# coding=utf-8 # Copyright 2022 The HuggingFace Datasets Authors and the current dataset script contributor. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ State of the art models using deep neural networks have become very good in learning an accurate mapping from inputs to outputs. However, they still lack generalization capabilities in conditions that differ from the ones encountered during training. This is even more challenging in specialized, and knowledge intensive domains, where training data is limited. To address this gap, we introduce MedNLI - a dataset annotated by doctors, performing a natural language inference task (NLI), grounded in the medical history of patients. As the source of premise sentences, we used the MIMIC-III. More specifically, to minimize the risks to patient privacy, we worked with clinical notes corresponding to the deceased patients. The clinicians in our team suggested the Past Medical History to be the most informative section of a clinical note, from which useful inferences can be drawn about the patient. The files comprising this dataset must be on the users local machine in a single directory that is passed to `datasets.load_datset` via the `data_dir` kwarg. This loader script will read the archive files directly (i.e. the user should not uncompress, untar or unzip any of the files). For example, if `data_dir` is `"mednli"` it should contain the following files: mednli ├── mednli-a-natural-language-inference-dataset-for-the-clinical-domain-1.0.0.zip """ import json import os from typing import Dict, List, Tuple import datasets from .bigbiohub import entailment_features from .bigbiohub import BigBioConfig from .bigbiohub import Tasks _LANGUAGES = ["English"] _PUBMED = False _LOCAL = True _CITATION = """\ @misc{https://doi.org/10.13026/c2rs98, title = {MedNLI — A Natural Language Inference Dataset For The Clinical Domain}, author = {Shivade, Chaitanya}, year = 2017, publisher = {physionet.org}, doi = {10.13026/C2RS98}, url = {https://physionet.org/content/mednli/} } """ _DATASETNAME = "mednli" _DISPLAYNAME = "MedNLI" _DESCRIPTION = """\ State of the art models using deep neural networks have become very good in learning an accurate mapping from inputs to outputs. However, they still lack generalization capabilities in conditions that differ from the ones encountered during training. This is even more challenging in specialized, and knowledge intensive domains, where training data is limited. To address this gap, we introduce MedNLI - a dataset annotated by doctors, performing a natural language inference task (NLI), grounded in the medical history of patients. As the source of premise sentences, we used the MIMIC-III. More specifically, to minimize the risks to patient privacy, we worked with clinical notes corresponding to the deceased patients. The clinicians in our team suggested the Past Medical History to be the most informative section of a clinical note, from which useful inferences can be drawn about the patient. """ _HOMEPAGE = "https://physionet.org/content/mednli/1.0.0/" _LICENSE = "PHYSIONET_LICENSE_1p5" _URLS = {} _SUPPORTED_TASKS = [Tasks.TEXTUAL_ENTAILMENT] _SOURCE_VERSION = "1.0.0" _BIGBIO_VERSION = "1.0.0" class MedNLIDataset(datasets.GeneratorBasedBuilder): """MedNLI""" SOURCE_VERSION = datasets.Version(_SOURCE_VERSION) BIGBIO_VERSION = datasets.Version(_BIGBIO_VERSION) BUILDER_CONFIGS = [ BigBioConfig( name="mednli_source", version=SOURCE_VERSION, description="MedNLI source schema", schema="source", subset_id="mednli", ), BigBioConfig( name="mednli_bigbio_te", version=BIGBIO_VERSION, description="MedNLI BigBio schema", schema="bigbio_te", subset_id="mednli", ), ] DEFAULT_CONFIG_NAME = "mednli_source" def _info(self) -> datasets.DatasetInfo: if self.config.schema == "source": features = datasets.Features( { "pairID": datasets.Value("string"), "gold_label": datasets.Value("string"), "sentence1": datasets.Value("string"), "sentence2": datasets.Value("string"), "sentence1_parse": datasets.Value("string"), "sentence2_parse": datasets.Value("string"), "sentence1_binary_parse": datasets.Value("string"), "sentence2_binary_parse": datasets.Value("string"), } ) elif self.config.schema == "bigbio_te": features = entailment_features return datasets.DatasetInfo( description=_DESCRIPTION, features=features, homepage=_HOMEPAGE, license=str(_LICENSE), citation=_CITATION, ) def _split_generators(self, dl_manager) -> List[datasets.SplitGenerator]: if self.config.data_dir is None: raise ValueError( "This is a local dataset. Please pass the data_dir kwarg to load_dataset." ) else: extract_dir = dl_manager.extract( os.path.join( self.config.data_dir, "mednli-a-natural-language-inference-dataset-for-the-clinical-domain-1.0.0.zip", ) ) data_dir = os.path.join( extract_dir, "mednli-a-natural-language-inference-dataset-for-the-clinical-domain-1.0.0", ) return [ datasets.SplitGenerator( name=datasets.Split.TRAIN, gen_kwargs={ "filepath": os.path.join(data_dir, "mli_train_v1.jsonl"), "split": "train", }, ), datasets.SplitGenerator( name=datasets.Split.TEST, gen_kwargs={ "filepath": os.path.join(data_dir, "mli_test_v1.jsonl"), "split": "test", }, ), datasets.SplitGenerator( name=datasets.Split.VALIDATION, gen_kwargs={ "filepath": os.path.join(data_dir, "mli_dev_v1.jsonl"), "split": "dev", }, ), ] def _generate_examples(self, filepath, split: str) -> Tuple[int, Dict]: with open(filepath, "r") as f: if self.config.schema == "source": for line in f: json_line = json.loads(line) yield json_line["pairID"], json_line elif self.config.schema == "bigbio_te": for line in f: json_line = json.loads(line) entailment_example = { "id": json_line["pairID"], "premise": json_line["sentence1"], "hypothesis": json_line["sentence2"], "label": json_line["gold_label"], } yield json_line["pairID"], entailment_example