dasf.pipeline.executors.ray
===========================

.. py:module:: dasf.pipeline.executors.ray

.. autoapi-nested-parse::

   Ray executor module. 



Attributes
----------

.. autoapisummary::

   dasf.pipeline.executors.ray.USE_RAY


Classes
-------

.. autoapisummary::

   dasf.pipeline.executors.ray.RayPipelineExecutor


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

.. py:data:: USE_RAY
   :value: True


.. py:class:: RayPipelineExecutor(address=None, port=6379, local=False, use_gpu=False, ray_kwargs=None)

   Bases: :py:obj:`dasf.pipeline.executors.base.Executor`


   A pipeline executor based on ray data flow.

   Parameters
   ----------
   address : str
       Address of the Dask scheduler, default=None.
   port : int
       Port of the Ray head, default=8786.
   local : bool
       Kicks off a new local Ray cluster, default=False.
   use_gpu : bool
       In conjunction with `local`, it kicks off a local CUDA Ray
       cluster, default=False.

   Constructor of the object RayPipelineExecutor. 


   .. py:attribute:: address


   .. py:attribute:: port


   .. py:property:: ngpus
      Return the number of GPUs in total.

      Returns
      -------
      ngpus : Number of GPUs in total



   .. py:property:: is_connected
      Return wether the executor is connected or not.

      Returns
      -------
      bool : if the executor is connected.



   .. py:method:: execute(fn, *args, **kwargs)

      Return wether the executor is connected or not.

      Parameters
      ----------
      fn : Callable
         Function to call when executor is performing a task.

      Returns
      -------
      ret : the same return of function `fn`.



   .. py:method:: __del__()

      Destructor of object.

      It also shutdowns Dask on Ray properly with `disable_dask_on_ray`.