liv4ever.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.
# TODO: Address all TODOs and remove all explanatory comments
"""Liv4ever dataset."""


import csv
import json
import os

import datasets


# TODO: Add BibTeX citation
# Find for instance the citation on arxiv or on the dataset repo/website
_CITATION = """\
@inproceedings{rikters-etal-2022,
    title = "Machine Translation for Livonian: Catering for 20 Speakers",
    author = "Rikters, Matīss  and
      Tomingas, Marili and
      Tuisk, Tuuli and
      Valts, Ernštreits  and
      Fishel, Mark",
    booktitle = "Proceedings of ACL 2022",
    year = "2022",
    address = "Dublin, Ireland",
    publisher = "Association for Computational Linguistics"
}
"""

# TODO: Add description of the dataset here
# You can copy an official description
_DESCRIPTION = """\
Livonian is one of the most endangered languages in Europe with just a tiny handful of speakers and virtually no publicly available corpora. 
In this paper we tackle the task of developing neural machine translation (NMT) between Livonian and English, with a two-fold aim: on one hand, 
preserving the language and on the other – enabling access to Livonian folklore, lifestories and other textual intangible heritage as well as 
making it easier to create further parallel corpora. We rely on Livonian's linguistic similarity to Estonian and Latvian and collect parallel 
and monolingual data for the four languages for translation experiments. We combine different low-resource NMT techniques like zero-shot translation, 
cross-lingual transfer and synthetic data creation to reach the highest possible translation quality as well as to find which base languages are 
empirically more helpful for transfer to Livonian. The resulting NMT systems and the collected monolingual and parallel data, including a manually 
translated and verified translation benchmark, are publicly released.

Fields:
- source: source of the data
- en: sentence in English
- liv: sentence in Livonian
"""

# TODO: Add a link to an official homepage for the dataset here
_HOMEPAGE = "https://huggingface.co/datasets/tartuNLP/liv4ever"

# TODO: Add the licence for the dataset here if you can find it
_LICENSE = "CC BY-NC-SA 4.0"

# TODO: Add link to the official dataset URLs here
# 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)
_URL = "https://huggingface.co/datasets/tartuNLP/liv4ever/raw/main/"
_URLS = {
    "train": _URL + "train.json",
    "dev": _URL + "dev.json",
    "test": _URL + "test.json",
}


# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
class liv4ever(datasets.GeneratorBasedBuilder):
    """Liv4ever dataset."""

    VERSION = datasets.Version("1.0.0")

    def _info(self):
        # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
        features = datasets.Features(
            {
                "source": datasets.Value("string"),
                "en": datasets.Value("string"),
                "liv": datasets.Value("string"),
                "lv": datasets.Value("string"),
                "et": datasets.Value("string")
                # These are the features of your dataset like images, labels ...
            }
        )
        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):
        urls_to_download = _URLS
        downloaded_files = dl_manager.download_and_extract(urls_to_download)

        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"]}),
            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
        ]

    def _generate_examples(self, filepath):
        with open(filepath, encoding="utf-8") as f:
            jsondata = json.load(f)
            n = 0
            for source in jsondata:
                for sentence in source["sentences"]:
                    # Yields examples as (key, example) tuples
                    n=n+1
                    if source["source"] in ["facebook", "satversme"]:
                        yield n, {
                            "source": source["source"],
                            "liv": sentence["liv"],
                            "lv": sentence["lv"],
                            "et": sentence["et"],
                            "en": sentence["en"],
                        }