omniloader.transforms.crop
==========================

.. py:module:: omniloader.transforms.crop

.. autoapi-nested-parse::

   Sequence cropping transforms (random window for train, center for eval).

   Both crops take a fixed-length window along the sequence axis. A **single shared
   offset** is applied across every targeted sequence key, so time-aligned features
   and per-step targets stay in sync. Keys shorter than the requested window are
   padded back to ``length`` with their placeholder value.



Classes
-------

.. autoapisummary::

   omniloader.transforms.crop.RandomCrop
   omniloader.transforms.crop.CenterCrop


Module Contents
---------------

.. py:class:: RandomCrop(length: int, keys: collections.abc.Sequence[str] | None = None, schema: omniloader.schema.spec.UnifiedSchema | None = None)

   Bases: :py:obj:`omniloader.transforms.base.Transform`


   Crop a random contiguous window of the sequence (training augmentation).

   :param length: Window length along the sequence axis.
   :param keys: Sequence keys to crop with a shared offset. When ``None``, every
                sequence key in the schema (features *and* aligned targets) is
                cropped together (requires ``schema``).
   :param schema: Schema used to resolve ``keys`` when it is ``None``.


   .. py:method:: apply(sample: omniloader.transforms.base.Sample, generator: torch.Generator | None) -> omniloader.transforms.base.Sample

      Crop all targeted keys at one random offset within the valid region.



.. py:class:: CenterCrop(length: int, keys: collections.abc.Sequence[str] | None = None, schema: omniloader.schema.spec.UnifiedSchema | None = None)

   Bases: :py:obj:`omniloader.transforms.base.Transform`


   Crop the centered window of the sequence (deterministic; used for eval).

   :param length: Window length along the sequence axis.
   :param keys: Sequence keys to crop. When ``None``, every sequence key in the
                schema is cropped (requires ``schema``).
   :param schema: Schema used to resolve ``keys`` when it is ``None``.


   .. py:method:: apply(sample: omniloader.transforms.base.Sample, generator: torch.Generator | None) -> omniloader.transforms.base.Sample

      Crop all targeted keys at the centered offset within the valid region.