foscat.xarray.statistics
========================

.. py:module:: foscat.xarray.statistics


Functions
---------

.. autoapisummary::

   foscat.xarray.statistics.backend_to_dict
   foscat.xarray.statistics.construct_backend
   foscat.xarray.statistics.stack_other_dims
   foscat.xarray.statistics.reference_statistics
   foscat.xarray.statistics.cross_statistics


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

.. py:function:: backend_to_dict(backend)

.. py:function:: construct_backend(backend, kwargs)

.. py:function:: stack_other_dims(arr, spatial_dim, batch_dim)

.. py:function:: reference_statistics(arr, *, parameters, spatial_dim='cells', variances=False, mask=None, norm=None, jmax=None)

   reference statistics for a single image

   :Parameters: * **arr** (:py:class:`xarray.DataArray`) -- Input image. For now, only 1D healpix is supported. Every dimension other than
                  the spatial dimension (see ``spatial_dim``) will be stacked.
                * **parameters** (:py:class:`Parameters`) -- The parameters for the scattering covariance transform.
                * **spatial_dim** (:py:class:`str`, *default*: ``"cells"``) -- The spatial dimension.
                * **variances** (:py:class:`bool`, *default*: :py:obj:`False`) -- Whether to compute the variances of the statistic values.
                * **mask** (:py:class:`xarray.DataArray`, *optional*) -- Mask out certain regions. Not implemented yet.
                * **norm** (``{"auto", "self"}`` or :py:obj:`None`, *default*: :py:obj:`None`) -- Normalization method:
                  - None: no normalization
                  - "auto": normalize by the reference S2
                  - "self": normalize by the current S2


.. py:function:: cross_statistics(arr1, arr2, *, parameters, spatial_dim='cells', variances=False, mask=None, norm=None)

   cross statistics between two images

   :Parameters: * **arr1, arr2** (:py:class:`xarray.DataArray`) -- Input images. Must align exactly. For now, only 1D healpix is supported. Every
                  dimension other than the spatial dimension (see ``spatial_dim``) will be stacked.
                * **parameters** (:py:class:`Parameters`) -- The parameters for the scattering covariance transform.
                * **spatial_dim** (:py:class:`str`, *default*: ``"cells"``) -- The spatial dimension.
                * **variances** (:py:class:`bool`, *default*: :py:obj:`False`) -- Whether to compute the variances of the statistic values.
                * **mask** (:py:class:`xarray.DataArray`, *optional*) -- Mask out certain regions. Not implemented yet.
                * **norm** (``{"auto", "self"}`` or :py:obj:`None`, *default*: :py:obj:`None`) -- Normalization method:
                  - None: no normalization
                  - "auto": normalize by the reference S2
                  - "self": normalize by the current S2