Base trait for setting/accessing debug flags.
Trait for defining whether interaction columns have been set for the application of Feature Interactions
Trait for defining whether interaction columns have been set for the application of Feature Interactions
0.6.2
0.6.1 trait for decorating all pipeline stages with pipeline ID. Helpful when troubleshooting logs with a given pipeline ID (eg, fetched from MLflow)
Transformer for creating interacted feature fields based on FeatureInteraction module
Transformer for creating interacted feature fields based on FeatureInteraction module
0.6.2
A WithNoopsStage transformer stage that is helpful to repartition a DataFrame coming out of any pipeline stage
Marker interface to signify any transformer extending this trait will not alter an input dataset.
Marker interface to signify any transformer extending this trait will not alter an input dataset. This is only for the edge cases where it is required to do an external Ops before pipeline execution can continue. An example would be to do Mlflow params Validation before training continues. Helpful in scenarios where fail-fast feature is needed Example transformers are DropTempTableTransformer, MlFlowLoggingValidationStageTransformer.
NOTE: Noops implies no changes to the input Dataset, but the implementation can result in a change to an external state
0.6.1 Utility for reporting pipeline progress to MLflow
Base trait for setting/accessing debug flags. Meant to be extended by all pipeline stages, which inherit pipeline stage logging by default