ezmsg.sigproc.scaler#

Adaptive standard scaling using exponentially weighted moving statistics.

Functions

scaler(time_constant=1.0, axis=None)[source]#

Create a RiverAdaptiveStandardScalerTransformer with the given parameters.

Parameters:

time_constant (float)
axis (str | None)

Return type:

RiverAdaptiveStandardScalerTransformer

scaler_np(time_constant=1.0, axis=None)[source]#

Parameters:

time_constant (float)
axis (str | None)

Return type:

AdaptiveStandardScalerTransformer

Classes

class AdaptiveStandardScaler(*args, settings=None, **kwargs)[source]#

Bases: BaseTransformerUnit[AdaptiveStandardScalerSettings, AxisArray, AxisArray, AdaptiveStandardScalerTransformer]

Parameters:: settings (Settings | None)

SETTINGS#: alias of AdaptiveStandardScalerSettings

INPUT_ACCUMULATE = InputStream:unlocated[bool]()#

async on_accumulate(accumulate)[source]#

Parameters:: accumulate (bool)
Return type:: None

class AdaptiveStandardScalerSettings(time_constant: float = 1.0, axis: str | None = None, accumulate: bool = True)[source]#

Bases: EWMASettings

Parameters:

time_constant (float)
axis (str | None)
accumulate (bool)

__init__(time_constant=1.0, axis=None, accumulate=True)#

Parameters:

time_constant (float)
axis (str | None)
accumulate (bool)

Return type:

None

class AdaptiveStandardScalerState[source]#

Bases: object

samps_ewma: EWMATransformer | None = None#

vars_sq_ewma: EWMATransformer | None = None#

alpha: float | None = None#

class AdaptiveStandardScalerTransformer(*args, **kwargs)[source]#

Bases: BaseStatefulTransformer[AdaptiveStandardScalerSettings, AxisArray, AxisArray, AdaptiveStandardScalerState]

NONRESET_SETTINGS_FIELDS: ClassVar[frozenset[str]] = frozenset({'accumulate'})#

update_settings(new_settings)[source]#

Apply new settings to this processor, requesting a state reset if needed.

Diffs new_settings against the current self.settings. If every changed field is listed in NONRESET_SETTINGS_FIELDS, self.settings is simply rebound and the processor keeps running. Otherwise, _request_reset() is called so the next message triggers a fresh _reset_state (for stateful subclasses). Override for finer-grained control than the class-level allow-list provides.

Parameters:: new_settings (AdaptiveStandardScalerSettings)
Return type:: None

property accumulate: bool#: Whether to accumulate statistics from incoming samples.

class RiverAdaptiveStandardScalerSettings(time_constant: float = 1.0, axis: str | None = None)[source]#

Bases: Settings

Parameters:

time_constant (float)
axis (str | None)

time_constant: float = 1.0#: Decay constant tau in seconds.

axis: str | None = None#: The name of the axis to accumulate statistics over.

__init__(time_constant=1.0, axis=None)#

Parameters:

time_constant (float)
axis (str | None)

Return type:

None

class RiverAdaptiveStandardScalerState[source]#

Bases: object

scaler: Any = None#

axis: str | None = None#

axis_idx: int = 0#

class RiverAdaptiveStandardScalerTransformer(*args, **kwargs)[source]#

Bases: BaseStatefulTransformer[RiverAdaptiveStandardScalerSettings, AxisArray, AxisArray, RiverAdaptiveStandardScalerState]

Apply the adaptive standard scaler from river.

This processes data sample-by-sample using River’s online learning implementation. For a vectorized EWMA-based alternative, see AdaptiveStandardScalerTransformer.