).__init__(**kwargs)
72
+
73
+
74
+ class ADECorpusV2(datasets.GeneratorBasedBuilder):
75
+ """ADE_Corpus_V2 Dataset: Adverse Drug Reaction Data for Classification and Relation Extraction tasks ."""
76
+
77
+ BUILDER_CONFIGS = [
78
+ ADE_Corpus_V2Config(
79
+ name="Ade_corpos_v2_classificaion",
80
+ version=datasets.Version("1.0.0"),
81
+ description="ADE_Corpus_V2 Dataset for Classification if a sentence is ADE-related or not.",
82
+ ),
83
+ ADE_Corpus_V2Config(
84
+ name="Ade_corpos_v2_drug_ade_relation",
85
+ version=datasets.Version("1.0.0"),
86
+ description="ADE_Corpus_V2 Dataset for Relation Extraction between Adverse Drug Event and Drug.",
87
+ ),
88
+ ADE_Corpus_V2Config(
89
+ name="Ade_corpos_v2_drug_dosage_relation",
90
+ version=datasets.Version("1.0.0"),
91
+ description="ADE_Corpus_V2 Dataset for Relation Extraction between Drug dosage and Drug.",
92
+ ),
93
+ ]
94
+
95
+ def _info(self):
96
+
97
+ if self.config.name == configs["classification"]:
98
+ features = datasets.Features(
99
+ {
100
+ "text": datasets.Value("string"),
101
+ "label": datasets.features.ClassLabel(names=["Not-Related", "Related"]),
102
+ }
103
+ )
104
+
105
+ if self.config.name == configs["RE_ade"]:
106
+ features = datasets.Features(
107
+ {
108
+ "text": datasets.Value("string"),
109
+ "drug": datasets.Value("string"),
110
+ "effect": datasets.Value("string"),
111
+ "indexes": {
112
+ "drug": datasets.Sequence(
113
+ {
114
+ "start_char": datasets.Value("int32"),
115
+ "end_char": datasets.Value("int32"),
116
+ }
117
+ ),
118
+ "effect": datasets.Sequence(
119
+ {
120
+ "start_char": datasets.Value("int32"),
121
+ "end_char": datasets.Value("int32"),
122
+ }
123
+ ),
124
+ },
125
+ }
126
+ )
127
+
128
+ if self.config.name == configs["RE_dosage"]:
129
+ features = datasets.Features(
130
+ {
131
+ "text": datasets.Value("string"),
132
+ "drug": datasets.Value("string"),
133
+ "dosage": datasets.Value("string"),
134
+ "indexes": {
135
+ "drug": datasets.Sequence(
136
+ {
137
+ "start_char": datasets.Value("int32"),
138
+ "end_char": datasets.Value("int32"),
139
+ }
140
+ ),
141
+ "dosage": datasets.Sequence(
142
+ {
143
+ "start_char": datasets.Value("int32"),
144
+ "end_char": datasets.Value("int32"),
145
+ }
146
+ ),
147
+ },
148
+ }
149
+ )
150
+
151
+ return datasets.DatasetInfo(
152
+ description=_DESCRIPTION,
153
+ features=features,
154
+ supervised_keys=None,
155
+ homepage="https://www.sciencedirect.com/science/article/pii/S1532046412000615",
156
+ citation=_CITATION,
157
+ )
158
+
159
+ def _split_generators(self, dl_manager):
160
+ """Returns SplitGenerators."""
161
+
162
+ DAE_path = dl_manager.download_and_extract(_DOWNLOAD_URL.format("DRUG", "AE", "rel"))
163
+ DD_path = dl_manager.download_and_extract(_DOWNLOAD_URL.format("DRUG", "DOSE", "rel"))
164
+ DAE_NEG_path = dl_manager.download_and_extract(_DOWNLOAD_URL.format("ADE", "NEG", "txt"))
165
+
166
+ return [
167
+ datasets.SplitGenerator(
168
+ name=datasets.Split.TRAIN,
169
+ gen_kwargs={"DRUG_AE_file": DAE_path, "DRUG_DOSAGE_file": DD_path, "NEG_DRUG_AE_file": DAE_NEG_path},
170
+ ),
171
+ ]
172
+
173
+ def _generate_examples(self, DRUG_AE_file, DRUG_DOSAGE_file, NEG_DRUG_AE_file):
174
+ """Generate ADE_Corpus_V2 examples."""
175
+
176
+ # For Classification task with ade dataset.
