faquad.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.
#
# Adapted from the SQuAD script.
#

# Lint as: python3
"""FaQuAD: Reading Comprehension Dataset in the Domain of Brazilian Higher Education."""


import json

import datasets
from datasets.tasks import QuestionAnsweringExtractive


logger = datasets.logging.get_logger(__name__)


_CITATION = """\
@INPROCEEDINGS{
    8923668,
    author={Sayama, Hélio Fonseca and Araujo, Anderson Viçoso and Fernandes, Eraldo Rezende},
    booktitle={2019 8th Brazilian Conference on Intelligent Systems (BRACIS)},
    title={FaQuAD: Reading Comprehension Dataset in the Domain of Brazilian Higher Education},
    year={2019},
    volume={},
    number={},
    pages={443-448},
    doi={10.1109/BRACIS.2019.00084}
}
"""

_DESCRIPTION = """\
Academic secretaries and faculty members of higher education institutions face a common problem: 
  the abundance of questions sent by academics 
  whose answers are found in available institutional documents. 
The official documents produced by Brazilian public universities are vast and disperse, 
  which discourage students to further search for answers in such sources.
In order to lessen this problem, we present FaQuAD: 
  a novel machine reading comprehension dataset 
  in the domain of Brazilian higher education institutions. 
FaQuAD follows the format of SQuAD (Stanford Question Answering Dataset) [Rajpurkar et al. 2016]. 
It comprises 900 questions about 249 reading passages (paragraphs), 
  which were taken from 18 official documents of a computer science college 
  from a Brazilian federal university 
  and 21 Wikipedia articles related to Brazilian higher education system. 
As far as we know, this is the first Portuguese reading comprehension dataset in this format.
"""

_URL = "https://raw.githubusercontent.com/liafacom/faquad/master/data/"
_URLS = {
    "train": _URL + "train.json",
    "dev": _URL + "dev.json",
}


class FaquadConfig(datasets.BuilderConfig):
    """BuilderConfig for FaQuAD."""

    def __init__(self, **kwargs):
        """BuilderConfig for FaQuAD.

        Args:
          **kwargs: keyword arguments forwarded to super.
        """
        super(FaquadConfig, self).__init__(**kwargs)


class Faquad(datasets.GeneratorBasedBuilder):
    """FaQuAD: Reading Comprehension Dataset in the Domain of Brazilian Higher Education. Version 1.0."""

    BUILDER_CONFIGS = [
        FaquadConfig(
            name="plain_text",
            version=datasets.Version("1.0.0", ""),
            description="Plain text",
        ),
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "id": datasets.Value("string"),
                    "title": datasets.Value("string"),
                    "context": datasets.Value("string"),
                    "question": datasets.Value("string"),
                    "answers": datasets.features.Sequence(
                        {
                            "text": datasets.Value("string"),
                            "answer_start": datasets.Value("int32"),
                        }
                    ),
                }
            ),
            # No default supervised_keys (as we have to pass both question
            # and context as input).
            supervised_keys=None,
            homepage="https://github.com/liafacom/faquad",
            citation=_CITATION,
            task_templates=[
                QuestionAnsweringExtractive(
                    question_column="question", context_column="context", answers_column="answers"
                )
            ],
        )

    def _split_generators(self, dl_manager):
        downloaded_files = dl_manager.download_and_extract(_URLS)

        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
            datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["dev"]}),
        ]

    def _generate_examples(self, filepath):
        """This function returns the examples in the raw (text) form."""
        logger.info("generating examples from = %s", filepath)
        key = 0
        with open(filepath, encoding="utf-8") as f:
            faquad = json.load(f)
            for article in faquad["data"]:
                title = article.get("title", "")
                for paragraph in article["paragraphs"]:
                    context = paragraph["context"]  # do not strip leading blank spaces GH-2585
                    for qa in paragraph["qas"]:
                        answer_starts = [answer["answer_start"] for answer in qa["answers"]]
                        answers = [answer["text"] for answer in qa["answers"]]
                        # Features currently used are "context", "question", and "answers".
                        # Others are extracted here for the ease of future expansions.
                        yield key, {
                            "title": title,
                            "context": context,
                            "question": qa["question"],
                            "id": qa["id"],
                            "answers": {
                                "answer_start": answer_starts,
                                "text": answers,
                            },
                        }
                        key += 1