MODULE REFERENCE

processing

anomx.data.processing

Reusable preprocessing for native time series and dataset frames.

Pipeline composes TimeSeries transformers, while ProcessingPipeline composes DataFrame Transform steps for materialized data and batch streams. Transformers, covariate encoders and dynamic time warping live in their named subpackages. The numerical baseline is derived from Unit8 Darts under Apache 2.0; its implementations are owned by Anomx and do not import the Darts package.

Public exports

Import these names directly from anomx.data.processing.

Pipeline

ProcessingPipeline

Fit transforms sequentially on training data and reuse them for inference.

Resample

Aggregate a complete time-indexed selection at an explicit frequency.

Transform

Transform materialized observations while retaining their index/metadata.