Preprocessing

Ver. 2.3.0 (2025-06-25)

This module provides a basic implementation of a boundary detector.

class mlpro.oa.streams.tasks.boundarydetectors.basic.BoundaryDetector(p_name: str = None, p_range_max=1, p_ada: bool = True, p_duplicate_data: bool = False, p_visualize: bool = False, p_logging=True, p_boundary_provider: BoundaryProvider = None, **p_kwargs)

Bases: OAStreamTask, BoundaryProvider

This class provides the functionality of boundary observation of incoming instances.

Parameters:

p_name (str, Optional.) – Name of the task.
p_range_max – Processing range of the task. Default is thread.
p_ada (bool) – True if the task has adaptivity. Default is True.
p_duplicate_data (bool) – If True, instances will be duplicated before processing. Default = False.
p_visualize (bool) – True to turn on the visualization.
p_logging – Logging level for the task, default is Log all.
p_boundary_provider (BoundaryProvider = None) – Optional boundary provider, used for service call on reverse adaptation.

C_TYPE = 'Boundary Detector'

C_PLOT_ND_XLABEL = 'Features'

C_PLOT_ND_YLABEL = 'Boundaries'

C_PLOT_STANDALONE: bool = True

C_PLOT_VALID_VIEWS: list = ['ND']

C_PLOT_DEFAULT_VIEW: str = 'ND'

_run(p_instances: Dict[int, Tuple[str, Instance]])

Method to run the boundary detector task

Parameters:: p_instances (InstDict) – Instances to be processed.

adapt(p_instances: Dict[int, Tuple[str, Instance]]) → bool

Adapts the model by calling the custom method _adapt().

Parameters:: **p_kwargs – All parameters that are needed for the adaption. Depends on the specific higher context.
Returns:: adapted – True, if something has been adapted. False otherwise.
Return type:: bool

_init_data_structures(p_instance: Instance)

_adapt(p_instance_new: Instance) → bool

Extends boundaries based on a new instance.

Parameters:: p_instance_new (Instance) – New instance to be processed.
Returns:: True, if at least one boundary value has been extended. False otherwise.
Return type:: bool

_adapt_reverse(p_instance_del: Instance) → bool

Determines the boundaries to be reduced based on an outdated instance.

Parameters:: p_instance_del (Instance) – Deleted instance to be processed.
Returns:: True, if at least one boundary value needs to be reduced. False otherwise.
Return type:: bool

_adapt_post() → OAStreamAdaptationType: …

get_boundaries(p_dim: int = None, p_side: BoundarySide = None) → ndarray[tuple[Any, ...], dtype[float64]] | float

Returns the current value boundaries of internally stored data. The result can be reduced by the optional parameters p_side, p_dim. If both parameters are specified, the result is a float.

Parameters:

p_dim (int = None) – Optionally reduces the result to a particular dimension.
p_side (BoundarySide = None) – Optionally reduces the result to upper or lower boundaries. See class BoundarySide for possible values.

Returns:

Returns the current boundaries of the data. The return value depends on the combination of the optional parameters: - If neither p_side nor p_dim is specified: returns the full 2×n array. - If only p_side is specified: returns a 1D array with values for all dimensions. - If only p_dim is specified: returns a 1D array with both lower and upper bounds. - If both p_side and p_dim are specified: returns a single float value.

Return type:

Union[Boundaries, float]

_init_plot_nd(p_figure: Figure, p_settings: PlotSettings)

Custom method to initialize plot for Boundary Detectors tasks for N-dimensional plotting.

Parameters:

p_figure (Figure) – Figure to host the plot
p_settings (PlotSettings) – PlotSettings objects with specific settings for the plot

_update_plot_nd(p_settings: PlotSettings, p_instances: Dict[int, Tuple[str, Instance]], **p_kwargs) → bool

N-dimensional plotting for Boundary Detector using vertical bars and mean markers.

Parameters:

p_settings (PlotSettings) – Plot settings object including matplotlib axes.
p_instances (InstDict) – Optional stream instances (not used here).
p_kwargs (dict) – Further optional parameters.

Returns:

True, if changes on the plot require a refresh of the figure. False otherwise.

