MODULE REFERENCE

transform

anomx.data.base.transform

Contracts for explicit, reusable DataFrame preprocessing.

Transform

Transform

Bases: ABC

Transform materialized observations while retaining their index/metadata.

Fitting receives training data only. Implementations decide explicitly how to handle missing values and schema changes. Stateful temporal transforms must override transform_batches to carry context across batch boundaries.

Members

fitmethod

# Transform.fit Source ↗

fit(frame: pd.DataFrame) -> Self

Fit on training observations; stateless transforms need no fitting.

Signature parameters
Parameter	Type	Default
`frame`	`pd.DataFrame`	`required`

Returns Self

transformmethod

# Transform.transform Source ↗

transform(frame: pd.DataFrame) -> pd.DataFrame

Transform a frame without modifying the caller's input.

Signature parameters
Parameter	Type	Default
`frame`	`pd.DataFrame`	`required`

Returns pd.DataFrame

fit_transformmethod

# Transform.fit_transform Source ↗

fit_transform(frame: pd.DataFrame) -> pd.DataFrame

Fit on a training frame and transform that same frame.

Signature parameters
Parameter	Type	Default
`frame`	`pd.DataFrame`	`required`

Returns pd.DataFrame

transform_batchesmethod

# Transform.transform_batches Source ↗

transform_batches(batches: Iterator[pd.DataFrame]) -> Iterator[pd.DataFrame]

Transform independent batches without refitting or materializing all rows.

Signature parameters
Parameter	Type	Default
`batches`	`Iterator[pd.DataFrame]`	`required`

Returns Iterator[pd.DataFrame]