bug(recipe): should we not transform the last step in recipe during recipe.fit()

[jitingxu1]

should we not transform the last step during fit(): https://github.com/ibis-project/ibis-ml/blob/main/ibis_ml/core.py#L512

    def _fit_table(
        self, table: ir.Table, targets: tuple[str, ...] = (), index: str | None = None
    ) -> None:
        metadata = Metadata(targets=targets)

        if index is not None:
            table = table.drop(index)

        for step in self.steps:
            step.fit_table(table, metadata)
            table = step.transform_table(table)

        self.metadata_ = metadata

[jitingxu1]

if it used in sklearn or other pipeline, I think it is needed. But it looks weird, if we use recipe independently...it already does all the transformation work in fit()