apache_beam.ml.anomaly.thresholds module
- class apache_beam.ml.anomaly.thresholds.FixedThreshold(*args, **kwargs)[source]
Bases:
ThresholdFnApplies a fixed cutoff value to anomaly scores.
This ThresholdFn is stateless and uses a pre-defined cutoff value to classify anomaly scores. Scores below the cutoff are considered normal, while scores at or above the cutoff are classified as outliers.
- Parameters:
cutoff (float) – The fixed threshold value. Anomaly scores at or above this value will be labeled as outliers.
**kwargs – Additional keyword arguments to be passed to the base ThresholdFn constructor.
- property is_stateful: bool
Indicates whether this ThresholdFn is stateful.
- Returns:
Always False for FixedThreshold as it is stateless.
- Return type:
- property threshold: float
Returns the fixed cutoff threshold value.
- Returns:
The fixed threshold value.
- Return type:
- apply(score: float | None) int | None[source]
Applies the fixed threshold to an anomaly score.
Classifies the given anomaly score as normal or outlier based on the predefined cutoff.
- Parameters:
score (Optional[float]) – The input anomaly score.
- Returns:
- The anomaly label:
normal_label if the score is less than the threshold.
outlier_label if the score is at or above the threshold.
missing_label if the score is NaN (detector not ready).
None if the score is None (detector ready, but unable to produce score).
- Return type:
Optional[int]
- FixedThreshold__spec_type = 'FixedThreshold'
- classmethod from_spec(spec: Spec, _run_init: bool = True) Self | type[Self]
Generate a Specifiable subclass object based on a spec.
- Parameters:
spec – the specification of a Specifiable subclass object
_run_init – whether to call __init__ or not for the initial instantiation
- Returns:
the Specifiable subclass object
- Return type:
Self
- run_original_init() None
Execute the original __init__ method with its saved arguments.
For instances of the Specifiable class, initialization is deferred (lazy initialization). This function forces the execution of the original __init__ method using the arguments captured during the object’s initial instantiation.
- classmethod spec_type()
- to_spec() Spec
Generate a spec from a Specifiable subclass object.
- Returns:
The specification of the instance.
- Return type:
- classmethod unspecifiable()
- class apache_beam.ml.anomaly.thresholds.QuantileThreshold(*args, **kwargs)[source]
Bases:
ThresholdFnApplies a quantile-based dynamic threshold to anomaly scores.
This ThresholdFn is stateful and uses a quantile tracker to dynamically determine the threshold for anomaly detection. It estimates the specified quantile of the incoming anomaly scores and uses this quantile value as the threshold.
The threshold adapts over time as more data is processed, making it suitable for scenarios where the distribution of anomaly scores may change.
- Parameters:
quantile (Optional[float]) – The quantile to be tracked (e.g., 0.95 for the 95th percentile). This value determines the dynamic threshold. Defaults to 0.95.
quantile_tracker (Optional[BufferedQuantileTracker]) – An optional pre-initialized quantile tracker. If provided, this tracker will be used; otherwise, a BufferedSlidingQuantileTracker will be created with a default window size of 100.
**kwargs – Additional keyword arguments to be passed to the base ThresholdFn constructor.
- property is_stateful: bool
Indicates whether this ThresholdFn is stateful.
- Returns:
Always True for QuantileThreshold as it is stateful.
- Return type:
- property threshold: float
Returns the current quantile-based threshold value.
- Returns:
- The dynamically calculated threshold value based on the quantile
tracker.
- Return type:
- QuantileThreshold__spec_type = 'QuantileThreshold'
- apply(score: float | None) int | None[source]
Applies the quantile-based threshold to an anomaly score.
Updates the quantile tracker with the given score and classifies the score as normal or outlier based on the current quantile threshold.
- Parameters:
score (Optional[float]) – The input anomaly score.
- Returns:
- The anomaly label:
normal_label if the score is less than the threshold.
outlier_label if the score is at or above the threshold.
missing_label if the score is NaN (detector not ready).
None if the score is None (detector ready, but unable to produce score).
- Return type:
Optional[int]
- classmethod from_spec(spec: Spec, _run_init: bool = True) Self | type[Self]
Generate a Specifiable subclass object based on a spec.
- Parameters:
spec – the specification of a Specifiable subclass object
_run_init – whether to call __init__ or not for the initial instantiation
- Returns:
the Specifiable subclass object
- Return type:
Self
- run_original_init() None
Execute the original __init__ method with its saved arguments.
For instances of the Specifiable class, initialization is deferred (lazy initialization). This function forces the execution of the original __init__ method using the arguments captured during the object’s initial instantiation.
- classmethod spec_type()
- to_spec() Spec
Generate a spec from a Specifiable subclass object.
- Returns:
The specification of the instance.
- Return type:
- classmethod unspecifiable()