quantlab.factor.predefined.fundamental_alpha.report_lags

quantlab.factor.predefined.fundamental_alpha.report_lags(values, key, lags, quarter_days, tolerance_days)

Return each column’s value k reports ago, keyed by its period end.

A report is a run of bars showing the same period end key (days since 1970, NaN where nothing is shown). On a bar of report r, the value k reports ago is the last value shown during report r - k, so it was public before the bar. It is NaN unless that report exists and its period ends k * quarter_days (within tolerance_days) before the current one.

Parameters:

valuesdict of str to numpy.ndarray

Panels on [time, symbol].

keynumpy.ndarray

The period end of each cell, [time, symbol].

lagsdict of str to tuple of int

The lags wanted per column of values.

quarter_days, tolerance_daysfloat

The spacing check.

Returns:

dict of str to numpy.ndarray

"<column>__q<k>" -> float32 panel, one per column and lag.

Examples

>>> key = np.array([[0.0], [0.0], [91.0], [182.0]])
>>> out = report_lags({"x": np.array([[1.0], [2.0], [3.0], [4.0]])}, key, {"x": (1,)}, 91.0, 10.0)
>>> out["x__q1"].ravel().tolist()
[nan, nan, 2.0, 3.0]