scientific_lay_summarisation.py

# coding=utf-8
# Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
#
# 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.

# Lint as: python3
"""Scientific Lay Summarization Datasets."""


import json
import os

import datasets


_CITATION = """
@misc{Goldsack_2022,
  doi = {10.48550/ARXIV.2210.09932},
  url = {https://arxiv.org/abs/2210.09932},
  author = {Goldsack, Tomas and Zhang, Zhihao and Lin, Chenghua and Scarton, Carolina},
  title = {Making Science Simple: Corpora for the Lay Summarisation of Scientific Literature},
  publisher = {arXiv},
  year = {2022},
  copyright = {arXiv.org perpetual, non-exclusive license}
}
"""

_DESCRIPTION = """
This repository contains the PLOS and eLife datasets, introduced in the EMNLP 2022 paper "[Making Science Simple: Corpora for the Lay Summarisation of Scientific Literature
](https://arxiv.org/abs/2210.09932)". 
Each dataset contains full biomedical research articles paired with expert-written lay summaries (i.e., non-technical summaries). PLOS articles are derived from various journals published by [the Public Library of Science (PLOS)](https://plos.org/), whereas eLife articles are derived from the [eLife](https://elifesciences.org/) journal. More details/anlaysis on the content of each dataset are provided in the paper.

Both "elife" and "plos" have 6 features:
    - "article": the body of the document (including the abstract), sections seperated by "/n".
    - "section_headings": the title of each section, seperated by "/n". 
    - "keywords": keywords describing the topic of the article, seperated by "/n".
    - "title" : the title of the article.
    - "year" : the year the article was published.
    - "summary": the lay summary of the document.
"""

_DOCUMENT = "article"
_SUMMARY = "summary"

_URLS = {
    "plos": "https://drive.usercontent.google.com/download?id=1lZ6PCAtXvmGjRZyp3vQQCEgO_yerH62Q&export=download&authuser=1&confirm=t&uuid=dc63dea1-0814-450f-9234-8bff2b9d1a94&at=APZUnTUfgwJ5Tdiin4ppFPPLWhMX%3A1716450460802",
    "elife": "https://drive.usercontent.google.com/download?id=1WKW8BAqluOlXrpy1B9mV3j3CtAK3JdnE&export=download&authuser=1&confirm=t&uuid=1332bc11-7cbf-4c4d-8561-85621060f397&at=APZUnTVLLKAGVSBpQlYKojrJ57xb%3A1716450570186",
}


class ScientificLaySummarisationConfig(datasets.BuilderConfig):
    """BuilderConfig for Scientific Papers."""

    def __init__(self, filename=None, **kwargs):
        """BuilderConfig for ScientificPapers
        Args:
          filename: filename of different configs for the dataset.
          **kwargs: keyword arguments forwarded to super.
        """
        super(ScientificLaySummarisationConfig, self).__init__(version=datasets.Version("1.0.0"), **kwargs)
        self.filename = filename


class ScientificLaySummarisation(datasets.GeneratorBasedBuilder):
    """Scientific Papers."""

    BUILDER_CONFIGS = [
        ScientificLaySummarisationConfig(name="plos", description="Documents and lay summaries from PLOS journals."),
        ScientificLaySummarisationConfig(name="elife", description="Documents and lay summaries from the eLife journal."),
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    _DOCUMENT: datasets.Value("string"),
                    _SUMMARY: datasets.Value("string"),
                    "section_headings": datasets.Value("string"),
                    "keywords": datasets.Value("string"),
                    "year": datasets.Value("string"),
                    "title": datasets.Value("string"),
                }
            ),
            supervised_keys=None,
            homepage="https://github.com/TGoldsack1/Corpora_for_Lay_Summarisation",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""
        dl_paths = dl_manager.download_and_extract(_URLS)
        path = dl_paths[self.config.name]
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"path": os.path.join(path, "train.json")},
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={"path": os.path.join(path, "val.json")},
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={"path": os.path.join(path, "test.json")},
            ),
        ]

    def _generate_examples(self, path=None):
        """Yields examples."""
        with open(path, encoding="utf-8") as f:
            f = json.loads(f.read())
            for d in f:
                # Possible keys are:
                # "id": str,                      # unique identifier
                # "year": int,                    # year of publication
                # "title": str,                   # title
                # "sections": List[List[str]],    # main text, divided in to sections/sentences
                # "headings" List[str],           # headings of each section
                # "abstract": List[str],          # abstract, in sentences
                # "summary": List[str],           # lay summary, in sentences
                # "keywords": List[str]           # keywords/topic of article

                sections = [" ".join(s).strip() for s in d["sections"]]
                abstract = " ".join(d['abstract']).strip()
                full_doc = [abstract] + sections
                summary = " ".join(d["summary"]).strip()

                yield d["id"], {
                    _DOCUMENT: "\n".join(full_doc),
                    _SUMMARY: summary,
                    "section_headings": "\n".join(["Abstract"] + d["headings"]),
                    "keywords": "\n".join(d["keywords"]),
                    "year": d["year"],
                    "title": d["title"]
                }