emmi_inference.preprocess
=========================

.. py:module:: emmi_inference.preprocess


Classes
-------

.. autoapisummary::

   emmi_inference.preprocess.Compose
   emmi_inference.preprocess.AddBatch
   emmi_inference.preprocess.NormalizeAny


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

.. py:class:: Compose(*ops)

   Chain multiple preprocessing ops.

   Each op must be a callable that accepts and returns the same data structure (e.g., a `torch.Tensor` or a
   `dict[str, torch.Tensor]`). Ops that are `None` are ignored, which makes conditional composition convenient.

   .. attribute:: ops

      A list of callables applied in order.


   .. py:attribute:: ops


.. py:class:: AddBatch(dim = 0)

   Insert a batch dimension.

   Works for a single `torch.Tensor` or a mapping of tensors; in the latter case, the dimension is inserted for every
   value.

   .. attribute:: dim

      Dimension index at which to insert the batch.

   Configure the batch insertion.

   :param dim: Dimension index along which to insert the new batch axis. Defaults to 0.


   .. py:attribute:: dim
      :value: 0



.. py:class:: NormalizeAny

   Normalize tensors or dicts of tensors using scalar or per-channel stats.

   This transform supports two modes:

   1. Tensor input: `mean`/`std` can be a scalar or 1-D sequences for per-channel normalization. Broadcasting is performed along `dim`.
   2. Dict input: `per_key` maps keys to their own `NormalizeAny` to apply to each value. Keys without a configured normalizer are left unchanged.

   .. attribute:: mean

      Scalar or 1-D sequence of means for tensor normalization. If `None`, no normalization is applied.

   .. attribute:: std

      Scalar or 1-D sequence of standard deviations matching `mean`. Zero values are clamped to 1 to avoid
      division by zero.

   .. attribute:: per_key

      Optional mapping from key to a `NormalizeAny` applied to dict inputs.

   .. attribute:: dim

      Feature/channel dimension for broadcasting vector stats.


   .. py:attribute:: mean
      :type:  float | collections.abc.Sequence[float] | None
      :value: None



   .. py:attribute:: std
      :type:  float | collections.abc.Sequence[float] | None
      :value: None



   .. py:attribute:: per_key
      :type:  collections.abc.Mapping[str, NormalizeAny] | None
      :value: None



   .. py:attribute:: dim
      :type:  int
      :value: -1