polqa.py

# Copyright 2020 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.


import csv
import json

import datasets


_CITATION = """\
@misc{rybak2022improving,
      title={Improving Question Answering Performance through Manual Annotation: Costs, Benefits and Strategies}, 
      author={Piotr Rybak and Piotr Przybyła and Maciej Ogrodniczuk},
      year={2022},
      eprint={2212.08897},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
"""

_DESCRIPTION = """\
PolQA is the first Polish dataset for OpenQA. It consists of 7,000 questions, 87,525 manually labeled evidence passages, and a corpus of over 7 million candidate passages.
"""

_HOMEPAGE = ""

_LICENSE = ""

_FEATURES_PAIRS = datasets.Features(
    {        
        "question_id": datasets.Value("int32"),
        "passage_title": datasets.Value("string"),
        "passage_text": datasets.Value("string"),
        "passage_wiki": datasets.Value("string"),
        "passage_id": datasets.Value("string"),
        "duplicate": datasets.Value("bool"),
        "question": datasets.Value("string"),
        "relevant": datasets.Value("bool"),
        "annotated_by": datasets.Value("string"),
        "answers": datasets.Value("string"),
        "question_formulation": datasets.Value("string"),
        "question_type": datasets.Value("string"),
        "entity_type": datasets.Value("string"),
        "entity_subtype": datasets.Value("string"),
        "split": datasets.Value("string"),
        "passage_source": datasets.Value("string"),
    }
)

_FEATURES_PASSAGES = datasets.Features(
    {        
        "id": datasets.Value("string"),
        "title": datasets.Value("string"),
        "text": datasets.Value("string"),
    }
)

_URLS = {
    "pairs": {
        "train": ["data/train.csv"],
        "validation": ["data/valid.csv"],
        "test": ["data/test.csv"],
    },
    "passages": {
        "train": ["data/passages.jsonl"],
    },
}


class PolQA(datasets.GeneratorBasedBuilder):
    """PolQA is the first Polish dataset for OpenQA. It consists of manually labeled QA pairs and a corpus of Wikipedia passages."""

    BUILDER_CONFIGS = list(map(lambda x: datasets.BuilderConfig(name=x, version=datasets.Version("1.0.0")), _URLS.keys()))
    DEFAULT_CONFIG_NAME = "pairs"

    def _info(self):
        if self.config.name == "pairs":
            features = _FEATURES_PAIRS
        else:
            features = _FEATURES_PASSAGES

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        urls = _URLS[self.config.name]
        data_dir = dl_manager.download_and_extract(urls)
        if self.config.name == "pairs":
            return [
                datasets.SplitGenerator(
                    name=datasets.Split.TRAIN,
                    gen_kwargs={
                        "filepaths": data_dir["train"],
                        "split": "train",
                    },
                ),
                datasets.SplitGenerator(
                    name=datasets.Split.VALIDATION,
                    gen_kwargs={
                        "filepaths": data_dir["validation"],
                        "split": "validation",
                    },
                ),
                datasets.SplitGenerator(
                    name=datasets.Split.TEST,
                    gen_kwargs={
                        "filepaths": data_dir["test"],
                        "split": "test",
                    },
                ),
            ]
        else:
            return [
                datasets.SplitGenerator(
                    name=datasets.Split.TRAIN,
                    gen_kwargs={
                        "filepaths": data_dir["train"],
                        "split": "train",
                    },
                ),
            ]

    @staticmethod
    def _parse_bool(text):
        if text == 'True':
            return True
        elif text == 'False':
            return False
        else:
            raise ValueError

    def _generate_examples(self, filepaths, split):
        if self.config.name == "pairs":
            boolean_features = [name for name, val in _FEATURES_PAIRS.items() if val.dtype == "bool"]

            for filepath in filepaths:
                with open(filepath, encoding="utf-8") as f:
                    data = csv.DictReader(f)
                    for i, row in enumerate(data):
                        for boolean_feature in boolean_features:
                            row[boolean_feature] = self._parse_bool(row[boolean_feature])
                        yield i, row
        else:
            for filepath in filepaths:
                with open(filepath, encoding="utf-8") as f:
                    for i, row in enumerate(f):
                        parsed_row = json.loads(row)
                        yield i, parsed_row