quantlab.runs.backtest_run.FoldArtifacts

class quantlab.runs.backtest_run.FoldArtifacts(index, window, metrics, trained_run, data_fingerprint)

Bases: object

What a run_cv run writes for one fold, a child run of kind "fold".

A fold is the run’s stitched account over the fold’s test segment (ADR 0033): its weights, equity, holdings and settlements are the run’s, cut to window, and are not written again.

Attributes:

indexint

The fold’s index in the walk.

windowtuple of str

The fold’s first and last bar, as bar labels.

metricsdict

The fold’s metrics.

trained_runPath or str

The fold’s trained unit.

data_fingerprintMapping

What the fold’s prediction read, as its DataRecorder recorded it.