Commit
·
373d79b
1
Parent(s):
4699c72
upload hubscripts/distemist_hub.py to hub from bigbio repo
Browse files- distemist.py +220 -0
distemist.py
ADDED
@@ -0,0 +1,220 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# coding=utf-8
|
2 |
+
# Copyright 2022 The HuggingFace Datasets Authors and the current dataset script contributor.
|
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 |
+
from pathlib import Path
|
17 |
+
from typing import Dict, List, Tuple
|
18 |
+
|
19 |
+
import datasets
|
20 |
+
import pandas as pd
|
21 |
+
|
22 |
+
from .bigbiohub import kb_features
|
23 |
+
from .bigbiohub import BigBioConfig
|
24 |
+
from .bigbiohub import Tasks
|
25 |
+
|
26 |
+
_LANGUAGES = ['English']
|
27 |
+
_PUBMED = False
|
28 |
+
_LOCAL = False
|
29 |
+
_CITATION = """\
|
30 |
+
@dataset{luis_gasco_2022_6458455,
|
31 |
+
author = {Luis Gasco and Eulàlia Farré and Miranda-Escalada, Antonio and Salvador Lima and Martin Krallinger},
|
32 |
+
title = {{DisTEMIST corpus: detection and normalization of disease mentions in spanish clinical cases}},
|
33 |
+
month = apr,
|
34 |
+
year = 2022,
|
35 |
+
note = {{Funded by the Plan de Impulso de las Tecnologías del Lenguaje (Plan TL).}},
|
36 |
+
publisher = {Zenodo},
|
37 |
+
version = {2.0.0},
|
38 |
+
doi = {10.5281/zenodo.6458455},
|
39 |
+
url = {https://doi.org/10.5281/zenodo.6458455}
|
40 |
+
}
|
41 |
+
"""
|
42 |
+
|
43 |
+
_DATASETNAME = "distemist"
|
44 |
+
_DISPLAYNAME = "DisTEMIST"
|
45 |
+
|
46 |
+
_DESCRIPTION = """\
|
47 |
+
The DisTEMIST corpus is a collection of 1000 clinical cases with disease annotations linked with Snomed-CT concepts.
|
48 |
+
All documents are released in the context of the BioASQ DisTEMIST track for CLEF 2022.
|
49 |
+
"""
|
50 |
+
|
51 |
+
_HOMEPAGE = "https://zenodo.org/record/6458455"
|
52 |
+
|
53 |
+
_LICENSE = 'Creative Commons Attribution 4.0 International'
|
54 |
+
|
55 |
+
_URLS = {
|
56 |
+
_DATASETNAME: "https://zenodo.org/record/6458455/files/distemist.zip?download=1",
|
57 |
+
}
|
58 |
+
|
59 |
+
_SUPPORTED_TASKS = [Tasks.NAMED_ENTITY_RECOGNITION]
|
60 |
+
|
61 |
+
_SOURCE_VERSION = "2.0.0"
|
62 |
+
_BIGBIO_VERSION = "1.0.0"
|
63 |
+
|
64 |
+
|
65 |
+
class DistemistDataset(datasets.GeneratorBasedBuilder):
|
66 |
+
"""
|
67 |
+
The DisTEMIST corpus is a collection of 1000 clinical cases with disease annotations linked with Snomed-CT
|
68 |
+
concepts.
|
69 |
+
"""
|
70 |
+
|
71 |
+
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
72 |
+
BIGBIO_VERSION = datasets.Version(_BIGBIO_VERSION)
|
73 |
+
|
74 |
+
BUILDER_CONFIGS = [
|
75 |
+
BigBioConfig(
|
76 |
+
name="distemist_source",
|
77 |
+
version=SOURCE_VERSION,
|
78 |
+
description="DisTEMIST source schema",
|
79 |
+
schema="source",
|
80 |
+
subset_id="distemist",
|
81 |
+
),
|
82 |
+
BigBioConfig(
|
83 |
+
name="distemist_bigbio_kb",
|
84 |
+
version=BIGBIO_VERSION,
|
85 |
+
description="DisTEMIST BigBio schema",
|
86 |
+
schema="bigbio_kb",
|
87 |
+
subset_id="distemist",
|
88 |
+
),
|
89 |
+
]
|
90 |
+
|
91 |
+
DEFAULT_CONFIG_NAME = "distemist_source"
|
92 |
+
|
93 |
+
def _info(self) -> datasets.DatasetInfo:
|
94 |
+
|
95 |
+
if self.config.schema == "source":
|
96 |
+
features = datasets.Features(
|
97 |
+
{
|
98 |
+
"id": datasets.Value("string"),
|
99 |
+
"document_id": datasets.Value("string"),
|
100 |
+
"passages": [
|
101 |
+
{
|
102 |
+
"id": datasets.Value("string"),
|
103 |
+
"type": datasets.Value("string"),
|
104 |
+
"text": datasets.Sequence(datasets.Value("string")),
|
105 |
+
"offsets": datasets.Sequence([datasets.Value("int32")]),
|
106 |
+
}
|
107 |
+
],
|
108 |
+
"entities": [
|
109 |
+
{
|
110 |
+
"id": datasets.Value("string"),
|
111 |
+
"type": datasets.Value("string"),
|
112 |
+
"text": datasets.Sequence(datasets.Value("string")),
|
113 |
+
"offsets": datasets.Sequence([datasets.Value("int32")]),
|
114 |
+
"concept_codes": datasets.Sequence(
|
115 |
+
datasets.Value("string")
|
116 |
+
),
|
117 |
+
"semantic_relations": datasets.Sequence(
|
118 |
+
datasets.Value("string")
|
119 |
+
),
|
120 |
+
}
|
121 |
+
],
|
122 |
+
}
|
123 |
+
)
|
124 |
+
elif self.config.schema == "bigbio_kb":
|
125 |
+
features = kb_features
|
126 |
+
|
127 |
+
return datasets.DatasetInfo(
|
128 |
+
description=_DESCRIPTION,
|
129 |
+
features=features,
|
130 |
+
homepage=_HOMEPAGE,
|
131 |
+
license=str(_LICENSE),
|
132 |
+
citation=_CITATION,
|
133 |
+
)
|
134 |
+
|
135 |
+
def _split_generators(self, dl_manager) -> List[datasets.SplitGenerator]:
|
136 |
+
"""Returns SplitGenerators."""
