abc
=============

.. py:module:: jaxns.samplers.abc

.. rubric:: :code:`jaxns.samplers.abc`


.. rubric:: Module Contents

.. py:data:: SamplerState

.. py:class:: EphemeralState

   Bases: :py:obj:`NamedTuple`


   .. py:attribute:: key
      :type:  jaxns.internals.types.PRNGKey


   .. py:attribute:: live_points_collection
      :type:  jaxns.nested_samplers.common.types.LivePointCollection


   .. py:attribute:: termination_register
      :type:  jaxns.nested_samplers.common.types.TerminationRegister


.. py:class:: AbstractSampler

   Bases: :py:obj:`abc.ABC`, :py:obj:`Generic`\ [\ :py:obj:`SamplerState`\ ]


   Helper class that provides a standard way to create an ABC using
   inheritance.


   .. py:method:: pre_process(ephemeral_state)

      Run this periodically on the state to produce a data pytree that can be used by the sampler, and
      updated quickly.

      :param ephemeral_state: the current state of the sampler

      :returns: any valid pytree



   .. py:method:: post_process(ephemeral_state, sampler_state)

      Post process the sampler state, after the sampler has been run. Should be quick.

      :param ephemeral_state: a sample collection post sample step
      :param sampler_state: data pytree produced by the sampler

      :returns: the updated sampler state



   .. py:method:: get_sample(key, log_L_constraint, sampler_state)

      Produce a single i.i.d. sample from the model within the log_L_constraint.

      :param key: PRNGkey
      :param log_L_constraint: the constraint to sample within
      :param sampler_state: the data pytree needed and produced by the sampler

      :returns: an i.i.d. sample, and phantom samples



   .. py:method:: num_phantom()
      :abstractmethod:


      The number of phantom samples produced by the sampler.

      :returns: number of phantom samples