177
+ if self.config.name == configs["classification"]:
178
+ texts, labels = [], []
179
+ with open(DRUG_AE_file, encoding="utf-8") as f:
180
+ for line in f:
181
+ pubmed_id, text = line.strip().split("|")[:2]
182
+ texts.append(text)
183
+ labels.append("Related")
184
+
185
+ with open(NEG_DRUG_AE_file, encoding="utf-8") as f:
186
+ for line in f:
187
+ pubmed_id, neg = line.strip().split(" ")[:2]
188
+ text = " ".join(line.strip().split(" ")[2:])
189
+ texts.append(text)
190
+ labels.append("Not-Related")
191
+
192
+ for i in range(len(labels)):
193
+ text, label = texts[i], labels[i]
194
+ yield i, {"text": text, "label": label}
195
+
196
+ # For Relation Extraction between drug and its effect.
197
+ elif self.config.name == configs["RE_ade"]:
198
+
199
+ texts, drugs, effects, drug_indexes, effect_indexes = [], [], [], [], []
200
+ with open(DRUG_AE_file, encoding="utf-8") as f:
201
+ for line in f:
202
+ value = line.strip().split("|")
203
+ text = value[1]
204
+ effect = value[2]
205
+ drug = value[5]
206
+
207
+ # add index of drug and effect from text
208
+ effect_matches, drug_matches = [], []
209
+ for match in re.finditer(effect, text):
210
+ effect_matches.append({"start_char": match.start(), "end_char": match.end()})
211
+ effect_indexes.append(effect_matches)
212
+
213
+ for match in re.finditer(drug, text):
214
+ drug_matches.append({"start_char": match.start(), "end_char": match.end()})
215
+ drug_indexes.append(drug_matches)
216
+
217
+ texts.append(text)
218
+ drugs.append(drug)
219
+ effects.append(effect)
220
+
221
+ for idx, (text, drug, effect, drug_index, effect_index) in enumerate(
222
+ zip(texts, drugs, effects, drug_indexes, effect_indexes)
223
+ ):
224
+
225
+ output = {
226
+ "text": text,
227
+ "drug": drug,
228
+ "effect": effect,
229
+ "indexes": {"drug": drug_index, "effect": effect_index},
230
+ }
231
+
232
+ yield idx, output
233
+
234
+ # For Relation Extraction between drug and its dosage.
235
+ elif self.config.name == configs["RE_dosage"]:
236
+
237
+ texts, drugs, dosages, drug_indexes, dosage_indexes = [], [], [], [], []
238
+ with open(DRUG_DOSAGE_file, encoding="utf-8") as f:
239
+ for line in f:
240
+ value = line.strip().split("|")
241
+ text = value[1]
242
+ dosage = value[2]
243
+ drug = value[5]
244
+
245
+ # add index of drug and effect from text
246
+ dosage_matches, drug_matches = [], []
247
+ for match in re.finditer(dosage, text):
248
+ dosage_matches.append({"start_char": match.start(), "end_char": match.end()})
249
+ dosage_indexes.append(dosage_matches)
250
+
251
+ for match in re.finditer(drug, text):
252
+ drug_matches.append({"start_char": match.start(), "end_char": match.end()})
253
+ drug_indexes.append(drug_matches)
254
+
255
+ texts.append(text)
256
+ drugs.append(drug)
257
+ dosages.append(dosage)
258
+
259
+ for idx, (text, drug, dosage, drug_index, dosage_index) in enumerate(
260
+ zip(texts, drugs, dosages, drug_indexes, dosage_indexes)
261
+ ):
262
+ output = {
263
+ "text": text,
264
+ "drug": drug,
265
+ "dosage": dosage,
266
+ "indexes": {"drug": drug_index, "dosage": dosage_index},
267
+ }
268
+ yield idx, output
dataset_infos.json ADDED
@@ -0,0 +1 @@
 
1
+ {"Ade_corpos_v2_classificaion": {"description": " ADE-Corpus-V2 Dataset: Adverse Drug Reaction Data.\n This is a dataset for Classification if a sentence is ADE-related (True) or not (False) and Relation Extraction between Adverse Drug Event and Drug.\n DRUG-AE.rel provides relations between drugs and adverse effects.\n DRUG-DOSE.rel provides relations between drugs and dosages.\n ADE-NEG.txt provides all sentences in the ADE corpus that DO NOT contain any drug-related adverse effects.\n", "citation": "@article{GURULINGAPPA2012885,\ntitle = \"Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports\",\njournal = \"Journal of Biomedical Informatics\",\nvolume = \"45\",\nnumber = \"5\",\npages = \"885 - 892\",\nyear = \"2012\",\nnote = \"Text Mining and Natural Language Processing in Pharmacogenomics\",\nissn = \"1532-0464\",\ndoi = \"https://doi.org/10.1016/j.jbi.2012.04.008\",\nurl = \"http://www.sciencedirect.com/science/article/pii/S1532046412000615\",\nauthor = \"Harsha Gurulingappa and Abdul Mateen Rajput and Angus Roberts and Juliane Fluck and Martin Hofmann-Apitius and Luca Toldo\",\nkeywords = \"Adverse drug effect, Benchmark corpus, Annotation, Harmonization, Sentence classification\",\nabstract = \"A significant amount of information about drug-related safety issues such as adverse effects are published in medical case reports that can only be explored by human readers due to their unstructured nature. The work presented here aims at generating a systematically annotated corpus that can support the development and validation of methods for the automatic extraction of drug-related adverse effects from medical case reports. The documents are systematically double annotated in various rounds to ensure consistent annotations. The annotated documents are finally harmonized to generate representative consensus annotations. In order to demonstrate an example use case scenario, the corpus was employed to train and validate models for the classification of informative against the non-informative sentences. A Maximum Entropy classifier trained with simple features and evaluated by 10-fold cross-validation resulted in the F1 score of 0.70 indicating a potential useful application of the corpus.