|
137 |
+
urls = _URLS[_DATASETNAME]
|
138 |
+
data_dir = dl_manager.download_and_extract(urls)
|
139 |
+
return [
|
140 |
+
datasets.SplitGenerator(
|
141 |
+
name=datasets.Split.TRAIN,
|
142 |
+
gen_kwargs={
|
143 |
+
"entities_mapping_file_path": Path(data_dir)
|
144 |
+
/ "training/subtrack1_entities/distemist_subtrack1_training_mentions.tsv",
|
145 |
+
"linking_mapping_file_path": Path(data_dir)
|
146 |
+
/ "training/subtrack2_linking/distemist_subtrack1_training1_linking.tsv",
|
147 |
+
"text_files_dir": Path(data_dir) / "training/text_files",
|
148 |
+
},
|
149 |
+
),
|
150 |
+
]
|
151 |
+
|
152 |
+
def _generate_examples(
|
153 |
+
self,
|
154 |
+
entities_mapping_file_path: Path,
|
155 |
+
linking_mapping_file_path: Path,
|
156 |
+
text_files_dir: Path,
|
157 |
+
) -> Tuple[int, Dict]:
|
158 |
+
"""Yields examples as (key, example) tuples."""
|
159 |
+
entities_mapping = pd.read_csv(entities_mapping_file_path, sep="\t")
|
160 |
+
linking_mapping = pd.read_csv(linking_mapping_file_path, sep="\t")
|
161 |
+
|
162 |
+
entity_file_names = set(entities_mapping["filename"])
|
163 |
+
linking_file_names = set(linking_mapping["filename"])
|
164 |
+
|
165 |
+
# entity_file_names = entity_file_names.difference(linking_file_names)
|
166 |
+
|
167 |
+
for uid, filename in enumerate(entity_file_names):
|
168 |
+
text_file = text_files_dir / f"{filename}.txt"
|
169 |
+
|
170 |
+
doc_text = text_file.read_text()
|
171 |
+
# doc_text = doc_text.replace("\n", "")
|
172 |
+
|
173 |
+
if filename in linking_file_names:
|
174 |
+
entities_df: pd.DataFrame = linking_mapping[
|
175 |
+
linking_mapping["filename"] == filename
|
176 |
+
]
|
177 |
+
else:
|
178 |
+
entities_df: pd.DataFrame = entities_mapping[
|
179 |
+
entities_mapping["filename"] == filename
|
180 |
+
]
|
181 |
+
|
182 |
+
example = {
|
183 |
+
"id": f"{uid}",
|
184 |
+
"document_id": filename,
|
185 |
+
"passages": [
|
186 |
+
{
|
187 |
+
"id": f"{uid}_{filename}_passage",
|
188 |
+
"type": "clinical_case",
|
189 |
+
"text": [doc_text],
|
190 |
+
"offsets": [[0, len(doc_text)]],
|
191 |
+
}
|
192 |
+
],
|
193 |
+
}
|
194 |
+
if self.config.schema == "bigbio_kb":
|
195 |
+
example["events"] = []
|
196 |
+
example["coreferences"] = []
|
197 |
+
example["relations"] = []
|
198 |
+
|
199 |
+
entities = []
|
200 |
+
for row in entities_df.itertuples(name="Entity"):
|
201 |
+
entity = {
|
202 |
+
"id": f"{uid}_{row.filename}_{row.Index}_entity_id_{row.mark}",
|
203 |
+
"type": row.label,
|
204 |
+
"text": [row.span],
|
205 |
+
"offsets": [[row.off0, row.off1]],
|
206 |
+
}
|
207 |
+
if self.config.schema == "source":
|
208 |
+
entity["concept_codes"] = []
|
209 |
+
entity["semantic_relations"] = []
|
210 |
+
if filename in linking_file_names:
|
211 |
+
entity["concept_codes"] = row.code.split("+")
|
212 |
+
entity["semantic_relations"] = row.semantic_rel.split("+")
|
213 |
+
|
214 |
+
elif self.config.schema == "bigbio_kb":
|
215 |
+
entity["normalized"] = []
|
216 |
+
|
217 |
+
entities.append(entity)
|
218 |
+
|
219 |
+
example["entities"] = entities
|
220 |
+
yield uid, example
|