dubfi.fluxes.dataprovider_mpi
=============================

.. py:module:: dubfi.fluxes.dataprovider_mpi

.. autoapi-nested-parse::

   MPI parent module for flux inversion using in-situ data.

   This module provides tools for orchestrating the MPI parallelized
   flux inversion. Worker processes read data and configuration from
   files, controlled by this module.

   .. codeauthor:: Valentin Bruch, DWD <valentin.bruch@dwd.de>
   .. versionchanged:: 0.1.2 (renamed module)
   .. versionadded:: 0.1.0 (initial release)



Classes
-------

.. autoapisummary::

   dubfi.fluxes.dataprovider_mpi.MpiInsituPostR
   dubfi.fluxes.dataprovider_mpi.MpiDistMecReader


Functions
---------

.. autoapisummary::

   dubfi.fluxes.dataprovider_mpi.segments_from_config
   dubfi.fluxes.dataprovider_mpi.configs_agree


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

.. py:function:: segments_from_config(config, coords, num_segments, debug_plot = False)

   Compute segments for distributing data to MPI worker processes.

   :param config: inversion configuration
   :type config: dict
   :param coords: coordinates as provided by
                  :func:`dubfi.fluxes.readobs.coordinates_from_config`
   :type coords: dict
   :param num_segments: number of segments (must be positive)
   :type num_segments: int
   :param debug_plot: show a plot to illustrate the result
   :type debug_plot: bool, default=False

   :returns: collection of segments
   :rtype: Segments


.. py:function:: configs_agree(cfg1, cfg2)

   Check whether two inversion configurations are compatible.


.. py:class:: MpiInsituPostR(seg)

   Bases: :py:obj:`dubfi.linalg.mpi_parent.MpiDensePostR`


   MPI parent class for R matrix generated from in-situ observations.

   This class seems trivial, but the corresponding worker class
   differs from the one for :class:`MpiDensePostR`.


.. py:class:: MpiDistMecReader(seg)

   Bases: :py:obj:`dubfi.linalg.mpi_parent.MpiObj`, :py:obj:`dubfi.fluxes.dataprovider.AbstractDataProvider`


   MPI parent data provider.

   Prepare the inversion including worker processes.


   .. py:property:: config
      :type: dict


      Inversion configuration.


   .. py:property:: coords
      :type: dict


      Data coordinates, see :func:`dubfi.fluxes.readobs.coordinates_from_config`.


   .. py:method:: fromconfig(config, target_cfg, num_segments = -1)
      :classmethod:


      Create from inversion configuration.

      This reads the data and prepares the inversion. The worker
      processes are also initialized, but they must be created
      before using CTX.init().

      :param config: inversion configuration
      :type config: dict
      :param target_cfg: path to target configuration file. The configuration will
                         be written to this file path for the worker processes.
      :type target_cfg: str
      :param num_segments: number of segments, defaults to the number of available
                           MPI worker processes
      :type num_segments: int, optional



   .. py:method:: get_Y()

      Create Y: observation or observation minus model vector.



   .. py:method:: get_H()

      Create H: parametrized model vector ("observation operator").



   .. py:method:: get_R()

      Create R: parametrized error covariance matrix of model data mismatch.