\"\n}\n", "homepage": "https://www.sciencedirect.com/science/article/pii/S1532046412000615", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 2, "names": ["Not-Related", "Related"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "ade_corpus_v2", "config_name": "Ade_corpos_v2_classificaion", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 3403711, "num_examples": 23516, "dataset_name": "ade_corpus_v2"}}, "download_checksums": {"https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-AE.rel": {"num_bytes": 1423024, "checksum": "542cdc483ccc94927762eaf2c9a8ecac49a6c10037dda2895be6a6e20160f75a"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-DOSE.rel": {"num_bytes": 59669, "checksum": "78b46dfcdc1325d7f81e5e01f5a424e380e4b38fafca02f6e8f67064ca73f2db"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/ADE-NEG.txt": {"num_bytes": 2308469, "checksum": "8f506c159042ce354fbf26981dc39971dde8f09b1158d94106eab1e516e53fcf"}}, "download_size": 3791162, "post_processing_size": null, "dataset_size": 3403711, "size_in_bytes": 7194873}, "Ade_corpos_v2_drug_ade_relation": {"description": " ADE-Corpus-V2 Dataset: Adverse Drug Reaction Data.\n This is a dataset for Classification if a sentence is ADE-related (True) or not (False) and Relation Extraction between Adverse Drug Event and Drug.\n DRUG-AE.rel provides relations between drugs and adverse effects.\n DRUG-DOSE.rel provides relations between drugs and dosages.\n ADE-NEG.txt provides all sentences in the ADE corpus that DO NOT contain any drug-related adverse effects.\n", "citation": "@article{GURULINGAPPA2012885,\ntitle = \"Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports\",\njournal = \"Journal of Biomedical Informatics\",\nvolume = \"45\",\nnumber = \"5\",\npages = \"885 - 892\",\nyear = \"2012\",\nnote = \"Text Mining and Natural Language Processing in Pharmacogenomics\",\nissn = \"1532-0464\",\ndoi = \"https://doi.org/10.1016/j.jbi.2012.04.008\",\nurl = \"http://www.sciencedirect.com/science/article/pii/S1532046412000615\",\nauthor = \"Harsha Gurulingappa and Abdul Mateen Rajput and Angus Roberts and Juliane Fluck and Martin Hofmann-Apitius and Luca Toldo\",\nkeywords = \"Adverse drug effect, Benchmark corpus, Annotation, Harmonization, Sentence classification\",\nabstract = \"A significant amount of information about drug-related safety issues such as adverse effects are published in medical case reports that can only be explored by human readers due to their unstructured nature. The work presented here aims at generating a systematically annotated corpus that can support the development and validation of methods for the automatic extraction of drug-related adverse effects from medical case reports. The documents are systematically double annotated in various rounds to ensure consistent annotations. The annotated documents are finally harmonized to generate representative consensus annotations. In order to demonstrate an example use case scenario, the corpus was employed to train and validate models for the classification of informative against the non-informative sentences. A Maximum Entropy classifier trained with simple features and evaluated by 10-fold cross-validation resulted in the F1 score of 0.70 indicating a potential useful application of the corpus.\"\n}\n", "homepage": "https://www.sciencedirect.com/science/article/pii/S1532046412000615", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "drug": {"dtype": "string", "id": null, "_type": "Value"}, "effect": {"dtype": "string", "id": null, "_type": "Value"}, "indexes": {"drug": {"feature": {"start_char": {"dtype": "int32", "id": null, "_type": "Value"}, "end_char": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}, "effect": {"feature": {"start_char": {"dtype": "int32", "id": null, "_type": "Value"}, "end_char": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}}, "post_processed": null, "supervised_keys": null, "builder_name": "ade_corpus_v2", "config_name": "Ade_corpos_v2_drug_ade_relation", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1546021, "num_examples": 6821, "dataset_name": "ade_corpus_v2"}}, "download_checksums": {"https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-AE.rel": {"num_bytes": 1423024, "checksum": "542cdc483ccc94927762eaf2c9a8ecac49a6c10037dda2895be6a6e20160f75a"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-DOSE.rel": {"num_bytes": 59669, "checksum": "78b46dfcdc1325d7f81e5e01f5a424e380e4b38fafca02f6e8f67064ca73f2db"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/ADE-NEG.txt": {"num_bytes": 2308469, "checksum": "8f506c159042ce354fbf26981dc39971dde8f09b1158d94106eab1e516e53fcf"}}, "download_size": 3791162, "post_processing_size": null, "dataset_size": 1546021, "size_in_bytes": 5337183}, "Ade_corpos_v2_drug_dosage_relation": {"description": " ADE-Corpus-V2 Dataset: Adverse Drug Reaction Data.