quantlab.dataset.config.MassiveTradeBarsDatasetConfig

class quantlab.dataset.config.MassiveTradeBarsDatasetConfig(*, zarr_file_path, start_date=None, end_date=None, symbols=None, kwargs=None, resample_freq=None, resample_how=None, name=None, raw_data_dir_path, market='us_equity', frequency='tick', vendor='massive', bar_interval='1m', session_start='09:30', session_end='16:00', tickers=None)

Bases: DatasetConfig

Config of the Trade bar panel built from Massive’s daily trade files (ADRs 0030, 0031).

frequency stays "tick" (the raw tier holds one row per trade); the size of the bars is bar_interval. The session window and the bar size are fixed at conversion. The panel’s symbol axis is the Massive ticker, exactly as the trade files spell it; reading it as securities goes through the Ticker map. The roster is tickers; the inherited symbols field is refused by the dataset.

Examples

>>> cfg = MassiveTradeBarsDatasetConfig(
...     zarr_file_path="/data/zarrs/massive_trade_bars_1m.zarr",
...     raw_data_dir_path="/data/downloads/massive",
...     start_date="2024-11-29",
...     end_date="2024-11-29",
... )
>>> cfg.frequency, cfg.vendor, cfg.bar_interval
('tick', 'massive', '1m')

One-second bars of a roster over a window, in a store of their own:

MassiveTradeBarsDatasetConfig(
    zarr_file_path="/data/zarrs/massive_trade_bars_1s.zarr",
    raw_data_dir_path="/data/downloads/massive",
    start_date="2024-11-29",
    end_date="2024-11-29",
    bar_interval="1s",
    tickers=("AAPL", "MSFT"),
)

bar_interval = '1m'

Bar size the trades are resampled to.

frequency = 'tick'

Always tick: the raw tier is one row per trade.

market = 'us_equity'

Always US equity for this vendor.

session_end = '16:00'

End of the session window, US/Eastern wall clock HH:MM.

session_start = '09:30'

Start of the session window, US/Eastern wall clock HH:MM.

tickers = None

The Massive tickers to convert (a roster); None converts every ticker of the trade files.

vendor = 'massive'

Always Massive for this dataset.