Utils

systematica.signals.utils

clean_signals_1d_nb

clean_signals_1d_nb(
    long_entries: numpy.ndarray,
    long_exits: numpy.ndarray,
    short_entries: numpy.ndarray,
    short_exits: numpy.ndarray,
) ‑> Tuple[numpy.ndarray, ...]

Cleans 1d entry and exit signals by ensuring that positions are entered and exited in a logical manner.

The function prevents overlapping or redundant signals, ensuring that a position is opened only when there is no existing position and closed only when an open position exists.

Parameters:

Name	Type	Default	Description
long_entries	tp.Array1d	--	Boolean array indicating long entry signals.
long_exits	tp.Array1d	--	Boolean array indicating long exit signals.
short_entries	tp.Array1d	--	Boolean array indicating short entry signals.
short_exits	tp.Array1d	--	Boolean array indicating short exit signals.

Returns:

Type	Description
tuple of tp.Array1d	Tuple containing cleaned boolean arrays for long entries, long exits, short entries, and short exits.

clean_signals_nb

clean_signals_nb(
    long_entries: numpy.ndarray,
    long_exits: numpy.ndarray,
    short_entries: numpy.ndarray,
    short_exits: numpy.ndarray,
) ‑> Tuple[numpy.ndarray, ...]

Cleans 2d entry and exit signals by ensuring that positions are entered and exited in a logical manner for each column (asset/time series).

The function prevents overlapping or redundant signals for each column, ensuring that a position is opened only when there is no existing position and closed only when an open position exists.

Parameters:

Name	Type	Default	Description
long_entries	tp.Array2d	--	Boolean array indicating long entry signals (time, assets).
long_exits	tp.Array2d	--	Boolean array indicating long exit signals (time, assets).
short_entries	tp.Array2d	--	Boolean array indicating short entry signals (time, assets).
short_exits	tp.Array2d	--	Boolean array indicating short exit signals (time, assets).

Returns:

Type	Description
tuple of tp.Array2d	Tuple containing cleaned boolean arrays for long entries, long exits, short entries, and short exits, all in the same shape as input.

fshift_1d_nb

fshift_1d_nb(
    long_entries: numpy.ndarray,
    long_exits: numpy.ndarray,
    short_entries: numpy.ndarray,
    short_exits: numpy.ndarray,
    n: int = 1,
) ‑> Tuple[numpy.ndarray, ...]

Shift forward 1d signal arrays by n positions.

Numba equivalent to pd.Series(arr).shift(n).

Parameters:

Name	Type	Default	Description
long_entries	tp.Array1d	--	Boolean array indicating long entry signals.
long_exits	tp.Array1d	--	Boolean array indicating long exit signals.
short_entries	tp.Array1d	--	Boolean array indicating short entry signals.
short_exits	tp.Array1d	--	Boolean array indicating short exit signals.
n	int	1	Number of steps. Defaults to 1.

Returns:

Type	Description
Signals	shifted signals.

fshift_nb

fshift_nb(
    long_entries: numpy.ndarray,
    long_exits: numpy.ndarray,
    short_entries: numpy.ndarray,
    short_exits: numpy.ndarray,
    n: int = 1,
) ‑> Tuple[numpy.ndarray, ...]

Shift forward 2d signal arrays by n positions.

Numba equivalent to pd.Series(arr).shift(n).

Parameters:

Name	Type	Default	Description
long_entries	tp.Array1d	--	Boolean array indicating long entry signals.
long_exits	tp.Array1d	--	Boolean array indicating long exit signals.
short_entries	tp.Array1d	--	Boolean array indicating short entry signals.
short_exits	tp.Array1d	--	Boolean array indicating short exit signals.
n	int	1	Number of steps. Defaults to 1.

Returns:

Type	Description
Signals	shifted signals.

prepare_signals_nb

prepare_signals_nb(
    signal_func: Callable,
) ‑> Callable

Wraps a signal function to automatically create threshold arrays for predictions.

The wrapper ensures that input arrays match the shape of the model_output array and converts them to floating-point type for further processing.

Parameters:

Name	Type	Default	Description
signal_func	tp.Callable	--	Function that processes signals using predefined thresholds.

Returns:

Type	Description
tp.Callable	A Numba JIT-compiled function that applies the given signal function with correctly shaped and typed input arrays.