\n This is a dataset for Classification if a sentence is ADE-related (True) or not (False) and Relation Extraction between Adverse Drug Event and Drug.\n DRUG-AE.rel provides relations between drugs and adverse effects.\n DRUG-DOSE.rel provides relations between drugs and dosages.\n ADE-NEG.txt provides all sentences in the ADE corpus that DO NOT contain any drug-related adverse effects.\n", "citation": "@article{GURULINGAPPA2012885,\ntitle = \"Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports\",\njournal = \"Journal of Biomedical Informatics\",\nvolume = \"45\",\nnumber = \"5\",\npages = \"885 - 892\",\nyear = \"2012\",\nnote = \"Text Mining and Natural Language Processing in Pharmacogenomics\",\nissn = \"1532-0464\",\ndoi = \"https://doi.org/10.1016/j.jbi.2012.04.008\",\nurl = \"http://www.sciencedirect.com/science/article/pii/S1532046412000615\",\nauthor = \"Harsha Gurulingappa and Abdul Mateen Rajput and Angus Roberts and Juliane Fluck and Martin Hofmann-Apitius and Luca Toldo\",\nkeywords = \"Adverse drug effect, Benchmark corpus, Annotation, Harmonization, Sentence classification\",\nabstract = \"A significant amount of information about drug-related safety issues such as adverse effects are published in medical case reports that can only be explored by human readers due to their unstructured nature. The work presented here aims at generating a systematically annotated corpus that can support the development and validation of methods for the automatic extraction of drug-related adverse effects from medical case reports. The documents are systematically double annotated in various rounds to ensure consistent annotations. The annotated documents are finally harmonized to generate representative consensus annotations. In order to demonstrate an example use case scenario, the corpus was employed to train and validate models for the classification of informative against the non-informative sentences. A Maximum Entropy classifier trained with simple features and evaluated by 10-fold cross-validation resulted in the F1 score of 0.70 indicating a potential useful application of the corpus.\"\n}\n", "homepage": "https://www.sciencedirect.com/science/article/pii/S1532046412000615", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "drug": {"dtype": "string", "id": null, "_type": "Value"}, "dosage": {"dtype": "string", "id": null, "_type": "Value"}, "indexes": {"drug": {"feature": {"start_char": {"dtype": "int32", "id": null, "_type": "Value"}, "end_char": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}, "dosage": {"feature": {"start_char": {"dtype": "int32", "id": null, "_type": "Value"}, "end_char": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}}, "post_processed": null, "supervised_keys": null, "builder_name": "ade_corpus_v2", "config_name": "Ade_corpos_v2_drug_dosage_relation", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 64725, "num_examples": 279, "dataset_name": "ade_corpus_v2"}}, "download_checksums": {"https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-AE.rel": {"num_bytes": 1423024, "checksum": "542cdc483ccc94927762eaf2c9a8ecac49a6c10037dda2895be6a6e20160f75a"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/DRUG-DOSE.rel": {"num_bytes": 59669, "checksum": "78b46dfcdc1325d7f81e5e01f5a424e380e4b38fafca02f6e8f67064ca73f2db"}, "https://raw.githubusercontent.com/trunghlt/AdverseDrugReaction/master/ADE-Corpus-V2/ADE-NEG.txt": {"num_bytes": 2308469, "checksum": "8f506c159042ce354fbf26981dc39971dde8f09b1158d94106eab1e516e53fcf"}}, "download_size": 3791162, "post_processing_size": null, "dataset_size": 64725, "size_in_bytes": 3855887}}
dummy/Ade_corpos_v2_classificaion/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36c086aede182ee2bdef54b207142b8d8c8eb0b774c4e6f5019a7b27ee1f8d88
3
+ size 1895
dummy/Ade_corpos_v2_drug_ade_relation/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b466530fc6122183765f162dae41d5672c811b3079c02862c665ca2209bda703
3
+ size 1895
dummy/Ade_corpos_v2_drug_dosage_relation/1.0.0/dummy_data.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6613df8eca9c7c0de5e5347e645120d1ce4f072ee2dd992eadabc20ff9dbe47
3
+ size 1895