quantlab.factor.config.IntradayCompositeConfig#

class quantlab.factor.config.IntradayCompositeConfig(*, warmup_bars=0, dataset=None, file_path=None, factor_names=None, kwargs=None, resample_freq=None, resample_how=None, name=None, factor, composites=<factory>, njobs=128)[source]#

Bases: BaseFactorConfig

Config of quantlab.factor.predefined.intraday.IntradayComposite.

factor is the factor whose stored daily values are combined, usually a resampled IntradayStats; it keeps its own config and store, which must cover the range asked for. composites maps each output name to its parts, each a string "[-]<output>:<statistic>:<window>": a trailing mean or cv (standard deviation over mean) of one of the wrapped factor’s outputs over window bars, negated by a leading -. dataset is filled with the wrapped factor’s dataset.

Examples

With daily an IntradayStats resampled to "1d":

>>> cfg = IntradayCompositeConfig(
...     factor=daily,
...     composites={"small_lit_trades": ("-trade_size:mean:60", "-offexchange_share:mean:60")},
... )
>>> cfg.warmup_bars, cfg.composites["small_lit_trades"][0]
(0, '-trade_size:mean:60')

composites#: Output name to its parts, each "[-]<output>:<mean|cv>:<window>".

dataset = None#: Filled with the wrapped factor’s dataset by the factor’s config setter.

factor#: The factor whose stored values are combined.

njobs = 128#: Threads used by the KunQuant executor.

warmup_bars = 0#: The windows’ history is read from the wrapped store; anything but 0 is refused.