quantlab.config.stock_kline_config#

quantlab.config.stock_kline_config(start_date=None, end_date=None, symbols=None, kwargs=None, market='us_equity', frequency='1d', subdir='nasdaq_data', store_name='stock.zarr', vendor='tiingo')[source]#

Build the DatasetConfig for daily US-equity bars.

Raw parquet is read from downloads/{market}/{frequency}/{subdir}/{vendor} and the panel is stored at data/{market}/{frequency}/{store_name}. subdir and store_name let a second roster (for example a full-market backfill) live beside the NASDAQ-only one instead of overwriting it. The raw path ends at the vendor segment on purpose: StockDataset checks that the directory basename equals vendor before scanning, which rules out an accidental scan one level up that would merge two vendors’ shards.

Parameters:

start_datestr | None, default None: First date to load, ISO format; None means unbounded.
end_datestr | None, default None: Last date to load, inclusive; None means unbounded.
symbolstuple[str, …] | None, default None: Symbols to keep; None leaves the selection to the dataset.
kwargsdict | None, default None: Extra dataset options.
marketMarket, default “us_equity”: Market label used in the storage paths.
frequencyFrequency, default “1d”: Bar frequency used in the storage paths.
subdirstr, default “nasdaq_data”: Raw-data subdirectory beneath the market/frequency root.
store_namestr, default “stock.zarr”: Zarr store filename.
vendorVendor, default “tiingo”: Vendor whose shards the raw directory holds, also recorded on the config. Defaults to "tiingo", which is what existing callers have on disk.

Examples

With the storage root set to /mnt/quant:

>>> cfg = stock_kline_config(start_date="2020-01-01", symbols=("AAPL",))
>>> cfg.raw_data_dir_path
'/mnt/quant/downloads/us_equity/1d/nasdaq_data/tiingo'
>>> cfg.zarr_file_path
'/mnt/quant/data/us_equity/1d/stock.zarr'
>>> stock_kline_config(subdir="us_all", store_name="us_all.zarr").zarr_file_path
'/mnt/quant/data/us_equity/1d/us_all.zarr'