Return type:

bool

Ver. 2.0.0 (2025-06-30)

This module provides implementation for adaptive normalizers for MinMax Normalization.

class mlpro.oa.streams.tasks.normalizers.minmax.NormalizerMinMax(p_name: str = None, p_range_max=1, p_ada: bool = True, p_duplicate_data: bool = False, p_visualize: bool = False, p_logging=True, p_dst_boundaries: list = [-1, 1], **p_kwargs)

Bases: NormalizerMinMax, OAStreamNormalizer

Class with functionality for adaptive normalization of instances using MinMax Normalization.

Parameters:

p_name (str = None,) – Optional name of the task.
p_range_max – Processing range of the task, default is a Thread.
p_ada (bool) – True if the task has adaptivity, default is true.
p_duplicate_data (bool) – If True, instances will be duplicated before processing. Default = False.
p_visualize – True for visualization, false by default.
p_logging – Logging level of the task. Default is Log.C_LOG_ALL
p_dst_boundaries (list = [-1,1]) – Explicit list of (low, high) destination boundaries. Default is [-1, 1].
**p_kwargs – Additional task parameters

C_NAME = 'MinMax'

_run(p_instances: Dict[int, Tuple[str, Instance]])

Runs MinMax Normalizer task for normalizing stream instances.

Parameters:: p_instances (InstDict) – Instances to be processed

_adapt_on_event(p_event_id: str, p_event_object: Event) → bool

Custom method to adapt the MinMax normalizer parameters based on event raised by Boundary object for changed boundaries.

Parameters:

p_event_id (str) – Event id of the raised event
p_event_obj (Event) – Event object that raises the corresponding event

Returns:

Returns True, if the task has adapted. False otherwise.

Return type:

bool

_update_plot_data_2d(): Updates the 2D plot data after parameter changes by renormalizing the existing points.

_update_plot_data_3d(): Method to update the 3d plot for Normalizer. Extended to renormalize the obsolete data on change of parameters.

_update_plot_data_nd(): Method to update the nd plot for Normalizer. Extended to renormalize the obsolete data on change of parameters.

Ver. 2.1.0 (2025-07-05)

This module provides implementation for adaptive normalizers for ZTransformation

class mlpro.oa.streams.tasks.normalizers.ztrans.NormalizerZTransform(p_name: str = None, p_range_max=1, p_ada: bool = True, p_duplicate_data: bool = False, p_visualize=False, p_logging=True, **p_kwargs)

Bases: NormalizerZTrans, OAStreamNormalizer

Online-adaptive normalization of instances with Z-transformation.

Parameters:

p_name (str, optional) – Name of the task.
p_range_max – Processing range of the task, default is a Thread.
p_ada – True if the task has adaptivity, default is true.
p_duplicate_data (bool) – If True, instances will be duplicated before processing. Default = False.
p_visualize – True for visualization, false by default.
p_logging – Logging level of the task. Default is Log.C_LOG_ALL
**p_kwargs – Additional task parameters

C_NAME = 'ZTrans'

_run(p_instances: Dict[int, Tuple[str, Instance]])

Custom method to for run Z-transform task for normalizing new instances and denormalizing deleted instances.

Parameters:: p_instances (InstDict) – Stream instances to be processed

_adapt(p_instance_new: Instance) → bool

Custom method to for adapting of Z-transform parameters on new instances.

Parameters:: p_instance_new (Instance) – Instance to be adapted on.
Returns:: adapted – Returns True, if task has adapted.
Return type:: bool

_adapt_reverse(p_instance_del: Instance) → bool

Custom method to for adapting of Z-transform parameters on deleted instances.

Parameters:: p_instance_del (Instance) – Instance to be adapted on.
Returns:: adapted – Returns True, if task has adapted.
Return type:: bool

_update_plot_data_2d(): Updates the 2D plot data after parameter changes by renormalizing the existing points.

_update_plot_data_3d(): Method to update the 3d plot for Normalizer. Extended to renormalize the obsolete data on change of parameters.

_update_plot_data_nd(): Method to update the nd plot for Normalizer. Extended to renormalize the obsolete data on change of parameters.

_update_plot_data(): Updates the plot data.