dataset.py

import csv, datasets, os

class Voaburmese(datasets.GeneratorBasedBuilder):
    VERSION = datasets.Version("1.0.0")

    _SLOT2IDX = {"morning": 0, "evening": 1}

    def _info(self):
        return datasets.DatasetInfo(
            description="VOA Burmese Morning & Evening audio index (2012-09-16 → today)",
            license="odc-pddl-1.0",
            features=datasets.Features(
                {
                    "id":           datasets.Value("string"),
                    "date":         datasets.Value("string"),
                    "slot":         datasets.ClassLabel(names=["morning", "evening"]),
                    "title":        datasets.Value("string"),
                    "article_url":  datasets.Value("string"),
                    "audio":        datasets.Audio(),   # remote MP3 streamed on access
                }
            ),
        )

    def _split_generators(self, dl_manager):
        # the CSV is shipped with the repo; no need to download
        csv_path = os.path.join(self._get_manual_data_dir(), "index.csv")
        return [datasets.SplitGenerator(name="train", gen_kwargs={"csv_path": csv_path})]

    def _generate_examples(self, csv_path):
        with open(csv_path, newline="", encoding="utf-8") as f:
            for i, row in enumerate(csv.DictReader(f)):
                yield i, {
                    "id":          f"{row['date']}_{row['slot']}",
                    "date":        row["date"],
                    "slot":        self._SLOT2IDX[row["slot"]],
                    "title":       row["title"],
                    "article_url": row["article_url"],
                    "audio":       row["mp3_url"],
                }