FaithDial.py

# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor (Nouha Dziri).
#
# 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.
"""FaithDial: A Faithful Benchmark for Information-Seeking Dialogue"""


import json

import datasets


# Find for instance the citation on arxiv or on the dataset repo/website
_CITATION = """\
@article{dziri2022faithdial,
  title={FaithDial: A Faithful Benchmark for Information-Seeking Dialogue},
  author={Dziri, Nouha and Kamalloo, Ehsan and Milton, Sivan and Zaiane, Osmar and Yu, Mo and Ponti, Edoardo and Reddy, Siva},
  journal={arXiv preprint, arXiv:2204.10757},
  year={2022},
  url={https://arxiv.org/abs/2204.10757}
}
"""

# You can copy an official description
_DESCRIPTION = """\
FaithDial is a new benchmark for hallucination-free dialogues, created by manually editing hallucinated and uncooperative responses in Wizard of Wikipedia.
"""

_HOMEPAGE = "https://mcgill-nlp.github.io/FaithDial/"

_LICENSE = "MIT"

# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
_URLS = {
    "train": "data/train.json",
    "valid": "data/valid.json",
    "valid_random_split": "data/valid_random_split.json",
    "valid_topic_split": "data/valid_topic_split.json",
    "test": "data/test.json",
    "test_random_split": "data/test_random_split.json",
    "test_topic_split": "data/test_topic_split.json",
}


class FaithDialDataset(datasets.GeneratorBasedBuilder):
    """FaithDial is a new benchmark for hallucination-free dialogues."""

    VERSION = datasets.Version("1.0.0")

    # This is an example of a dataset with multiple configurations.
    # If you don't want/need to define several sub-sets in your dataset,
    # just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.

    # If you need to make complex sub-parts in the datasets with configurable options
    # You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
    # BUILDER_CONFIG_CLASS = MyBuilderConfig

    # You will be able to load one or the other configurations in the following list with
    # data = datasets.load_dataset('my_dataset', 'first_domain')
    # data = datasets.load_dataset('my_dataset', 'second_domain')
    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name="plain_text", version=VERSION, description="Plain text"),
    ]

    DEFAULT_CONFIG_NAME = (
        "plain_text"  # It's not mandatory to have a default configuration. Just use one if it make sense.
    )

    def _info(self):
        features = datasets.Features(
            {
                "dialog_idx": datasets.Value("int32"),
                "response": datasets.Value("string"),
                "original_response": datasets.Value("string"),
                "history": datasets.features.Sequence(datasets.Value("string")),
                "knowledge": datasets.Value("string"),
                "BEGIN": datasets.features.Sequence(datasets.Value("string")),
                "VRM": datasets.features.Sequence(datasets.Value("string")),
            }
        )

        return datasets.DatasetInfo(
            # This is the description that will appear on the datasets page.
            description=_DESCRIPTION,
            # This defines the different columns of the dataset and their types
            features=features,  # Here we define them above because they are different between the two configurations
            # If there's a common (input, target) tuple from the features, uncomment supervised_keys line below and
            # specify them. They'll be used if as_supervised=True in builder.as_dataset.
            # supervised_keys=("sentence", "label"),
            # Homepage of the dataset for documentation
            homepage=_HOMEPAGE,
            # License for the dataset if available
            license=_LICENSE,
            # Citation for the dataset
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        # If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name

        # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLS
        # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
        # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
        downloaded_files = dl_manager.download_and_extract(_URLS)

        split_dict = {
            "train": datasets.Split.TRAIN,
            "valid": datasets.Split.VALIDATION,
            "test": datasets.Split.TEST,
        }

        return [
            datasets.SplitGenerator(
                name=split_dict.get(split, split),
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": downloaded_file,
                    "split": split,
                },
            )
            for split, downloaded_file in sorted(downloaded_files.items(), key=lambda x: x[0])
        ]

    # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
    def _generate_examples(self, filepath, split):
        # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
        with open(filepath, encoding="utf-8") as f:
            data = json.load(f)
            key = 0
            for dialogue in data:
                for utterance in dialogue["utterances"]:
                    yield key, {
                        "dialog_idx": dialogue["dialog_idx"],
                        "response": utterance["response"],
                        "original_response": utterance["original_response"],
                        "history": utterance["history"],
                        "knowledge": utterance["knowledge"],
                        "BEGIN": utterance["BEGIN"],
                        "VRM": utterance["VRM"],
                    }
                    key += 1