regulatory_comments.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.

import json
import datasets

# Description of the dataset
_DESCRIPTION = """\
United States governmental agencies often make proposed regulations open to the public for comment. 
Proposed regulations are organized into "dockets". This project will use Regulation.gov public API 
to aggregate and clean public comments for dockets that mention opioid use. 

Each example will consist of one docket, and include metadata such as docket id, docket title, etc. 
Each docket entry will also include information about the top 10 comments, including comment metadata
and comment text. 
"""

# Homepage URL of the dataset
_HOMEPAGE = "https://www.regulations.gov/"

# URL to download the dataset
_URLS = {"url": "https://huggingface.co/datasets/ro-h/regulatory_comments/raw/main/docket_comments_v6.json"}

# Citation of repository
_CITATION = """@misc{ro_huang_regulatory_2023-1,
	author = {{Ro Huang}},
	date = {2023-03-19},
	publisher = {Hugging Face},
	title = {Regulatory Comments},
	url = {https://huggingface.co/datasets/ro-h/regulatory_comments},
	version = {1.1.4},
	bdsk-url-1 = {https://huggingface.co/datasets/ro-h/regulatory_comments}}
"""

# Class definition for handling the dataset
class RegComments(datasets.GeneratorBasedBuilder):

    # Version of the dataset
    VERSION = datasets.Version("1.1.4")
    
    # Method to define the structure of the dataset
    def _info(self):
        # Defining the structure of the dataset
        features = datasets.Features({
            "id": datasets.Value("string"),
            "agency": datasets.Value("string"), #Added In
            "title": datasets.Value("string"),
            "update_date": datasets.Value("string"), #Added In
            "update_time": datasets.Value("string"), #Added In
            "purpose": datasets.Value("string"),
            "keywords": datasets.Sequence(datasets.Value("string")),
            "comments": datasets.Sequence({
                "text": datasets.Value("string"),
                "comment_id": datasets.Value("string"),
                "comment_url": datasets.Value("string"),
                "comment_date": datasets.Value("string"),
                "comment_time": datasets.Value("string"),
                "commenter_fname": datasets.Value("string"),
                "commenter_lname": datasets.Value("string"),
                "comment_length": datasets.Value("int32")
            })
        })

        # Returning the dataset structure
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            citation = _CITATION
        )

    # Method to handle dataset splitting (e.g., train/test)
    def _split_generators(self, dl_manager):
        urls = _URLS["url"]
        data_dir = dl_manager.download_and_extract(urls)
        # Defining the split (here, only train split is defined)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "filepath": data_dir,
                },
            ),
        ]
    
    # Method to generate examples from the dataset
    def _generate_examples(self, filepath):
        """This function returns the examples in the raw (text) form."""
        key = 0
        with open(filepath, 'r', encoding='utf-8') as f:
            data = json.load(f)
            for docket in data:
                # Extracting data fields from each docket
                docket_id = docket["id"]
                docket_agency = docket["agency"]
                docket_title = docket["title"]
                docket_update_date = docket["update_date"]
                docket_update_time = docket["update_time"]
                docket_purpose = docket.get("purpose", "unspecified")
                docket_keywords = docket.get("keywords", [])
                comments = docket["comments"]

                # Yielding each docket with its information
                yield key, {
                    "id": docket_id,
                    "agency": docket_agency,
                    "title": docket_title,
                    "update_date": docket_update_date,
                    "update_time": docket_update_time,
                    "purpose": docket_purpose, 
                    "keywords": docket_keywords, 
                    "comments": comments
                }
                key += 1