apache_beam.ml.anomaly.univariate.base module

class apache_beam.ml.anomaly.univariate.base.BaseTracker

Bases: ABC

Abstract base class for all univariate trackers.

abstract push(x)

Push a new value to the tracker.

Parameters:

x – The value to be pushed.

abstract get()

Get the current tracking value.

Returns:

The current tracked value, the type of which depends on the specific tracker implementation.

class apache_beam.ml.anomaly.univariate.base.WindowMode(value)

Bases: Enum

Enum representing the window mode for windowed trackers.

LANDMARK = 1

operating on all data points from the beginning.

SLIDING = 2

operating on a fixed-size sliding window of recent data points.

class apache_beam.ml.anomaly.univariate.base.WindowedTracker(window_mode, **kwargs)

Bases: BaseTracker

Abstract base class for trackers that operate on a data window.

This class provides a foundation for trackers that maintain a window of data, either as a landmark window or a sliding window. It provides basic push and pop operations.

Parameters:

• window_mode – A WindowMode enum specifying whether the window is LANDMARK or SLIDING.

• **kwargs – Keyword arguments. For SLIDING window mode, window_size can be specified to set the maximum size of the sliding window. Defaults to 100.

push(x)

Adds a new value to the data window.

Parameters:

x – The value to be added to the window.

pop()

Removes and returns the oldest value from the data window (FIFO).

Returns:

The oldest value from the window.