Register a dataset with the name, update type, and only the options its update planner needs.
from bagelquant_data import DatasetSpec
lake.admin.datasets.register(DatasetSpec("trade_cal", "general"))
lake.admin.datasets.register(DatasetSpec("daily", "by_daily", calendar="trade_cal"))
lake.admin.datasets.register(DatasetSpec("balancesheet", "by_asset", asset_list="stock_basic", primary_key_extra=("period",)))
The pipeline derives the incremental key from time, asset_id, and optional
primary_key_extra fields. General datasets do not require a canonical key.
Store the same compact mapping in TOML and register it with register_toml.