Skip to main content

limen.transforms

Hold stateless transform helpers for target construction and data cleanup.

Canonical docs​

What this package owns​

Owns stateless DataFrame transforms. Does not own fitted feature scaling, higher-level feature engineering, experiment orchestration, or post-model helpers such as probability calibration and threshold optimization (see limen.calibration).

Key entry points​

Entry pointUse caseNotes
shift_column_transformForward-looking target columnCommon inside manifest target blocks
zscore_transformFrame-wide normalization stepStateless per call
winsorize_transform, mad_transform, quantile_trim_transformClipping or row filtering before modelingThese do not carry their own train-only state

Adjacent modules​

  • limen.features and limen.indicators produce the columns these transforms often operate on.
  • limen.scalers handles fitted train-only scaling, which is a separate concern from this package.
  • limen.calibration handles probability calibration and threshold optimization after model training.
  • limen.experiment.Manifest provides the most common place to attach target transforms.

Quick orientation​

transforms/
├── shift_column_transform.py        # Forward-looking target shifts
├── zscore_transform.py              # Stateless DataFrame normalization
├── winsorize_transform.py           # Percentile clipping
├── mad_transform.py                 # MAD-based clipping
└── quantile_trim_transform.py       # Quantile-based row trimming

Things to know​

  • All helpers in this package are intentionally stateless per call.
  • For fit-on-train and apply-to-all semantics, use a manifest fit parameter or a scaler rather than assuming a transform will remember state.
  • shift_column_transform with shift=-1 creates a future-looking label and therefore introduces a trailing null row that later prep stages drop.