medical_institutions_reviews.py

import datasets
import json

class HfreviewsConfig(datasets.BuilderConfig):
    def __init__(self, features, **kwargs):
        super(HfreviewsConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
        self.features = features

class MedicalInstitutionsReviews(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [
        HfreviewsConfig(
            name="simple",
            description="Simple config",
            features=["review_id", "content", "general", "quality", "service", "equipment", "food", "location"],
        )
    ]
    
    def _info(self):
        return datasets.DatasetInfo(
            description='Healthcare facilities reviews dataset.',
            features=datasets.Features(
                {
                    "review_id": datasets.Value("string"),
                    "content": datasets.Value("string"),
                    "general": datasets.Value("string"),
                    "quality": datasets.Value("string"),
                    "service": datasets.Value("string"),
                    "equipment": datasets.Value("string"),
                    "food": datasets.Value("string"),
                    "location": datasets.Value("string"),
                    "Idx": datasets.Value("int32"),
                }
            ),
        )
    
    def _split_generators(self, dl_manager: datasets.DownloadManager):
        urls_to_download = {
            "train": "medical_institutions_reviews.jsonl"
        }
        downloaded_files = dl_manager.download_and_extract(urls_to_download)
        
        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]})
        ]
            
    def _generate_examples(self, filepath):
        """Yields examples."""
        with open(filepath, encoding="utf-8") as f:
            for uid, row in enumerate(f):
                data = json.loads(row)
                data["Idx"] = uid
                yield uid, data