jomleh.py

# coding=utf-8
# Copyright 2021 The TensorFlow Datasets Authors and the HuggingFace NLP 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
"""Jomleh The Farsi sentence dataset."""


import collections
import io
import zstandard
import json

from dataclasses import dataclass

import datasets


logger = datasets.logging.get_logger(__name__)


@dataclass
class Identification:
    label: str
    prob: float


_DESCRIPTION = """\
Jomleh is a Farsi (Persian) monolingual dataset composed of one sentence per \
sample. It's focused on quality over quantity and it's curated mostly based \
on the OSCAR project (https://oscar-project.com) among other data sources.\
"""

_URL = "https://mlengineer.ai"

_LICENSE = """
    These data are released under this licensing scheme
    We do not own any of the text from which these data has been extracted.
    We license the actual packaging of these data under the Creative Commons CC0 license \
    (\"no rights reserved\") http://creativecommons.org/publicdomain/zero/1.0/
    To the extent possible under law, Inria has waived all copyright \
    and related or neighboring rights to OSCAR
    This work is published from: France.
    Should you consider that our data contains material that is owned by you \
    and should therefore not be reproduced here, please:
    * Clearly identify yourself, with detailed contact data such as an address, \
    telephone number or email address at which you can be contacted.
    * Clearly identify the copyrighted work claimed to be infringed.
    * Clearly identify the material that is claimed to be infringing and \
    information reasonably sufficient to allow us to locate the material.
    We will comply to legitimate requests by removing the affected sources \
    from the next release of the corpus. \
"""

_CITATION = """\
"""

_BASE_DATA_PAT_FORMAT_STR = "files/"
_BASE_CHECKSUM_FILE_NAME = "checksum.sha256"

class JomlehConfig(datasets.BuilderConfig):
    """OSCAR corpus."""

    def __init__(self, **kwargs):
        """BuilderConfig for Jomleh.
        Args:
            **kwargs: Keyword arguments forwarded to super.
        """
        description = (
            f"Jomleh dataset from April 2023"
        )
        super(JomlehConfig, self).__init__(
            name="Jomleh", description=description, **kwargs
        )

        # Additional attributes
        self.base_data_path = _BASE_DATA_PAT_FORMAT_STR


class Jomleh(datasets.GeneratorBasedBuilder):
    """Jomleh The Farsi text dataset based on OSCAR project."""

    BUILDER_CONFIGS = [
        JomlehConfig(  # pylint: disable=g-complex-comprehension
            version=datasets.Version("2023.4.0"),
        )
    ]
    BUILDER_CONFIG_CLASS = JomlehConfig

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "id": datasets.Value("int64"),
                    "text": datasets.Value("string"),
                    "source": datasets.Value("string"),
                }
            ),
            supervised_keys=None,
            homepage=_URL,
            citation=_CITATION,
            license=_LICENSE,
        )

    def _split_generators(self, dl_manager):
        checksum_path = self.config.base_data_path + _BASE_CHECKSUM_FILE_NAME
        checksum_file = dl_manager.download(checksum_path)
        with open(checksum_file, encoding="utf-8") as f:
            data_filenames = [line.split()[1] for line in f if line]
            data_urls = [
                self.config.base_data_path + data_filename
                for data_filename in data_filenames
            ]
        doc_files = dl_manager.download(
            [url for url in data_urls if url.endswith(".jsonl.zst")]
        )
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN, gen_kwargs={"doc_files": doc_files}
            ),
        ]

    def _generate_examples(self, doc_files):
        """This function returns the examples in the raw (text) form by iterating on all the files."""
        for doc_path in doc_files:
            logger.info("generating examples from = %s", doc_path)

            with open(doc_path, "rb") as fh:
                dctx = zstandard.ZstdDecompressor()
                stream_reader = dctx.stream_reader(fh)
                buffered_reader = io.BufferedReader(stream_reader)
                text_stream = io.TextIOWrapper(buffered_reader, encoding="utf-8")
                for line in text_stream:
                    doc = json.loads(line)
                    yield doc["id"], {"id": doc["id"], "text": doc["text"], "source": doc["source"]}