mediasum.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
"""MediaSum dataset"""

import os
import json

import datasets


logger = datasets.logging.get_logger(__name__)


_HOMEPAGE = "https://github.com/zcgzcgzcg1/MediaSum"

_DESCRIPTION = """\
This large-scale media interview dataset contains 463.6K transcripts with abstractive summaries, 
collected from interview transcripts and overview / topic descriptions from NPR and CNN.
"""

_CITATION = """\
@article{zhu2021mediasum,
  title={MediaSum: A Large-scale Media Interview Dataset for Dialogue Summarization},
  author={Zhu, Chenguang and Liu, Yang and Mei, Jie and Zeng, Michael},
  journal={arXiv preprint arXiv:2103.06410},
  year={2021}
}
"""

_DOWNLOAD_URLS = {
    "train": "https://huggingface.co/datasets/nbroad/mediasum/resolve/main/train.json",
    "validation": "https://huggingface.co/datasets/nbroad/mediasum/resolve/main/validation.json",
    "test": "https://huggingface.co/datasets/nbroad/mediasum/resolve/main/test.json",
}


class MediaSumConfig(datasets.BuilderConfig):
    """BuilderConfig for MediaSum."""

    def __init__(self, **kwargs):
        """BuilderConfig for MediaSum.
        Args:
          **kwargs: keyword arguments forwarded to super.
        """
        super().__init__(**kwargs)


class MediaSum(datasets.GeneratorBasedBuilder):
    """MediaSum summarization dataset."""

    BUILDER_CONFIGS = [MediaSumConfig(name="mediasum", description="Plain text")]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "id": datasets.Value("string"),
                    "program": datasets.Value("string"),
                    "date": datasets.Value("string"),
                    "url": datasets.Value("string"),
                    "title": datasets.Value("string"),
                    "summary": datasets.Value("string"),
                    "utt": datasets.features.Sequence(
                            datasets.Value("string")
                        ),
                    "speaker": datasets.features.Sequence(
                            datasets.Value("string")
                        ),
                }
            ),
            supervised_keys=None,
            homepage=_HOMEPAGE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        dl_path = dl_manager.download(_DOWNLOAD_URLS)

        return [
            datasets.SplitGenerator(
                name=split,
                gen_kwargs={
                    "filepath": dl_path[split],
                },
            )
            for split in [
                datasets.Split.TRAIN,
                datasets.Split.VALIDATION,
                datasets.Split.TEST,
            ]
        ]

    def _generate_examples(self, filepath):

        with open(filepath, "r") as fp:
            for idx, line in enumerate(fp):
                data = json.loads(line)
                
                # Some do not have titles
                if "title" not in data:
                    data["title"] = ""
                yield idx, data