legal_lama.py

# coding=utf-8
# 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.
"""LegalLAMA: Legal LAnguage Model Analysis (LAMA) (LAMA) dataset."""

import json
import os
import datasets


MAIN_CITATION = """
@inproceedings{chalkidis-garneau-etal-2023-lexlms,
    title = {{LeXFiles and LegalLAMA: Facilitating English Multinational Legal Language Model Development}},
    author = "Chalkidis*, Ilias and 
              Garneau*, Nicolas and
              Goanta, Catalina and 
              Katz, Daniel Martin and 
              Søgaard, Anders",
    booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics",
    month = july,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/xxx",
}
"""

_DESCRIPTION = """LegalLAMA: Legal LAnguage Model Analysis (LAMA) (LAMA) dataset."""
MAIN_PATH = 'https://huggingface.co/datasets/lexlms/legal_lama/resolve/main'


class LegalLAMAConfig(datasets.BuilderConfig):
    """BuilderConfig for XAI - Fairness."""

    def __init__(
        self,
        data_url,
        **kwargs,
    ):
        """BuilderConfig for LegalLAMA.

        Args:
          data_url: `string`, url to download the zip file from
          **kwargs: keyword arguments forwarded to super.
        """
        super(LegalLAMAConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
        self.data_url = data_url


class LegalLAMA(datasets.GeneratorBasedBuilder):
    """LegalLAMA: A multilingual benchmark for evaluating fairness in legal text processing. Version 1.0"""

    BUILDER_CONFIGS = [
        LegalLAMAConfig(
            name="canadian_crimes",
            data_url=os.path.join(MAIN_PATH, "canadian_crimes.jsonl"),
        ),
        LegalLAMAConfig(
            name="canadian_sections",
            data_url=os.path.join(MAIN_PATH, "canadian_sections.jsonl"),
        ),
        LegalLAMAConfig(
            name="cjeu_terms",
            data_url=os.path.join(MAIN_PATH, "cjeu_terms.jsonl"),
        ),
        LegalLAMAConfig(
            name="ecthr_terms",
            data_url=os.path.join(MAIN_PATH, "ecthr_terms.jsonl"),
        ),
        LegalLAMAConfig(
            name="ecthr_articles",
            data_url=os.path.join(MAIN_PATH, "ecthr_articles.jsonl"),
        ),
        LegalLAMAConfig(
            name="us_crimes",
            data_url=os.path.join(MAIN_PATH, "us_crimes.jsonl"),
        ),
        LegalLAMAConfig(
            name="us_terms",
            data_url=os.path.join(MAIN_PATH, "us_terms.jsonl"),
        ),
        LegalLAMAConfig(
            name="contract_types",
            data_url=os.path.join(MAIN_PATH, "contract_types.jsonl"),
        ),
        LegalLAMAConfig(
            name="contract_sections",
            data_url=os.path.join(MAIN_PATH, "contract_sections.jsonl"),
        ),
    ]

    def _info(self):
        features = {"text": datasets.Value("string"), "label": datasets.Value("string"), "category": datasets.Value("string")}
        return datasets.DatasetInfo(
            description=self.config.description,
            features=datasets.Features(features),
            homepage="https://huggingface.co/datasets/lexlms/legal_lama",
            citation=MAIN_CITATION,
        )

    def _split_generators(self, dl_manager):
        data_dir = dl_manager.download(self.config.data_url)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": data_dir,
                    "split": "test",
                },
            ),
        ]

    def _get_category(self, sample):
        if 'canadian_article' in sample:
            category = sample['canadian_article']
        elif 'legal_topic' in sample:
            category = sample['legal_topic']
        elif 'echr_article' in sample:
            category = sample['echr_article']
        else:
            category = sample['obj_label']
        return category
        
    def _generate_examples(self, filepath, split):
        """This function returns the examples in the raw (text) form."""
        with open(filepath, encoding="utf-8") as f:
            for id_, row in enumerate(f):
                data = json.loads(row)
                example = {
                    "text": data["masked_sentences"][0],
                    "label": data["obj_label"],
                    "category": self._get_category(data)
                }
                yield id_, example